Search is not available for this dataset
article
stringlengths 4.36k
149k
| summary
stringlengths 32
3.35k
| section_headings
listlengths 1
91
| keywords
listlengths 0
141
| year
stringclasses 13
values | title
stringlengths 20
281
|
|---|---|---|---|---|---|
Active DNA demethylation plays crucial roles in the regulation of gene expression in both plants and animals . In Arabidopsis thaliana , active DNA demethylation is initiated by the ROS1 subfamily of 5-methylcytosine-specific DNA glycosylases via a base excision repair mechanism . Recently , IDM1 and IDM2 were shown to be required for the recruitment of ROS1 to some of its target loci . However , the mechanism ( s ) by which IDM1 is targeted to specific genomic loci remains to be determined . Affinity purification of IDM1- and IDM2- associating proteins demonstrated that IDM1 and IDM2 copurify together with two novel components , methyl-CpG-binding domain protein 7 ( MBD7 ) and IDM2-like protein 1 ( IDL1 ) . IDL1 encodes an α-crystallin domain protein that shows high sequence similarity with IDM2 . MBD7 interacts with IDM2 and IDL1 in vitro and in vivo and they form a protein complex associating with IDM1 in vivo . MBD7 directly binds to the target loci and is required for the H3K18 and H3K23 acetylation in planta . MBD7 dysfunction causes DNA hypermethylation and silencing of reporter genes and a subset of endogenous genes . Our results suggest that a histone acetyltransferase complex functions in active DNA demethylation and in suppression of gene silencing at some loci in Arabidopsis .
DNA methylation is an important epigenetic mark conserved in many eukaryotes . Many studies have demonstrated that DNA methylation plays central roles in genome organization , genomic imprinting , transposon silencing , and gene expression [1 , 2 , 3] . DNA methylation levels are coordinately determined by methylation and demethylation reactions during development and reproduction in both plants and animals . In the model plant Arabidopsis thaliana , levels of symmetric CG and CHG methylation are maintained by DNA METHYLTRANSFERASE 1 ( MET1 ) and CHROMOMETHYLASE 3 ( CMT3 ) , respectively , during DNA replication [1] . In contrast , asymmetric CHH methylation needs to be established de novo during each cell cycle [1] . DNA methylation is antagonized by an active DNA demethylation pathway catalyzed by a subfamily of bi-functional DNA glycosylase/lyases represented by REPRESSOROF SILENCING 1 ( ROS1 ) and DEMETER ( DME ) [4 , 5 , 6 , 7 , 8] . Although the RNA-directed DNA methylation pathway and its regulation are well-established ( Law and Jacobsen , 2010 ) , little is known about the locus-specific targeting of proteins involved in active DNA demethylation in plants . Previous studies suggested that the RNA-binding protein ROS3 is required for the recruitment of ROS1 to some target loci [9] . INCREASED DNA METHYLATION 1 ( IDM1 ) , a histone acetyltransferase , was recently identified as another important regulator of DNA demethylation that functions upstream of ROS1 [10] . With the assistance of IDM2 , an alpha-crystallin domain protein , IDM1 recognizes chromatin regions that are marked by CG methylation and low levels of histone 3 lysine 4 ( H3K4 ) and arginine 2 ( H3R2 ) methylation and subsequently catalyzes H3K18 and H3K23 acetylation , which triggers ROS1-mediated DNA demethylation [10 , 11] . However , the molecular mechanism ( s ) by which IDM1 is targeted to these specific genomic loci remains largely unknown . In plants and animals , a set of proteins containing a methyl-CpG-binding domain ( MBD ) are capable of specifically recognizing and binding methylated DNA [12 , 13] . They are believed to function as an interpreter of DNA methylation signals [13] . In mammals , MBD proteins MeCP2 , MBD1 , MBD2 and MBD4 have methyl-CpG-binding activity and function in targeting of maintenance DNA methylation machinery , H3K9 methylation , transcriptional silencing and X chromosome inactivation [14] . They usually execute their functions by binding to other proteins . For example , MeCP2 was demonstrated to interact with Sin3 and histone deacetylases to transcriptionally silence methylated chromosome , suggesting a direct relationship between histone modification and methylated DNA [15] . In Zebrafish , MBD4 , a protein that possesses DNA glycosylase activity , was reported be involved in active DNA demethylation [16] . The Arabidopsis genome contains 13 genes encoding putative MBD proteins orthologous to their mammalian counterparts [17] . Only AtMBD5 , AtMBD6 and AtMBD7 bind methylated CpG nucleotide in vitro [17] . AtMBD5 , AtMBD6 and AtMBD7 localize in the nuclei , and assemble at highly methylated chromocenters . while , in DNA hypomethylation mutants ddm1 and met1 they dissociate from chromocenters , indicating their methyl-CpG-binding activity in vivo [18] . Different from AtMBD7 , AtMBD5 and AtMBD6 mainly localize to chromatin regions covered by ribosomal DNA ( rDNA ) gene clusters [18 , 19] . Snf2 family nucleosome remodeler DDM1 is able to bind AtMBD5 and AtMBD6 in vitro and facilitate their heterochromatin localization [18] . AtMBD7 , belonging to a unique class of plant MBD proteins that have multiple MBD domains , was hypothesized to bind multiple methylated sites that are not close to each other and promote the formation and/or maintainance of a microenvironment in chromatin [13] . To find out how IDM1 is targeted to specific loci , we identified IDM1 and IDM2 interacting proteins by an approach that combines affinity purification and mass spectrometry . We found that IDM1 and IDM2 copurify with each other and with two novel proteins , IDL1 and MBD7 . MBD7 directly binds to the target loci and is required for the H3K18 and H3K23 acetylation in planta . The mbd7 mutants exhibit DNA hypermethylation at thousands of genomic loci and show increased transcriptional silencing of reporter genes and some endogenous genes . Our results reveal novel components involved in DNA demethylation and novel functions for MBD proteins in plants .
To better understand the mechanism of IDM1 targeting in active DNA demethylation , we set out to purify the protein complex associated with IDM1 . Transgenic Arabidopsis plants expressing IDM1-3HA-YFP driven by native IDM1 promoter in idm1-3 mutant background were generated . Previous studies showed that IDM1 dysfunction caused the kanamycin sensitive growth phenotype and silencing of NPT II transgene [20] . The IDM1-3HA-YFP was able to complement the kanamycin sensitive growth phenotype ( S1A Fig ) . Real time PCR results showed that the NPT II expression was also restored in the transgenic plants ( S1B Fig ) , suggesting that the IDM1-3HA-YFP fusion protein was functional in vivo . The IDM1-3HA-YFP expression in the transgenic lines was confirmed by Western blot with the HA antibody ( S1C Fig ) . Co-immunoprecipitation ( co-IP ) of IDM1-3HA-YFP and its interacting proteins was performed using anti-HA antibodies and flower extracts . Tryptic peptides derived from the purified proteins were sequenced by Velos Pro Orbitrap Elite tandem mass spectrometry ( MS/MS ) ( Table 1 and S1 Table ) . As expected , the MS analysis identified peptides corresponding to a known IDM1 interacting protein IDM2 [11] . Peptides corresponding to the IDM2-related protein IDM2 like 1 ( IDL1 ) and methyl-CpG-binding domain protein 7 ( MBD7 ) were also identified ( Table 1 ) , suggesting these two proteins may be new components of active DNA demethylation pathway . Meanwhile , we carried out affinity purification of 6Myc-IDM2 using transgenic plants overexpressing 6Myc-IDM2 in idm2-1 mutants [11] and 6Myc-IDL1 using transgenic plants overexpressing 6Myc-IDL1 in wild type plants . Mass spectrometric analysis revealed that IDM2 and IDL1 copurify with each other . We also identified a large number of peptides corresponding to IDM1 and MBD7 ( Table 1 ) . Our data suggest that IDM1 , IDM2 , IDL1 , and MBD7 form a tight complex . The interaction of IDL1 and MBD7 with IDM1 and IDM2 was corroborated using the yeast two-hybrid assay ( Fig 1A ) . Interestingly , IDL1 can interact with the C-terminal of IDM1 while IDM2 can interact with the N-terminal of IDM1 ( Fig 1A ) . MBD7 also interacts with IDM2 and IDL1 , but not with IDM1 ( Fig 1A ) . The association between IDM2 , IDL1 , and MBD7 was also confirmed by a pull down assay ( Fig 1B ) . To confirm these protein-protein interactions in vivo , we performed two assays . First , results from the split-LUC assay demonstrated that IDL1 interacts with IDM2 , and that MBD7 interacts with both IDM2 and IDL1 ( Fig 1C ) . Second , experiments using GFP-tagged IDM2 and Flag-tagged MBD7 transiently expressed in Nicotiana benthamiana leaves revealed that MBD7 co-IP’d with IDM2 from IDM2-GFP-expressing leaves , but not from leaves of control plants ( Fig 1D ) . MBD7 also co-IP’d with IDL1 ( Fig 1D ) . The interaction between MBD7 and IDM1 was also detected by co-IP in plants harboring both MBD7-Myc and IDM1-HA-YFP transgenes driven by their native promoters ( Fig 1D ) . Taken together , our results suggest that IDM1 , IDM2 , IDL1 and MBD7 form a tight multiprotein complex in vivo . IDL1 encodes an α-crystallin domain protein that belongs to the IDM2 protein family . Besides IDM2 , the IDM2 protein family includes three IDM2-like proteins including IDL1 ( At1g20870 ) , IDL2 ( At1g54850 ) , and IDL3 ( At1g76640 ) ( S2 Fig ) . According to sequence alignment and phylogenetic analysis , IDL1 belongs to the same clade as IDM2 and shares maximal sequence similarity with IDM2 [21] . MBD7 encodes a methyl-CpG-binding domain protein , which binds methylated CpG dinucleotides in vitro and localizes to highly CpG-methylated chromocenters in vivo [17 , 18] . The tight association of IDL1 and MBD7 with IDM1 and IDM2 suggests that IDL1 and MBD7 may also be required for active DNA demethylation . In order to examine the possible role of MBD7 in active DNA demethylation , two T-DNA insertion lines were isolated and RT-PCR was done to confirm that they have a complete loss of mRNA expression ( S3 Fig ) . Whole genome bisulfite sequencing was done using DNA purified from 12-day-old mbd7-1 and wild type seedlings . The overall genome methylation level was slightly increased in mbd7-1mutant plants compared with wild type control ( S4 and S5A Figs ) . Hundreds of differentially methylated regions ( DMRs ) were apparent in mbd7-1 knockout plants , including 748 loci with DNA hypermethylation and 198 loci with DNA hypomethylation , respectively ( S2 and S3 Tables ) . Six hypermethylated loci in mbd7 , idm1 , and ros1 plants were selected for validation by PCR-based DNA methylation analysis , and all of them confirmed to be hypermethylated in mbd7 mutants ( Fig 2A and 2B and S5B and S5C Fig ) , as in idm1-1 or ros1-4 . However , the chop PCR marker locus DT-77 , which was hypermethylated in idm1 and ros1 [10 , 11] , was not hypermethylated in mbd7 mutant plants ( Fig 2B ) , suggesting not all IDM1 target loci were affected by MBD7 loss-of-function . Same as the results in idm1-1 and ros1-4 mutants , the hypermethylated loci in mbd7-1 mutants spread evenly across the five chromosomes and there was no enrichment at the chromocenters ( S5D Fig ) , although MBD7 are mainly localized at highly methylated chromocenters in nuclei [18] . To determine whether the MBD7 mutation affects DNA demethylation in gene body regions or in TEs , DNA regions were classified into gene body regions , intergenic regions , TEs out of gene regions , and TEs overlapping with gene regions . The ratios of DMRs in each class were calculated . Interestingly , we found the ratios of hypermethylated loci in each class of mbd7-1 were similar to those of ros1-4 ( S5E Fig ) . Both of these two mutants have a small number of hyper DMRs distributed in gene body regions . We also examined the methylation levels of ros1-4 and idm1-1 in those regions that are hypermethylated in mbd7-1 in CG , CHG , and CHH contexts and found most of them are hypermethylated to different degrees ( S5F Fig ) . Approximately 18% and 48% of the 748 hyper-DMRs in mbd7-1 were also hypermethylated in the idm1-1 and ros1-4 mutants , respectively ( Fig 2C ) . We found the hyper-DMRs identified from ros1-4 or idm1-1 that overlap with mbd7-1 mutant showed higher density of CG methylation compared with the non-overlapped loci ( S5G Fig ) . This result suggests MBD7 dysfunction preferentially affects the genomic regions with high CG methylation density . The overlapped loci between mbd7-1 and ros1-4 were more frequently located in TE regions while the overlapped loci between mbd7-1 and idm1-1 were preferentially located in gene body regions ( S5H and S5I Fig ) . The overlapped loci in these mutants are all hypermethylated in all sequence contexts ( Fig 2C ) . The DNA methylation level at the mbd7-specific loci also appeared to be slightly increased in idm1-1 and ros1-4 ( Fig 2C ) . However , the increases in ros1-4 and idm1-1 were not significant to make the loci counted as hyper-DMRs according to the parameters defined in this study , suggesting possible underestimation of the ratio of overlapping ( Fig 2C ) . Taken together , our whole genome bisulfite sequencing results indicate that MBD7 , IDM1 and ROS1 may function in the same pathway in active DNA demethylation at some loci . Previous studies indicated that IDM1 and IDM2 were required for preventing certain transgenes and endogenous genes from transcriptional silencing [10 , 11 , 20 , 22] . To determine whether MBD7 functions in anti-silencing of transgenes , we introduced the RD29A-LUC and 35S-NPTII into mbd7-1 mutant plants by crossing the mbd7-1 mutant with wild type C24 plants expressing the reporter gene . As shown in Fig 3A , the mbd7-1 mutant exhibited a kanamycin sensitive growth phenotype but normal LUC expression . Real time PCR analysis also showed reduced expression of the NPTII transgene , but the expression level of RD29A-LUC in mbd7-1 was comparable to that of the wild type control ( Fig 3B ) . Thus , the mbd7 mutation preferentially causes the silencing of 35S-NPTII but not RD29A-LUC , consistent with the idm1 or idm2 mutation [20 , 22] . Changes in the DNA methylation level in or near a gene may modulate the expression of that gene . To determine whether DNA hypermethylation affects the expression of adjacent genes in mbd7 mutant plants , we examined the expression levels of 14 genes that have detectable levels of transcript by real time PCR and show increased DNA methylation near or within the gene in mbd7-1 , idm1-1 and ros1-4 plants ( Fig 3C–3E and S6 Fig ) . Eleven of the tested genes showed a significant reduction in their transcript levels in these mutants ( Fig 3C–3E and S6 Fig ) . Our results suggest that , like ROS1 and IDM1 , MBD7 is critical for preventing the transcriptional silencing of some loci . MBD7 can bind to methylated DNA through the methyl-CpG-binding domain in vitro [17] , suggesting that MBD7 may recognize hypermethylated DNA regions and recruit IDM1 complex to the target loci . Previous data showed that IDM1 is critical for H3K18 and H3K23 acetylation in vivo [10] . Chromatin immunoprecipitation ( ChIP ) assays revealed that the acetylated histone H3K18ac and H3K23ac markers were reduced in idm1-1 and mbd7 mutant plants at hyper-DMRs , although the reduction in mbd7-1 was not as dramatic as in idm1-1 ( Fig 4A ) . ChIP assays also indicated that MBD7 was enriched at all of the hyper-DMRs tested , but not the control regions that contain pretty high CpG DNA methylation in the gene body regions ( like At1g01260 and At1g10950 ) but did not show DNA hypermethylation in mbd7-1 mutant compared with wild type ( Fig 4B ) .
ROS1/DME family of 5mC DNA glycosylases initiated DNA demethylation is the major active DNA demethylation pathway in plants [23] . Unlike the well-established DNA methylation pathway , in which enzymes responsible for methylating DNA are guided to specific loci by base-pairing between small RNAs and the scaffold transcripts [1 , 24] , how the DNA demethylation enzymes are recruited to specific genomic loci is poorly understood . Our findings demonstrate that , in addition to IDM1 and IDM2 , MBD7 and IDL1 are required for ROS1-mediated active DNA demethylation and suppression of gene silencing at some loci in Arabidopsis . DNA methylation is a conserved epigenetic marker that is important for TE and gene silencing , whereas DNA demethylation positively regulates gene expression . Some stress responsive genes are silenced by DNA methylation under normal conditions , but are induced by ROS1/DME mediated DNA demethylation under stressed conditions . The dynamic regulation of these genes is important for stress responses in Arabidopsis [25] . For example , the ros1 and rdd ( ros1 dml2 dml3 ) triple mutant shows increased susceptibility to the fungal pathogen Fusarium oxysporum [25] . In light of the important roles of MBD7 , IDM1 and IDM2 in DNA demethylation , we speculate that this complex is important for activating stress responsive genes via recruiting ROS1 for active DNA demethylation . IDM2 encodes an α-crystallin domain ( ACD ) protein that interact with IDM1 and is required for the full activity of IDM1 in vivo [11] . IDL1 is also an ACD protein and shares the maximal sequence identity with IDM2 . Due to the lack of mutants for IDL1 in the public database , we were unable to investigate the effects of IDL1 mutation on DNA methylation level . IDL1 directly interacts with IDM1 , IDM2 , and MBD7 . Our co-IP and MS/MS results suggested that equal amounts of IDM1 , IDM2 , IDL1 , and MBD7 polypeptides are present in the IDM1 complex ( Table 1 ) , indicating that IDL1 may be as important as MBD7 , IDM1 and IDM2 for active DNA demethylation and gene activation . Our results also revealed that dysfunction of MBD7 causes DNA hypermethylation at hundreds of genomic loci and silencing of a reporter gene and lots of endogenous genes . Because MBD7 binds to the methylated DNA , the role of MBD7 in DNA demethylation may be facilitating the recruitment of IDM1 to the specific loci . Since the ratio of overlapped hyper-DMRs between mbd7 and idm1 was low , MBD7 may also recruit histone modification enzymes other than IDM1 to create chromatin environment favorable for active DNA demethylation . In addition to MBD7 , MBD5 and MBD6 also bind to methylated DNA in vitro [18] . Interestingly , MBD6 was also identified during MS/MS analyses of IDM1- and IDL1-associated proteins ( S7A Fig and S1 Table ) and our yeast two-hybrid and split-LUC results confirmed that MBD6 can interact with IDM2 and IDL1 ( S7B and S7C Fig ) . Chop-PCR results shown that the DNA methylation level of two loci were increased in mbd6 and idm1-1 mutants but not in mbd7-1 mutant plants ( S7D Fig ) , suggesting that MBD6 may also involved in recruitment IDM1 to specific DNA regions , promoting active DNA demethylation . This is consistent with the fact only a subset of genomic regions demethylated by ROS1 and other related 5mC DNA glycosylases were influenced in the mbd7 knockout plants . When this manuscript was under review , Lang et al . and Wang et al . reported that MBD7 and IDM3 were required for active DNA demethylation at some loci in Arabidopsis [26 , 27] . IDM3 is synonymous with IDL1 . The identification of the same two proteins by three independent studies further underlines the importance of MBD7 and IDL1/IDM3 in the active DNA demethylation pathway . More importantly , in our study , we identified MBD7 and IDL1 from affinity purification and mass-spectrometric analysis . We characterized the IDM1-IDM2-IDL1-MBD7 complex and found the ratio of these four proteins should be 1:1:1:1 in the complex . Moreover , in addition toMBD7 , MBD6 was found to be required for the recruitment of IDM1 to some specific loci in our study . In summary , we propose a model in which MBD7 or MBD6 binds to the DNA hypermethylated region and recruits the histone acetyltransferase complex ( IDM1 , IDM2 and IDL1 ) to specific loci . Histone modification ( i . e . , H3K18 and H3K23 acetylation ) creates the favorable chromatin environment for targeting of ROS1 ( and other members of the ROS1 subfamily of 5-methycytosine DNA glycosylases ) to the loci for active DNA demethylation ( Fig 5 ) . In accordance with this model , MBD7 binds the target loci and H3K18ac and H3K23ac markers are reduced in the mbd7 mutant plants , as in idm1 mutants , at the target regions ( Fig 4A ) . Not all the hypermethylated CG sites exhibited MBD7 accumulation ( Fig 4B ) , suggesting that in addition to high CpG DNA methylation , other features such as heterochromatic histone marks may also contribute to MBD7 binding . MBD7 interacting proteins may also be involved in IDM1 targeting . As previous reported , IDM1 also contain a MBD domain at the N-terminal that can bind CG methylated nucleotide in vitro [10] . It seems that single MBD domain in IDM1 is not sufficient to recruit IDM1 to the methylated DNA and three additional MBD domains in MBD7 are required for the recognition of some of the IDM1 target loci . In addition , only some of the ROS1 target loci are affected in idm1 , idm2 and mbd7 , indicating the existence of IDM1-independent mechanisms for recruiting the ROS1/DME glycosylases .
Two T-DNA insertion mutants of MBD7 , mbd7-1 and mbd7-2 , were obtained from Arabidopsis Biological Resource Center . The 35S::6Myc-IDM2 transgenic plants were as described [11] . Arabidopsis seedlings were cultivated on Murashige-Skoog ( MS ) nutrient agar plates at 23°C with 16 h of light and 8 h of darkness for 12 days before DNA or RNA analysis . The IDM1promoter::IDM1-3HA-YFP construct was as described [10] . For the 35S::6Myc-IDL1 construct , full length of IDL1 genomic DNA was amplified from wild type genomic DNA by PCR and sub-cloned into pCAMBIA1307 ( 6Myc tag ) . For complementation of mbd7 mutants , a 3 . 7-kb genomic DNA fragment containing the MBD7 gene was amplified from Col-0 genomic DNA by PCR and cloned into the pCAMBIA1305 vector for plant transformation . Agrobacterium tumefacines strain GV3101 carrying different constructs was used to transform mutant or wild type plants via the standard floral dipping method . Primary transformants were selected on MS plates containing hygromycin . Approximately 5 g of flower tissue collected from IDM1-HA , IDM2-GFP and 6MYC-IDL1 transgenic plants was ground in liquid N2 with a mortar and pestle . The fine powder was suspended in 25 ml of lysis buffer ( 50 mM Tris , pH8 . 0 , 230 mM NaCl , 5 mM MgCl2 , 10% glycerol , 0 . 2% NP-40 , 0 . 5 mM DTT , 1 mM PMSF , and 1 protease inhibitor cocktail tablet ( Roche , 14696200 ) ) and centrifuged for 15 min at 18 , 300 g at 4°C . The supernatant was incubated with 50 μl of anti-HA agarose beads ( Sigma , A4865 ) , GFP-Trap A beads ( ChromoTek ) , or anti-c-Myc agarose beads ( Sigma , A7470 ) for 5 h at 4°C . The beads were washed twice with 25 ml of lysis buffer and four times with 1 ml of lysis buffer . Proteins bound to HA beads were eluted with HA peptide ( Genescript ) , whereas proteins bound to Myc beads were eluted with 0 . 1 M ammonium hydroxide ( pH 11 . 5 ) . The GFP-Trap A beads were directly boiled in SDS-sample buffer . The mass spectrometric identification of purified proteins was done using the method described by Li et al . [28] . Briefly , purified proteins were separated by SDS-PAGE . Then the entire gel lane was cut into 6 equal portions and dehydrated . Proteins were digested in-gel with endoproteinase trypsin ( 0 . 5 ng/μL trypsin in 50 mM ammonium bicarbonate , pH 8 . 5 ) . Extracted peptides were sequenced by LC-MS/MS on the Velos Pro Orbitrap Elite mass spectrometer ( Thermo Scientific , USA ) equipped with a nano-ESI source . The mass spectrometer was run in data-dependent mode with one MS scan in FT mode at a resolution of 120000 . Ten CID ( Collision Induced Dissociation ) MS/MS scans were applied in the ion trap for each cycle [28] . Mascot server ( Matrix Science Ltd . , London , UK ) and IPI ( International Protein Index ) Arabidopsis protein database were used as a searching platform . The coding sequences of MBD7 and IDL1 were amplified by PCR . After validation of the sequences , the genes were subcloned into pGBK-T7 or pGAD-T7 ( Clontech ) to generate DNA binding or activation domain fusion constructs , respectively . For protein interaction analysis , two combinatory constructs were transformed simultaneously into the yeast strain AH109 ( Clontech ) and tested for Leu- , Trp- , and His- auxotrophy according to the manufacturer’s protocols . The coding sequences of MBD7 and IDL1 were amplified by PCR and subcloned into pCAMBIA1300-NLUC or pCAMBIA1300-CLUC to generate N-terminal or C-terminal luciferase-fusion constructs , respectively . For protein interaction analysis , two combinatory constructs were transformed simultaneously into Nicotiana benthamiana leaves . To prevent the silencing of those genes , a construct encoding virus p19 protein was transformed at the same time [29 , 30] . Pull down assays among IDM2 , IDL2 and MBD7 were performed as described [31] . Briefly , Maltose binding protein ( MBP ) either alone or fused to IDM2 ( MBP-IDM2 ) was expressed in E . coli and fixed to an amylose resin . His-tagged IDL1 purified from E . coli ( His-IDL1 ) was incubated with either MBP-IDM2 or MBP bound to the resin . After washes , the proteins associated to the resin were separated by SDS-PAGE , transferred to a membrane , and immunoblotted with antibodies against His6- tag . To determin the interaction betweem MBD7 and IDM2 or IDL1 , MBP alone or MBP-IDM2 or MBP-IDL1 was expressed in E . coli and fixed to an amylose resin . His-tagged MBD7 purified from E . coli ( His-MBD7 ) was incubated with either MBP-IDM2 or MBP-IDL1 or MBP bound to the resin . After washes , the proteins associated to the resin were separated by SDS-PAGE , transferred to a membrane , and immunoblotted with antibodies against His6- tag . Over-expression binary constructs were transformed into Agrobaterium tumefaciens GV3101 and cultured overnight . After resuspending in 10 mM MgCl2 , 150 μM acetosyringone and 10 mM MES ( pH 5 . 7 ) at an OD600 of 0 . 6 , and incubating for 3 h at room temperature in dark , the microbodies were syringed into leaves of N . Benthamiana . The infiltrated plants were grown in the greenhouse for 2–3 days . For co-IP between MBD7 and IDM1 , 1 g of flower tissue from MBD7-MYC and IDM1-HA-YFP and MBD7-MYC F1 hybrids were collected and ground in liquid N2 . Lysis buffer ( 4 ml; 50 mM Tris pH8 . 0 , 230 mM NaCl , 5 mM MgCl2 , 10% glycerol , 0 . 2% NP-40 , 0 . 5 mM DTT , 1 mM PMSF , 100 μM MG132 ( Gene operation ) and 1 protease inhibitor cocktail tablet ( Roche , 14696200 ) was added and mixed carefully . After centrifugation for 15 min at 18 , 200 g and 4°C , the supernatants were incubated with 8 μL of anti-HA agarose ( Sigma , A4865 ) overnight at 4°C while rotating . After washing the beads 5 times with 1 ml lysis buffer for 5 min each , the beads were resuspended in 60 μL SDS sample buffer and boiled for 5 min . Input and eluted proteins ( 20 μL and 15 μL , respectively ) were resolved by 10% SDS-PAGE and electroblotted onto PVDF membranes for Western blotting . The primary antibodies against HA and MYC were diluted at 1:2000 and the second antibody goat anti-mouse IgG ( Cwbio ) was diluted at 1:5000 . For the co-IP between MBD7 and IDM2 , MBD7 and IDL1 , 0 . 3 g of co-transformed N . Benthamiana leaves were ground in liquid N2 and incubated in l ml lysis buffer without MG132 . After centrifugation , 5 μL GFP-Trap A ( ChromoTek ) or anti-c-Myc agarose ( Sigma , A 7470 ) were added and incubated for 4 h . The beads were washed and then boiled in 50 μL SDS sample buffer . About 20 μL of input and 10 μl of eluate were analyzed by SDS-PAGE and Western blotting . The primary antibody against GFP ( EASYBIO ) , Flag ( Sigma , F1804 ) and MYC ( Cwbio ) were diluted at 1:2000 and the second antibody goat anti-mouse IgG or goat anti-rabbit IgG ( Cwbio ) were diluted at 1:5000 . DNA was extracted from 2 g of 12-d-old seedlings using the DNeasy Plant Mini Kit ( Qiagen ) and sent to the Center for High throughput sequencing ( Biodynamic Optical Imaging Center , Peking University ) for bisulfite treatment , library preparation , and sequencing . For data analysis , adapter , and low-quality sequences ( q < 20 ) were trimmed and clean reads were mapped to Arabidopsis thaliana TAIR 10 ( 10th release of the Arabidopsis genome sequence from the Arabidopsis Information Resource ) genome using Bisulfite Sequence Mapping Program allowing two mismatches . Identification of differentially methylated regions ( DMRs ) was conducted according to [32] . Heat maps were drawn according to [22] . Genomic DNA ( 100 ng ) was cleaved with the methylation-sensitive restriction enzymes HhaI or the methylation-dependent restriction enzyme McrBC , and the products subjected to PCR using the loci-specific primers ( S4 Table ) . Total RNA was extracted from 12-d-old seedlings in plates using the RNeasy plant mini kit ( Qiagen ) , and contaminating DNA removed with RNase-free DNase ( Qiagen ) . About 2 μg mRNA was used for first-strand cDNA synthesis with the Super scriptTM III First-Strand Synthesis System ( Invitrogen ) for RT-PCR following the manufacturer’s instructions . The cDNA reaction mixture was then diluted five times , and 1 μl used as template in a 20 μl PCR reaction with iQ SYBR Green Supermix ( Bio-Rad ) . Chromatin immunoprecipitation ( ChIP ) assays were performed according to a published protocol [33] . The following antibodies were used for ChIP assays: anit-H3K18ac ( ab1191 , Abcam ) , anti-H3K23ac ( 07–355 , Millipore ) , and anti-MYC ( M4439 , Sigma ) . ChIP products were eluted into 50 μl of TE buffer , and a 2 μl aliquot was used for each qPCR reaction . We used whole genome bisulfite sequencing data to analyze the genome-wide methylation status in WT and mbd7-1 mutant plants . The dataset was deposited at NCBI ( Col-0:SRX747290 , mbd7-1:SRX833671 ) . The ros1-4 and idm1-1 whole genome bisulfite sequencing data were from GEO accession GSE33071 [10] .
|
DNA cytosine methylation ( 5-methylcytosine , 5-meC ) plays important roles in genome organization , genomic imprinting , transposon silencing and gene expression . DNA methylation patterns are dynamically controlled by methylation and demethylation reactions during development and reproduction in eukaryotes . In Arabidopsis thaliana , ROS1 , a 5-methylcytosine DNA glycosylase , is responsible for active DNA demethylation via a base excision repair pathway . Our previous work showed that a histone acetyltransferase , IDM1 , is an important regulator of active DNA demethylation . Here we identified two components , IDL1 and MBD7 , as additional novel regulators of active DNA demethylation . Our data suggest that a histone acetyltransferase complex , composed of MBD7 , IDL1 , IDM2 and IDM1 , is required for the recruitment of ROS1 to specific loci for active DNA demethylation . Our findings significantly expand our understanding of the regulatory mechanisms underlying active DNA demethylation in plants .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Regulation of Active DNA Demethylation by a Methyl-CpG-Binding Domain Protein in Arabidopsis thaliana
|
Animal control measures in Latin America have decreased the incidence of urban human rabies transmitted by dogs and cats; currently most cases of human rabies are transmitted by bats . In 2004–2005 , rabies outbreaks in populations living in rural Brazil prompted widespread vaccination of exposed and at-risk populations . More than 3 , 500 inhabitants of Augusto Correa ( Pará State ) received either post-exposure ( PEP ) or pre-exposure ( PrEP ) prophylaxis . This study evaluated the persistence of rabies virus-neutralizing antibodies ( RVNA ) annually for 4 years post-vaccination . The aim was to evaluate the impact of rabies PrEP and PEP in a population at risk living in a rural setting to help improve management of vampire bat exposure and provide additional data on the need for booster vaccination against rabies . This prospective study was conducted in 2007 through 2009 in a population previously vaccinated in 2005; study participants were followed-up annually . An RVNA titer >0 . 5 International Units ( IU ) /mL was chosen as the threshold of seroconversion . Participants with titers ≤0 . 5 IU/mL or Equivalent Units ( EU ) /mL at enrollment or at subsequent annual visits received booster doses of purified Vero cell rabies vaccine ( PVRV ) . Adherence of the participants from this Amazonian community to the study protocol was excellent , with 428 of the 509 ( 84% ) who attended the first interview in 2007 returning for the final visit in 2009 . The long-term RVNA persistence was good , with 85–88 . 0% of the non-boosted participants evaluated at each yearly follow-up visit remaining seroconverted . Similar RVNA persistence profiles were observed in participants originally given PEP or PrEP in 2005 , and the GMT of the study population remained >1 IU/mL 4 years after vaccination . At the end of the study , 51 subjects ( 11 . 9% of the interviewed population ) had received at least one dose of booster since their vaccination in 2005 . This study and the events preceding it underscore the need for the health authorities in rabies enzootic countries to decide on the best strategies and timing for the introduction of routine rabies PrEP vaccination in affected areas .
Rabies is a viral zoonosis that affects mammals . It is caused by neurotropic viruses belonging to the family Rhabdoviridae , genus Lyssavirus . The International Committee on Taxonomy of Viruses ( ICTV ) recognizes today 14 species [1 , 2]; this taxonomy is rapidly evolving and the two more recently accepted Lyssaviruses isolated from a bat in Germany ( Bokeloh bat lyssavirus ) and from a civet in Africa ( Ikoma lyssavirus ) have been included as new species [3 , 4 , 5] . Most lyssavirus variants are found in bats and are known to cause rabies in humans and in domestic animals [6] . Interestingly , the isolates detected until now on the American continent all belong to the classical rabies virus ( RABV ) , the species used in rabies vaccine . Lyssaviruses are neurotropic , causing acute encephalitis or “furious rabies” in about 70% of cases and a paralytic form of rabies in 30% . Not all exposures lead to illness , but once symptoms occur , rabies is almost always fatal . Therefore , proper prophylaxis to prevent infection must be administered promptly after exposure . Approximately 26 , 400 [95% confidence interval ( CI ) 15 , 200–45 , 200] human rabies deaths are estimated to occur worldwide each year using the “Cause of Death Ensemble” model , but the estimate rises to 61 , 000 ( 95% CI 37 000–86 000 ) when a probability decision-tree approach is used [7] . Rabies reservoirs and vectors include domestic as well as wild mammals , but human infection mostly results from bites from rabies-infected dogs . Animal control measures have decreased the incidence of urban human rabies transmitted by dogs and cats; and currently , in Latin American and Caribbean countries , most cases of human rabies are transmitted by bats [8 , 9] . The Pan American Health Organization ( PAHO ) implemented a multinational program against rabies in 1983 , supporting intensive dog vaccination programs . The results have been very effective . In the Americas , canine cases decreased by 93% ( from 15 , 686 to 1 , 131 ) and human cases decreased by 91% ( from 355 to 35 ) between 1990 and 2003 [10] . In the countries where the circulation of canine rabies has been controlled , incidences of canine and human rabies continue to decrease in parallel , with 400 cases reported in dogs in 2010 and 10 in humans in 2012 in Latin America [11] . However , the number of human rabies cases caused by bats began to increase in Latin America in 2004 , when more than half of the 87 reported cases were transmitted by vampire bats . Most cases were caused by outbreaks in Brazil ( 21 cases ) , Colombia ( 14 cases ) and Peru ( 8 cases ) . In 2005 , of the 60 reported cases of bat-transmitted human cases in Latin America , 42 were in Brazil and 7 in Peru ( Amazonian area ) [8] . Although human rabies cases have declined since 2006 , cattle rabies in the region continues to increase , and a recent report from Peru estimated that the rabies seroprevalence in bats varied from 3 to 28% depending on the geographical region [12 , 13] . Although many human rabies outbreaks have been reported in northern Amazonian Brazil , few epidemiological studies have been performed . In 2004 , a total of 21 people died during rabies outbreaks in two villages , Portel and Viseu , in the region of Pará State , Brazil , following bat bites ( or as a result of bat rabies ) . In May 2005 , 15 cases occurred in Augusto Correa , another rural municipality in the same region . These outbreaks , affecting populations living in remote areas , were of great concern to health authorities , prompting widespread vaccination of the exposed or at risk populations [14] . Following the rabies outbreak in 2005 , more than 3 , 500 inhabitants of Augusto Correa received either post-exposure ( PEP ) or a pre-exposure ( PrEP ) prophylaxis . A few people were given booster vaccinations after possible rabies re-exposure , mostly following dog , bat , and monkey bites . As per national guidelines for PrEP in Brazil , if antibody titers < 0 . 5 IU/mL , the recommendation is to administer 1 booster dose via the IM route and to perform serological testing at D14 . For re-exposed individuals who have previously received PEP no serological testing is done . Within 90 days of completing PEP , no vaccine is administered while within 90 days of incomplete PEP , the missing doses have to be given . More than 90 days after completing full PEP , 2 doses of vaccine ( D0 , D3 ) are recommended while if the PEP is incomplete , the recommendation is to administer the full 5-dose schedule based on the nature of rabies exposure [15] .
This was a single-site , prospective epidemiological study designed to evaluate the persistence of RVNA in a population at risk of vampire bat rabies and who had previously received either PrEP or PEP regimens . The study also aimed at providing additional data on the need for booster vaccination against rabies . The results of 3 years of follow-up are presented here . Outbreaks of human rabies cases occurred in 2004 and in 2005 in Augusto Correa , a rural municipality of approximately 27 , 000 inhabitants in Para state , northern Brazil . After the second outbreak , approximately 3 , 500 local residents of Augusto Correa were given either the standard five-dose intramuscular ( IM ) PEP ( with or without rabies immunoglobulin administration ) on Days 0 , 3 , 7 , 14 and 28 , or a three-dose PrEP vaccination series on Days 0 , 7 and 21 or 28 with purified Vero cell rabies vaccine ( PVRV , Verorab; Sanofi Pasteur , France ) . This prospective study was conducted in 2007 through 2009 at the Arai health unit ( USF Arai 3 ) in Augusto Correa , to evaluate the persistence of RVNAs in those who had been vaccinated in 2005 . Each study participant was followed-up annually for 3 years ( in 2007 , 2008 and 2009 ) . As recommended by WHO , an RVNA titer >0 . 5 International Units ( IU ) /mL was chosen as the threshold of seroconversion [16] . Participants with RVNA titers ≤0 . 5 IU/mL or Equivalent Units ( EU ) /mL at enrollment or at subsequent annual visits received booster doses of PVRV . Anyone who had been vaccinated in 2005 was eligible to participate . Written informed consent was given by participants aged 18 years and above or by parents or legal guardians if younger . The study was conducted in accordance with the Edinburgh revision of the Declaration of Helsinki , International Conference on Harmonization ( ICH ) good clinical practice and applicable national and local requirements regarding ethical committee review . The primary objective was to evaluate the persistence of RVNA following PrEP or PEP . Secondary objectives included describing RVNA titers following receipt of PVRV booster doses , estimating the incidence of clinical cases of rabies in the study population , and determining the correlation between the anti-rabies antibody titers obtained by the rapid fluorescent focus inhibition test ( RFFIT ) and a commercially available enzyme-linked immunosorbent assay ( ELISA ) . Blood specimens ( 5 mL ) were collected from each study participant at enrollment and at each of the three annual follow-up visits ( when the patient came to the health center ) for testing by RFFIT and enzyme-linked immunosorbent assay ( ELISA ) . Blood serum specimens were divided into four 0 . 5 mL aliquots for testing . The RFFIT method was adapted from the original one [17] while changing both the cell line support ( BHK21 cells instead of MNA cells ) and the rabies virus strain used . RVNA titers of all specimens against the Pasteur virus strain PV ( instead of the Challenge Virus Strain CVS ) were measured by RFFIT at the Centro de Controle de Zoonoses ( CCZ ) laboratory in São Paulo , Brazil . Ten percent of those specimens were randomly selected for RFFIT re-testing at Institut Pasteur laboratory in Paris , France , using a vampire bat virus strain ( instead of the CVS strain ) . In addition , the concentration of rabies virus anti-glycoprotein antibodies ( EU/mL ) in each blood sample was determined by ELISA ( Pasteur virus strain ) at Institut Pasteur laboratory in Paris , France , using the Bio-Rad Platelia assay as per the manufacturer’s instructions . The correlation between the RVNA titers measured by RFFIT and by fluorescent antibody virus neutralization ( FAVN ) assay ( CVS in BHK21 cells ) [18] was estimated at the CCZ laboratory , São Paulo , Brazil , using the specimens collected in 2007 , the first year of the study . The immunogenicity analysis was descriptive; no hypotheses were tested . Seroconversion ( RFFIT titer >0 . 5 IU/mL ) rates and geometric mean antibody titers ( GMTs ) were calculated with their 95% confidence intervals ( CIs ) . The sample size calculation was based on an expected seroconversion rate of 90% at 5 years after the primary vaccination series . A total of 140 subjects were required to ensure a 95% precision for a two-sided CI of 5% . Assuming 30% of the participants would be lost to follow-up at 5 years after primary vaccination , a total of 200 subjects had to be included . However , to anticipate additional dropouts , subgroup analyses and insufficient sera for laboratory testing , the planned enrollment was 500 participants . The study populations included in the analysis comprised: 1 ) all the evaluable study participants in each follow-up year , 2 ) participants who received a booster dose of vaccine at enrollment or during a follow up year , and 3 ) participants who did not receive booster doses of vaccine either at study entry or in any follow-up year . Missing data were not replaced . The primary study endpoint was the number and percentage of subjects with RVNA titers >0 . 5 IU/mL each year using the RFFIT assay . We performed an analysis by gender and age group ( i . e . , 2–5 , 6–17 , 18–40 , 41–60 , and >60 years of age ) . The number and percentage of subjects with RVNA titers >0 . 5 EU/mL using the ELISA test was calculated in the overall study population for each of their follow up visits . For inter-group comparisons , quantitative variables and ordinal qualitative variables were compared using Student’s t-test or ANOVA ( parametric data ) and the Wilcoxon or Kruskal–Wallis test ( nonparametric data ) . Qualitative variables were compared using the Chi square test ( or Fisher exact test when frequencies were less than five for at least one category ) . The correlations of GMTs measured by two different assays were determined by Pearson’s correlation coefficient ( r ) . The correlations between percentages of participants with titers >0 . 5 IU/mL or EU/mL measured by RFFIT , FAVN or ELISA were calculated using the Kappa coefficient ( κ ) .
A total of 509 participants were enrolled in 2007 ( Fig 1 ) . Two of the 509 participants were excluded because their vaccination dates could not be confirmed . Among the 507 participants included , 496 ( 97 . 8% ) were immunized ( either PEP or PrEP ) in 2005 , eight in 2004 and three in 2006 . In 2008 , 42 participants were discontinued and 465 ( 91 . 7% ) returned for evaluation . In 2009 , 53 of the 465 remaining participants were discontinued and 16 of the 42 subjects who had been discontinued in 2008 came back so that a total of 428 ( 84 . 4% ) participants were present in 2009 . The mean ( ±SD ) follow-up duration was 22 . 6 ± 6 . 9 months ( range 0 . 0–26 . 5 months ) . Among the 95 participants who did not complete the study or did not attend all of the visits , 91 ( 95 . 8% ) were lost to follow-up and four died ( one following an epileptic coma and three of different cancers ) . In 2007 , nine participants were excluded from analysis because they had not received a complete PEP schedule ( i . e . , <5 vaccine doses ) . Four additional participants were excluded from analysis in 2008 because of missing data ( no booster dose information ) . One subject who had previously been excluded from analysis in 2008 withdrew from the study in 2009 . The 507 participants who were evaluated at the start of the study ranged from 2 to 83 years of age , with a mean ± SD of 21 . 4 ± 16 . 8 years , and 288 ( 56 . 8% ) were male . The age and gender distributions are shown in Table 1 . The mean time ± SD between the last vaccine dose and enrollment was 23 . 7 ± 1 . 7 months . At enrollment , PEP had been given to 448 of the 507 participants ( 88 . 4% ) ; 58 ( 11 . 4% ) had received PrEP , and 1 ( 0 . 2% ) had received a re-exposure PEP vaccination . In 2005 , 340 subjects ( 78 . 0% ) received rabies immunoglobulin . The number of vaccine doses administered in 2005 and participant age at inclusion are given in Table 1 . To be eligible for the immunogenicity analysis , participants had to receive three vaccine doses for PrEP , five doses for PEP , or two booster doses for PEP following a suspected re-exposure . Most participants ( 439 , 86 . 6% ) had received five doses , six subjects ( 1 . 2% ) had received four doses , 59 subjects ( 11 . 6% ) received three doses , and only three subjects ( 0 . 6% ) had received two doses . The number of doses does not exactly match the number and type of prophylaxis regimens given in 2005 because nine subjects who reported being given PEP had received fewer than five injections . Two of them had received only two vaccine doses , one received three doses and six received four doses . Those subjects were excluded from the analysis of both the boosted and non-boosted populations . Participants with antibody levels <0 . 5 IU/mL or EU/mL at inclusion or at one of the annual study visits were considered no longer seroconverted against rabies and were boosted . At enrollment in 2007 , 2 years after vaccination , nine of the 507 participants ( 1 . 8% ) had been boosted since receiving their PrEP or PEP regimens; six were given one or two booster injections , but the number of doses was not known for the three others . In 2008 , 3 years after vaccination , 43 of the 461 participants with booster dose information ( 9 . 3% ) had been boosted in the previous year . Forty of the 43 received one or two booster dose injections , one received five doses , and the number of doses was not known for two participants . In 2009 , 4 years after vaccination , 14 of 428 remaining participants ( 3 . 3% ) had received booster doses since their 2008 follow-up visit . Thirteen received one or two booster dose injections and one received three doses ( Table 2 ) . Twenty-six participants ( 5 . 1% ) reported being bitten by an animal between vaccination in 2005 and enrollment in 2007; 21 of them ( 80 . 8% ) had an RFFIT titer >0 . 5 IU/mL and 5 ( 19 . 2% ) had a titer ≤0 . 5 IU/mL . Additionally , 9 ( 34 . 6% ) had received a booster after the vaccination in 2005 and 17 ( 65 . 4% ) had not . In the following year , 2007–2008 , 34 participants ( 7 . 3% ) were bitten; 32 ( 94 . 1% ) had an RFFIT titer >0 . 5 IU/mL and 12 ( 35 . 3% ) had received a booster after enrollment . Between their 2008 and 2009 study visits , 29 participants ( 6 . 8% ) were bitten; 24 ( 82 . 8% ) had an RFFIT titer >0 . 5 IU/mL , and 7 ( 24 . 1% ) had received booster dose after enrollment . There was a total of 89 cases of re-exposure to rabies resulting from bites from rabid animals , mostly dogs ( 52 cases ) but also bats , cats , and monkeys . No cases of rabies occurred among the study participants . The serology results for both the non-boosted and boosted populations are shown in Table 3 . In 2007 , 2 years after vaccination , 413 ( 84 . 6% ) of the 488 non-boosted participants had RFFIT RVNA titers >0 . 5 IU/mL . In 2008 , three years after vaccination , 352 ( 88 . 0% ) of the 400 evaluable , non-boosted participants had titers >0 . 5 IU/mL , while in 2009 , four years after vaccination , 312 ( 85 . 7% ) of the 364 evaluable non-boosted participants had RFFIT titers >0 . 5 IU/mL ( Fig 2 ) . Nine ( 1 . 8% ) of the 507 participants had received rabies vaccine booster doses between vaccination in 2005 and enrollment in 2007 ( Table 3 ) . The time since the last vaccination was not known for five of them , but it was 0–6 months for one , 12–18 months for two , and 18–24 months for one . Additionally , 43 ( 9 . 3% ) of the 465 participants present at their follow up visit in 2008 ( 3 years after vaccination ) were boosted according to their titer measured during the 2007 campaign . The interval since the last vaccination was not known for 12 of the boosted participants , but was 0–6 months for 31 , and 6–36 months for the remaining five . Fourteen ( 3 . 3% ) of 428 participants received a booster at the 4-year follow up in 2009 . In 2009 , 41 ( 80 . 4% ) of the 51 evaluable boosted participants had RFFIT titers >0 . 5 IU/mL; the mean GMT was 1 . 33 IU/mL . The interval from the last vaccination was 12–18 months for 28 of the participants , 18–24 months for six , 42–48 months for one , and was not known for 16 . In the non-boosted population , GMTs ( Table 4 ) were significantly higher in young participants 2–5 and 6–15 years of age and the proportion of subjects with RFFIT titers >0 . 5 IU/mL ( Fig 3 ) was only slightly decreasing at each year of follow-up . In subjects aged 60 years or older , GMTs were lower although mostly >1 IU/mL , except for a drop between 2008 and 2009 where the seroconversion rate also decreased from 83 . 3% to 66 . 7% . However , the number of subjects was limited and the proportion of those with RFFIT titers >0 . 5 IU/mL was not significantly lower compared to the other age groups . In the 16–40 years age group , both the GMTs ( around 1 IU/mL ) and the proportion of individuals with RFFIT titers >0 . 5 IU/mL was stable over the 4 years of follow up . In the 41–60 years age group , the situation was far more contrasted with significantly lower GMTs ( 0 . 53 to 0 . 77 IU/mL ) and proportion of subjects with RFFIT titers >0 . 5 IU/mL at inclusion and at the follow up visit in 2008 , 3 years after vaccination ( P <0 . 05 , Fisher exact test ) compared to the general study population . However , both values tended to increase over the years , thus suggesting that poor responders were progressively removed from the non-boosted population . Males had lower seroconversion rates than females at each follow up visit , with significant differences observed in 2008 ( P <0001 ) and 2009 ( P <0 . 008 , Chi squared test ) , 3 and 4 years after vaccination ( Fig 4 ) . Significant gender differences were also observed , with males having lower RFFIT GMTs than females at each year of follow-up . GMTs ranged from 1 . 27 [95% CI: 1 . 11–1 . 44] in 2007 to 1 . 13 [95% CI: 1 . 02–1 . 24] in 2009 in females and from 0 . 98 [95% CI: 0 . 87–1 . 11] to 0 . 91 [95% CI: 0 . 83–0 . 99] in males over the same years ( Fig 4 ) . At each study visit , similar percentages of neutralizing antibody ( RFFIT ) titers >0 . 5 IU/mL were observed in the non-boosted participants who were given PrEP ( 3 vaccine doses , n = 58 ) in 2005 and in those receiving a PEP regimen ( five vaccine doses , n = 448 ) at each study visit ( Fig 5 ) . Similar GMTs were also observed in the PrEP and PEP groups using the RFFIT assay , ranging from 1 . 0 to 1 . 1 IU/mL each year of follow up ( Fig 6 ) . All specimens collected in 2007 were retested with the FAVN assay to determine the correlations with the RFFIT assay and ELISA ( Table 5 ) . In both the non-boosted and boosted populations , strong correlations of the GMT values obtained with the FAVN and RFFIT assays were observed , r = 0 . 92 for the non-boosted ( Table 5 ) and r = 0 . 99 for the boosted participants ( Table 6 ) . There was a good concordance of the seroconversion rates determined by FAVN ( 86 . 5% ) and the RFFIT ( 84 . 6% ) assays , with κ = 0 . 86 . In the non-boosted population , the Pearson’s correlation coefficient for FAVN and ELISA was 0 . 83 , ( 95% CI: 0 . 80–0 . 86 ) . The result in the boosted population was similar ( r = 0 . 95 ) . There was a strong correlation between RFFIT and ELISA results ( Pearson’s correlation coefficient ) in the non-boosted population r = 0 . 82 at inclusion , which however progressively decreased over the years to 0 . 71 at the 1-year follow-up and 0 . 62 at 2-year follow-up . There was a good concordance of the proportion of titers >0 . 5 determined by the RFFIT ( IU/mL ) or the ELISA ( EU/mL ) assays; however the same trend was observed . The Kappa coefficient ( κ ) in the non-boosted population was 0 . 61 at inclusion , 0 . 54 at the 1-year follow-up and 0 . 42 at 2-year follow-up ( Table 7 ) . In summary , the strength of the association between RFFIT and ELISA decreased with time , as the GMTs obtained by RFFIT remained relatively unchanged over the duration of follow-up; and , unexpectedly , the ELISA values increased in the second and third years .
Overall long-term persistence of RVNAs was good , with 85 to 88% of the non-boosted study population remaining seroconverted ( RFFIT titer >0 . 5 IU/mL ) over the 3 years of follow-up ending in 2009 . The GMT of the population remained >1 IU/mL ( twice the WHO-recommended threshold ) at the end of follow-up . Persistence of RVNA following vaccination in 2005 was similar in participants given PrEP and those given PEP . These results are consistent with those reported in previous studies [21 , 22] , and are discussed below in the context of routine PrEP vaccination . Our results are in accordance with other serological studies demonstrating that RVNA titers equal to or greater than 0 . 5 IU/mL , which is the WHO-recommended threshold of seroconversion , can persist for several years after administration of a complete vaccination series [23] . These results therefore highlight the need to maintain and intensify rabies PrEP and PEP . There were gender- and age-related differences in RVNA persistence . Overall , females had significantly higher GMTs and higher seroconversion rates than males in 2008 and 2009 . These results are in line with some previous reports [24 , 25] , however a correlation between gender and immune response to rabies vaccine has not been established [26] . While some gender differences in this study were statistically significant , their clinical significance remains doubtful because the seroconversion rates remained above 80% and the GMTs above 0 . 90 IU/mL in both genders . Also , the persistence of RVNA , as measured by the seroconversion rate , was shorter in the population >60 years of age than in younger participants , but the difference was not significant , and GMTs decreased only slightly . Participants 16–40 years of age had lower immune responses than the other age groups , but the observed GMTs and seroconversion rates among that age group , at 0 . 91–1 . 0 IU/mL and 80 . 6–84 . 5% , respectively were similar to those observed in previous studies [22] . The GMT and seroconversion rate point values were lower in those 41–60 years of age than in the other age groups , and both increased over the duration of follow-up . These values may have been influenced by a relatively small sample size and broad 95% CIs . They also indicated the progressive removal of the poor responders from the study which mostly focused on the non-boosted population . One of the limitations of the study is that only those subjects who responded well to the initial vaccination , i . e . remained seroconverted throughout follow up , were evaluated for antibody persistence . Subjects whose RFFIT antibody titer fell below or equal to 0 . 5 IU/mL were boosted and were excluded from the analysis to avoid any bias in evaluating antibody titers during subsequent follow up visits . Ideally , the analysis should have included all study subjects; however , it would have been both unethical and contrary to the design of our study ( based on the recommendations presented in the leaflet of the rabies vaccine Verorab ) not to vaccinate those with low antibody levels and expose them to the risk of rabies disease . RVNA titers are generally measured by RFFIT [17] or FAVN , the gold standard assays recommended by the WHO [16] . Nevertheless , for additional analyses , an ELISA using rabies virus glycoprotein as antigen ( Platelia Rabies II ) is available [27 , 28] . Although this ELISA method does not measure human RVNA but all anti-glycoprotein G antibodies , it is easier and more rapid to perform . Rapid assays should be encouraged to facilitate diagnosis in rural settings lacking sophisticated techniques and qualified personnel . Increasingly discordant results were obtained with the RFFIT and ELISA assays from 2007 to 2009 . The reason for these discordant results and for the decrease in correlation and concordance is not clear and deserves further investigation using well characterized proficiency panels . A previous comparison of these two assays found that the results of each corresponded closely except in samples with high RFFIT titers [27] . This study and the dramatic events preceding it underscore the need for the health authorities in rabies enzootic countries to decide on the best timing for the introduction of routine rabies PrEP vaccination in affected areas even if regular titer checks and boosters , may not appear affordable for the developing economies . This introduction would also prevent the need for serotherapeutic treatment , a real advantage in developing countries where human or equine rabies immunoglobulins are scarce and expensive . Ideally , routine pre-exposure rabies vaccination should be included in the Expanded Program on Immunization ( EPI ) schedule , given concomitantly with other pediatric vaccines . Two studies have evaluated the concomitant administration of rabies and DTP vaccines in Vietnam , and a third evaluated the concomitant administration of rabies and Japanese encephalitis vaccines . The first Vietnamese study in infants showed that PVRV can be administered concomitantly with DTP-IPV as 2 IM doses at 2 and 4 months of age and a booster dose 1 year later with satisfactory safety and immunogenicity results and with no interference between the 2 vaccines [29] [30] [31] . Similar findings were drawn from another study conducted in Vietnam where PVRV was co-administered with DTP-IPV as 3 ID or 2 IM injections . The study showed that there was no apparent interference between the 2 vaccines and confirmed that their co-administration was safe in infants and toddlers [32] [33] . Finally , a study conducted in Thailand confirmed that the co-administration of a purified chick embryo cell vaccine ( PCECV ) with Japanese encephalitis vaccine ( JEV ) is safe and confers satisfactory immune response without interference between the two vaccines [34] . These clinical studies strongly suggest that rabies vaccine may be co-administered with routine pediatric vaccines and support integration of rabies PrEP vaccination into the childhood immunization schedules of countries where rabies is enzootic . This would minimize the costs and practical difficulties associated with the introduction of rabies PrEP into routine immunization practice . Shortened PEP vaccination regimens that require less than 1 month to complete are also particularly relevant in rural populations in rabies-endemic countries . They require fewer visits to the vaccination center , potentially resulting in better compliance . One option is abbreviated 4-dose IM schedule , which requires 2 weeks for completion . Preliminary data from studies conducted in Thailand [35] and India [36] suggest that a 1-week 4-4-4 intradermal ( ID ) PEP regimen is an alternative option to consider . An ongoing study in The Philippines ( ClinicalTrials ID no . NCT01622062 ) is evaluating the 1-week 4-4-4 ID PEP regimen followed by a single-visit four-site ID booster vaccination at five years . The surveillance results obtained in this study should encourage health authorities in rabies-enzootic countries to investigate the best strategies and timing for introduction of routine rabies PrEP vaccination in affected areas . In terms of PEP regimens , our observation that a complete 3-dose PrEP schedule induced similar GMTs and similar percentages of vaccinees with RVNA titers >0 . 5 IU/mL compared to a complete 5-dose PEP schedule is in favor of abbreviated schedules .
|
Animal control measures have decreased the incidence of human rabies in urban regions of Latin America . Currently , most cases of human rabies occur in rural areas and are transmitted by bats . In 2004–2005 , rabies outbreaks affecting populations living in remote areas of Brazil prompted the widespread vaccination of exposed and at-risk populations . We evaluated the persistence of the humoral immune response for 4 years after vaccination in a rural population at daily risk of rabies exposure . Our aim was to evaluate the impact of vaccination in a rural setting to help improve management of vampire bat exposure . The participation of this Amazonian community was excellent , with 428 of the 509 ( 84% ) who attended the first interview in 2007 returning for the final visit in 2009 . The long-term RVNA persistence was good , with 85–88% of the participants evaluated at each yearly follow-up visit remaining seroconverted . Similar neutralizing antibody persistence levels were observed in participants originally given post-exposure or pre-exposure prophylaxis in 2005 . This study and the events preceding it underscore the need for the health authorities in rabies enzootic countries to decide on the best strategies and timing for the introduction of routine rabies PrEP vaccination in affected areas .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"enzyme-linked",
"immunoassays",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"tropical",
"diseases",
"microbiology",
"vertebrates",
"animals",
"mammals",
"vaccines",
"preventive",
"medicine",
"viruses",
"rabies",
"rna",
"viruses",
"neglected",
"tropical",
"diseases",
"vaccination",
"and",
"immunization",
"antibodies",
"immunologic",
"techniques",
"research",
"and",
"analysis",
"methods",
"rabies",
"virus",
"public",
"and",
"occupational",
"health",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"booster",
"doses",
"zoonoses",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"immunoassays",
"biochemistry",
"lyssavirus",
"viral",
"pathogens",
"physiology",
"bats",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"amniotes",
"organisms"
] |
2016
|
Persistence of Rabies Virus-Neutralizing Antibodies after Vaccination of Rural Population following Vampire Bat Rabies Outbreak in Brazil
|
We carried out an admixture analysis of a sample comprising 1 , 019 individuals from all the provinces of Cuba . We used a panel of 128 autosomal Ancestry Informative Markers ( AIMs ) to estimate the admixture proportions . We also characterized a number of haplogroup diagnostic markers in the mtDNA and Y-chromosome in order to evaluate admixture using uniparental markers . Finally , we analyzed the association of 16 single nucleotide polymorphisms ( SNPs ) with quantitative estimates of skin pigmentation . In the total sample , the average European , African and Native American contributions as estimated from autosomal AIMs were 72% , 20% and 8% , respectively . The Eastern provinces of Cuba showed relatively higher African and Native American contributions than the Western provinces . In particular , the highest proportion of African ancestry was observed in the provinces of Guantánamo ( 40% ) and Santiago de Cuba ( 39% ) , and the highest proportion of Native American ancestry in Granma ( 15% ) , Holguín ( 12% ) and Las Tunas ( 12% ) . We found evidence of substantial population stratification in the current Cuban population , emphasizing the need to control for the effects of population stratification in association studies including individuals from Cuba . The results of the analyses of uniparental markers were concordant with those observed in the autosomes . These geographic patterns in admixture proportions are fully consistent with historical and archaeological information . Additionally , we identified a sex-biased pattern in the process of gene flow , with a substantially higher European contribution from the paternal side , and higher Native American and African contributions from the maternal side . This sex-biased contribution was particularly evident for Native American ancestry . Finally , we observed that SNPs located in the genes SLC24A5 and SLC45A2 are strongly associated with melanin levels in the sample .
The post-Columbian history of the Caribbean has been marked by the encounter of people from different continents . This is reflected in the gene pool of the present inhabitants of the Caribbean archipelago , as shown in recent studies using autosomal , mtDNA and Y-chromosome markers [1]–[7] . However , very few studies have focused on Cuba , the largest island of the Greater Antilles [8] , [9] . Evidence of human habitation in Cuba goes back to approximately 7 , 000 years BP [10] , [11] . It has been estimated that at the arrival of the Spaniards there were around 110 , 000 indigenous people living on the island [12] . At the time of contact there were two indigenous groups in Cuba . The Guanahatabey were hunter-gatherers living in western Cuba . They comprised approximately 10% of the indigenous Cuban population , spoke a non-Arawak language and have been considered to be the descendants of the earliest settlers of the island . The Taino were Arawak-speaking agriculturalists inhabiting the rest of the island , and comprised 90% of the indigenous population . The most accepted hypothesis is that both groups migrated from South America ( lower Orinoco Valley ) [11] , [12] . However , North American ( Florida ) and Mesoamerican ( Yucatan , Honduras and Nicaragua ) migrations have also been postulated by some authors , particularly for the earliest settlers of the island [11] , [12] . Within 50 years of the arrival of Columbus , the indigenous Cuban population had been decimated to a few thousand people . The Spaniards then started to relocate indigenous people from North America and Mesoamerica to Cuba , as well as enslaved Africans , primarily from the West Coast of Africa [9] , [13] . It has been estimated that between 700 , 000 and 1 , 300 , 000 Africans were brought to Cuba during the slave trade period [14] , [15] . Immigration from Spain took place throughout the colonial and post-colonial periods , until the first half of the 20th century . Historical records indicate that most of the immigrants from Spain were male ( 60–85% ) , and that mixing between European males and indigenous and African females occurred since the early stages of the colonization of the island [12] . Therefore , the present genetic structure of the Cuban population has been shaped by the history of admixture between indigenous Americans , Europeans and Africans . Today , the Cuban census classifies the population into three categories: “Blancos” ( “White” ) , “Mestizos” ( “Mixed” ) and “Negros” ( “Black” ) [16] . Here , we present an admixture analysis of a large sample comprising 1 , 019 individuals from the 16 provinces of Cuba . We used a panel of 128 Ancestry Informative Markers ( AIMs ) to estimate the admixture proportions . In addition to the AIMs , we also characterized a number of haplogroup diagnostic SNPs in the mtDNA and Y-chromosome in order to evaluate admixture using uniparental markers . Finally , we also evaluated the association of 16 single nucleotide polymorphisms ( SNPs ) with skin pigmentation . This study is relevant from different points of view . Understanding the distribution of admixture proportions throughout Cuba is important from an anthropological perspective , and this is the most extensive effort carried out to date in terms of the size and representativeness of the sample . Additionally , the study of uniparental markers provides interesting evidence regarding the directionality of gene flow . The elucidation of admixture proportions is also of interest for future application of admixture mapping studies or genome-wide association studies in Cuba . Finally , we show that SNPs located in the genes SLC24A5 and SLC45A2 are strongly associated with melanin levels in the sample .
Estimates of admixture proportions were obtained with the program ADMIXMAP , using data from 128 AIMs . In the total sample , the average European , African and Native American contributions were 72% ( range 4 . 3% to 98 . 2% ) , 20% ( range 0 . 8% to 95 . 2% ) and 8% ( range 0 . 4% to 34% ) , respectively ( Figure 1 ) . By province , the average proportion of European ancestry ranged from 51% in Santiago de Cuba to 84% in Mayabeque , the average proportion of African ancestry ranged from 11% in Mayabeque and Sancti Spíritus to 40% in Guantánamo , and the average proportion of Native American ancestry from 4% in Matanzas to 15% in Granma ( Figure 1 ) . There are significant differences in admixture proportions between provinces ( ANOVA: Africans F = 11 . 54 , P<0 . 001; Native American F = 13 . 06 , P<0 . 001 ) . Post-hoc tests indicate that , in terms of African proportions , the differences are driven by the higher African proportions in the provinces of Santiago de Cuba ( 39% ) and Guantánamo ( 40% ) , with respect to the other provinces ( 11% to 24% ) . With respect to the Native American contributions , a clear pattern is also present , with higher average contributions in the Eastern provinces , particularly Granma ( 15% ) , Las Tunas ( 12% ) and Holguín ( 12% ) than in the Western provinces . As indicated above , 55% of the participants self-reported to be “blanco” , 33% “mestizo” and 12% “negro” . These proportions are similar to those based on the report of an external observer; there were discrepancies for only 65 out of the 1019 individuals . Several measures of concordance indicated excellent agreement between both classifications ( Cohen's kappa = 0 . 8873 [17] , Ciccheti-Allison's kappa = 0 . 9091 [18] and Fleiss-Cohen's kappa = 0 . 9345 [19] ) . Age did not have a significant effect on melanin levels ( M ) , measured quantitatively with the reflectometer ( melanin index ) [20] , but there were significant differences in melanin index by sex ( males M = 40 . 68±10 . 7; females M = 39 . 17±9 . 45; P = 0 . 015 ) . The average melanin index of the total sample was 39 . 8 , but there was a broad distribution of values , from 23 . 4 to 85 . 9 . In individuals who self-reported to be “blanco” , the average melanin index was 34 . 06±3 . 70 ( mean ± SD ) , in those who self-reported to be “mestizo” 41 . 69±6 . 29 and in those who self-reported to be “negro” 60 . 59±8 . 87 ( Figure 2 ) . The differences in melanin levels between census groups were significant ( ANOVA with sex as a covariate: F = 4 . 30 , P<0 . 001 ) . The average European , African and Native American ancestry in those self-reporting to be “blanco” were 86% , 6 . 7% and 7 . 8% , in those self-reporting to be “mestizo” 63 . 8% , 25 . 5% and 10 . 7% , and in those self-reporting to be “negro” 29% , 65 . 5% , 5 . 5% ( Figure 3 ) . ( ANOVA European: F = 1048 . 04 , P<0 . 001; African: F = 1138 . 97 , P<0 . 001; Native American: F = 34 . 19 , P<0 . 001 ) . The levels of pigmentation show a strong correlation with the estimates of individual ancestry proportions obtained with the panel of AIMs . African ancestry was positively correlated with the melanin index ( Spearman's rho = 0 . 632 , P<0 . 001 ) , and European ancestry was inversely correlated with melanin index ( rho = −0 . 659 , P<0 . 001 ) . No significant correlation was observed between Native American ancestry and melanin index ( rho = 0 . 0547 , P = 0 . 0809 ) . The analysis of melanin index distribution by province revealed that the samples from Guantánamo ( GT ) and Santiago de Cuba ( SC ) have significantly higher melanin index values ( GT average M = 47 . 51 , SC average M = 46 . 77 ) ( Figure 4 ) . The data show clear geographic trends in admixture proportions in Cuba . For example , the average African ancestry in the provinces of Guantánamo and Santiago de Cuba is higher than in the other provinces . In principle , this could be explained by two different scenarios , which are not mutually exclusive: ( i ) African admixture proportions are higher in Guantánamo and Santiago de Cuba because these provinces have higher proportions of individuals self-reporting to be “negro” or “mestizo” , who on average have higher African contributions than individuals reporting to be “blanco” , or ( ii ) . There are no differences in the proportion of individuals self-reporting to be “blanco” , “mestizo” or “negro” between Guantánamo and Santiago de Cuba and the other provinces , but the average African admixture contributions in at least some of the census categories are higher in Guantánamo and Santiago de Cuba than in the other provinces . In order to evaluate these two scenarios , we explored the relationships between African admixture proportions and the proportion of individuals in each province reporting to be “negro” , “mestizo” , or “blanco” . We observed a strong positive relationship between average African ancestry in each province and the proportion of individuals reporting to be “negro” or “mestizo” ( r2 = 0 . 69 , P = 7×10−5 , and r2 = 0 . 63 , P = 0 . 63 , P = 2×10−4 , see also Figure S1 ) . Therefore , the higher African admixture proportions in Guantánamo and Santiago de Cuba are due , to a considerable extent , to the higher proportions of self-reported “negro” and “mestizo” in these provinces . We also observed a positive relationship between the proportion of individuals reporting to be “mestizo” and Native American ancestry across provinces , although this relationship is not as strong as that observed for African ancestry ( r2 = 0 . 43 , P = 6×10−3 ) . In addition to the relationship of ancestry and census proportions by province , we also explored to which extent there are differences in admixture proportions within each census category ( “blanco” , “mestizo” and “negro” ) between provinces ( Figure S2 ) . The presence of differences in ancestry proportions within each census category would indicate that provincial differences in ancestry proportions are not only due to differences in the relative proportions of individuals from each census category . We observed some differences in ancestry proportions within census categories . For example , within individuals self-reporting to be “blanco” , the average African admixture proportions are significantly higher in Guantánamo , Santiago de Cuba and Granma than in many other provinces , and within individuals self-reporting to be “negro” , the average African admixture proportions are significantly lower in Las Tunas , Holguín and Granma than in Guantánamo , Santiago de Cuba , Camagüey and La Habana ( data not shown ) . We explored if there are differences in ancestry proportions estimated with AIMs between rural and urban areas . For the total sample , we observed that the African ancestry proportions were significantly higher in urban than rural areas ( P = 0 . 003 ) , and conversely , the Native American ancestry proportions were significantly higher in rural than urban areas ( P = 2×10−6 ) ( Figure S3 ) . A plot showing ancestry proportions in rural and urban areas by province is depicted in Figure S4 . The results of a two-way ANOVA and post-hoc tests indicate that the difference in African ancestry proportions between urban and rural areas is primarily driven by the higher African ancestry in individuals reporting to be “negro” living in urban areas vs . those living in rural areas . In contrast , the average Native American contribution in individuals self-reporting to be “negro” living in rural areas is higher than in those living in urban areas , and this is the main factor explaining the higher Native American ancestry in rural vs . urban areas . No significant differences between rural and urban areas were observed for African or Native American ancestry for individuals reporting to be “blanco” or “mestizo” . A total of 943 mtDNA haplotypes could be allocated to a specific branch of the mtDNA phylogeny resolved by the mtSNPs genotyped in the present study ( see the mtDNA phylogeny of Figure S5 ) . A detailed list of the haplogroup assignations based on the 18 markers genotyped in this study is presented in Table S5 . The analysis of mtSNPs indicates that 34 . 5% of the mtDNA haplotypes have Native American ancestry , 38 . 8% African ancestry , and 26 . 7% Eurasian ancestry ( Figure 5 ) . The highest maternal Eurasian proportions were found in the provinces of Matanzas ( 58% ) , Artemisa ( 53% ) , and Pinar del Rio ( 49% ) and the lowest in Santiago de Cuba ( 6% ) , Granma ( 7% ) and Holguín ( 7 . 5% ) . The highest maternal African proportions were observed in the provinces of Santiago ( 57% ) and Granma ( 52% ) , and the lowest in Las Tunas ( 21% ) and Camagüey ( 24% ) . With respect to the maternal Native American proportions , the highest were found in Holguín ( 59% ) and Las Tunas ( 58% ) , and the lowest in Matanzas ( 13% ) , Cienfuegos ( 13% ) and Pinar del Río ( 13% ) . An analysis of contingency tables using exact tests ( Table S6 ) indicates that many of the Western provinces have significantly higher Eurasian proportions than some of the Eastern provinces , in particular Holguín , Granma and Santiago de Cuba . These tests also show that the province of Santiago de Cuba has significantly higher African proportions than other Cuban provinces , and that the provinces of Holguín , Las Tunas and to some extent , Granma , have significantly higher Native American proportions than most of the Western provinces . Y-chromosome SNPs could be genotyped in 384 males and haplotypes were classified into haplogroups following the phylogeny of Figure S6 . A detailed list of the haplogroup assignations based on the 12 Y-SNPs genotyped in this study is presented in Table S7 . Most of the haplotypes are of Eurasian ancestry ( 81 . 8% ) , while 17 . 7% have African ancestry and only two haplotypes are of Native American ancestry ( 0 . 5% ) ( Figure 6 ) . The Native American haplotypes belong to two individuals , one from the province of Camagüey and the other from Santiago de Cuba . Regarding Eurasian and African ancestry , the highest Eurasian paternal contributions were found in Matanzas , and Pinar del Río , and the highest African paternal contributions correspond to the province of Santiago de Cuba . Although the size of the Y-chromosome sample was substantially smaller than the mtDNA sample , the contingency table analysis ( Table S8 ) identified significant differences in paternal Eurasian contributions between Matanzas and Villa Clara , Cienfuegos and Santiago , and also between Pinar del Río and Guantánamo and Santiago . The province of Santiago showed a significantly higher African paternal contribution than Pinar del Río , Matanzas and Guantánamo . Sixteen genetic markers located within or nearby genes that previously have demonstrated association with skin pigmentation ( APBA2 –linked to OCA2– , ASIP , BNC2 , GATA3 , GRM5 –linked to TYR– , HERC2–linked to OCA2– , IRF4 , KITLG , MC1R , OCA2 , SLC24A5 , SLC45A2 –also known as MATP– , TYR , TYRP1 and UGT1A1 ) [21]–[37] were analyzed for association with melanin levels measured quantitatively . The program ADMIXMAP was used to run a linear regression analysis conditioning on individual ancestry . Of the 16 markers analyzed , four were significantly associated with melanin index after Bonferroni correction ( P<0 . 0031 ) : rs1426654 located on the SLC24A5 gene ( P = 1 . 2×10−25 ) , rs16891982 and rs35395 located on the SLC45A2 ( MATP ) gene ( P = 1 . 7×10−20 and P = 2 . 8×10−11 , respectively ) , and rs12913832 located on the HERC2 gene , linked to OCA2 ( P = 0 . 0018 ) . In order to evaluate if there was evidence of residual stratification unaccounted for in the analysis based on the three-parental model , we used the P-values obtained for 86 AIMs located more than 5 cM apart from the 16 pigmentation markers to estimate the lambda inflation factor . We observed evidence of residual stratification ( lambda = 1 . 38 ) . Therefore , we implemented genome control ( GC ) [38] methods to correct for type I error inflation . After GC-correction , SLC24A5 rs1426654 ( P = 5 . 1×10−19 ) , SLC45A2 rs16891982 ( P = 2 . 9×10−15 ) and SLC45A2 rs35395 ( P = 1 . 5×10−8 ) remained significant after Bonferroni correction . However , the P-value for HERC2 rs12913832 ( P = 0 . 0078 ) slightly exceeded the Bonferroni-corrected threshold . Table 2 reports the GC-corrected P-values for all the pigmentation markers . Assuming an additive model , we estimated that each copy of rs1426654 allele A and rs16891982 allele G decrease the melanin index by 5 . 04 and 3 . 40 units , respectively . The HERC2 SNP rs12913832 has a substantially smaller effect , with each copy of the G allele , which has been associated with blue iris color in previous studies [25]–[27] , decreasing melanin index by approximately 1 . 11 units . Finally , we repeated the analysis including the genotypes of rs1426654 , rs16891982 and rs12913832 as covariates . This analysis showed that the P-value observed for rs35395 at the SLC45A2 locus was no longer significant , indicating that the significant result for this marker is primarily due to its linkage with rs16891982 , which is located approximately 3 kb apart from rs35395 on chromosome 5 . None of the other 12 SNPs surveyed had significant effects on melanin levels after conditioning for the rs1426654 , rs16891982 and rs12913832 polymorphisms .
By genotyping a panel of autosomal AIMs in combination with mtDNA and Y-chromosome markers in a large sample representative of all Cuban provinces , we were able to identify very clear patterns in the distribution of admixture proportions throughout Cuba . The analysis using AIMs indicated that the average European , African and Native American contributions were 72% , 20% and 8% , respectively . However , the African and Native American contributions were relatively higher , and the European contributions lower , in the Eastern provinces than in the Western provinces . In particular , the Southeastern provinces , such as Santiago de Cuba and Guantánamo , showed the highest African proportions , and the highest Native American proportions were found in the Eastern provinces of Granma , Holguín and Las Tunas . Similar geographic patterns were observed in the analyses of the uniparental markers . Additionally , by comparing the autosomal and uniparental admixture proportions , we identified a clear sex-biased pattern in the process of gene flow , with a substantially higher European contribution from the paternal side than the maternal side , and conversely higher Native American and African contributions from the maternal side than the paternal side . The geographic patterns observed for the admixture proportions are consistent with historical and archaeological evidence . The identification of sex-biased gene flow is also in agreement with historical information indicating that most of the European immigrants throughout Cuban history were male and that the process of admixture took place primarily between European males and Native American and African females . Finally , we observed that SNPs located in the genes SLC24A5and SLC45A2 are significantly associated with skin pigmentation in the sample , in accordance with what has been reported in other admixed populations .
The study was approved by the Research Ethics Committee of the National Centre of Medical Genetics of Cuba . Each individual in this study gave written informed consent prior to the interview , physical examination and blood sample collection . The final sample comprised 1 , 019 individuals representing all the provinces of Cuba . The selection of the individuals was made in collaboration with the National Statistics Office from Cuba . The individuals were selected based on the demographic characteristics of the Cuban population in terms of population density , age , gender and census category ( “Blanco” , “Mestizo” , “Negro” ) . Individuals were recruited from 1 , 229 households , located in 137 of the 168 Cuban municipalities . Selection of individuals from each household was based on the Kish grid , in order to ensure that all the members of the household had the same probability of being selected for the study . The final sample represents quite well the current distribution of the Cuban population in terms of sex , age , census category ( “Blanco” , “Mestizo” , and “Negro” ) , provincial population density and rural/urban residence . A detailed comparison of the relative proportions of each category in the study sample and the Cuban census is provided as supplementary information ( see Table S1 ) . Researchers visited 1 , 182 of the 1 , 229 selected households and 1 , 031 individuals volunteered to participate in the study . Due to problems with DNA quality , 12 samples were excluded from the final analyses . Information about individual , parent and grandparents place of birth , demographics , education level , physical health , mental disorders , non-communicable disease risk factors and anthropometry was collected via questionnaire and physical examination . Information about census category was obtained in two ways: self-reported by the participants and independently classified by one trained researcher ( EFS ) for all the individuals included in the study . The concordance between the two classifications was evaluated using Cohen's kappa coefficient [17] , and also the Ciccheti-Allison [18] and Fleiss-Cohen [19] weighted kappa coefficients . Melanin content of the skin was measured with a narrow band reflectometer ( DSM II ColorMeter , Cortex Technologies , Hadsund , Denmark ) [20] . This instrument provides quantitative estimates of melanin levels ( e . g . melanin index ) . The measurements were taken at the medial side of the upper inner arm , an area of the body not exposed to the sun ( constitutive pigmentation ) , and also at the dorsum of the hand , an area with substantial exposure to the sun ( facultative pigmentation ) . Average admixture proportions , the sum of intensities parameter ( equivalent to the average number of generations since the admixture event ) and the individual ancestry proportions were estimated using the software ADMIXMAP v3 . 8 for Windows . This is a general purpose program for modeling population admixture with genotype and phenotype data , based on a combination of Bayesian and classical methods . If information for a quantitative trait ( such as skin pigmentation ) is provided , ADMIXMAP fits a linear regression model of the trait conditioning upon individual admixture . Covariates such as sex and age can be included in this model . Detailed information about this program can be found in Hoggart et al . [55] , [56] . In order to estimate admixture proportions; we used the prior allele frequency model , which requires information about the prior distribution of allele frequencies in each ancestral population . Under this model , the program estimates the allele frequencies from unadmixed and admixed population samples simultaneously , allowing for sampling error . ADMIXMAP implements a diagnostic test for variation of allele frequencies between the unadmixed populations that were sampled to obtain prior parameters and the corresponding ancestry-specific allele frequencies in the admixed sample . The program was run with 20 , 000 iterations , including 1 , 000 iterations for burn-in of the Markov chain . Differences between provinces and between sexes for the melanin index and the ancestral genetic proportions were assessed using one-way ANOVA . The relationship between age and skin pigmentation was assessed by the parametric Pearson correlation test and also the non-parametric Spearman's rho test . A two-way ANOVA was conducted in order to evaluate the relationship between melanin index and skin color using sex as a covariate . Finally , potential differences in the distributions of mtDNA and Y-chromosome haplogroups among provinces were evaluated using exact tests . The above described statistical analyses were performed in Statistic 7 . 0 and SPSS 20 . 0 .
|
Cuba is the largest island of the Greater Antilles and its most populous country . The post-Columbian history of the Caribbean has been marked by the encounter of people from different continents . Here , we present an admixture analysis of 1 , 019 individuals from all the provinces of Cuba , using autosomal , mtDNA and Y-chromosome markers . We also analyzed the association of 16 single nucleotide polymorphisms ( SNPs ) with quantitative estimates of skin pigmentation ( melanin index ) . The highest proportions of African ancestry were observed in the Southeastern provinces of Santiago de Cuba and Guantánamo , and the highest proportions of Native American ancestry were found in the Eastern provinces of Granma , Holguín and Las Tunas . Similar geographic patterns were observed in the analyses of the uniparental markers . Additionally , by comparing the autosomal and uniparental admixture proportions , we identified a clear sex-biased pattern in the process of gene flow , with a substantially higher European contribution from the paternal side than the maternal side , and conversely higher Native American and African contributions from the maternal side than the paternal side . Finally , we observed that SNPs located in the genes SLC24A5 and SLC45A2 show a strong association with skin pigmentation in the sample .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"haplotypes",
"medicine",
"and",
"health",
"sciences",
"genetic",
"polymorphism",
"genetic",
"association",
"studies",
"epidemiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"population",
"genetics",
"gene",
"flow",
"human",
"genetics",
"evolutionary",
"biology",
"genetic",
"epidemiology"
] |
2014
|
Cuba: Exploring the History of Admixture and the Genetic Basis of Pigmentation Using Autosomal and Uniparental Markers
|
Lyme disease , caused by Borrelia burgdorferi , is a vector-borne illness that requires the bacteria to adapt to distinctly different environments in its tick vector and various mammalian hosts . Effective colonization ( acquisition phase ) of a tick requires the bacteria to adapt to tick midgut physiology . Successful transmission ( transmission phase ) to a mammal requires the bacteria to sense and respond to the midgut environmental cues and up-regulate key virulence factors before transmission to a new host . Data presented here suggest that one environmental signal that appears to affect both phases of the infective cycle is osmolarity . While constant in the blood , interstitial fluid and tissue of a mammalian host ( 300 mOsm ) , osmolarity fluctuates in the midgut of feeding Ixodes scapularis . Measured osmolarity of the blood meal isolated from the midgut of a feeding tick fluctuates from an initial osmolarity of 600 mOsm to blood-like osmolarity of 300 mOsm . After feeding , the midgut osmolarity rebounded to 600 mOsm . Remarkably , these changes affect the two independent regulatory networks that promote acquisition ( Hk1-Rrp1 ) and transmission ( Rrp2-RpoN-RpoS ) of B . burgdorferi . Increased osmolarity affected morphology and motility of wild-type strains , and lysed Hk1 and Rrp1 mutant strains . At low osmolarity , Borrelia cells express increased levels of RpoN-RpoS-dependent virulence factors ( OspC , DbpA ) required for the mammalian infection . Our results strongly suggest that osmolarity is an important part of the recognized signals that allow the bacteria to adjust gene expression during the acquisition and transmission phases of the infective cycle of B . burgdorferi .
Borrelia burgdorferi , the Lyme disease agent , survives and grows in mammals and various vertebrate hosts . However , the bacteria are not transmitted directly to a new host . Instead they are acquired by a hematophagous arthropod ( Ixodes scapularis ) and transmitted to a new host . Cycling between a host and the vector requires the bacteria to adapt and survive in both mileus . The environment of the host is well defined: nutrient-rich , constant temperature , stable pH , established ion concentrations and osmolarity [1] . Overall , the mammalian host provides the bacteria with a very steady environment for survival provided that they can successfully evade an aggressive host immune system . In contrast , the tick presents a more variable environment with parameters that are gradually changing before , during and after feeding . Acquisition of B . burgdorferi begins when uninfected ticks begin feeding on infected mammals . Initially , this colonization is characterized by rapid growth of the bacteria and regulation of gene expression by the two-component system ( TCS ) , Response Regulator 1 ( Rrp1 ) and Histidine Kinase 1 ( Hk1 ) [2–4] . As the blood meal is consumed , a feast-famine succession that lasts for several weeks slowly converts B . burgdorferi from rapid growth to stationary phase . During this progression , Borrelia adjusts gene expression for long-term survival via regulatory networks mediated by RelBbu ( RelA/SpoT homolog ) , the Borrelia oxidative stress regulator ( BosR ) and σS ( RpoS ) [5–7] . After molting to the next developmental stage , the ticks begin the next feeding and parameters in the midgut revert . B . burgdorferi , localized specifically to the midgut , begin to grow and alter gene expression according to reconstituted feeding conditions ( replenished nutrients , temperature , etc . ) . Some of these conditions act as signals to upregulate key virulence and transmission factors ( OspC , DbpA , BBA66 , etc . ) via the Rrp2/RpoN/RpoS regulatory cascade [8 , 9] . The way the midgut environmental conditions affect the expression of this regulatory system and virulence factors required for the successful transmission has been extensively studied [10] . In several cases , in vitro conditions have been used to mimic parameters that are suspected to exist or have been measured in the tick midgut or the blood meal [11–15] . In addition , the transcription of virulence related genes has been assayed directly from B . burgdorferi RNA extracted from feeding ticks [16] . Interestingly , one parameter that has been completely overlooked is osmolarity . Because of the extended feeding time of Ixodes ticks ( 5 to 6 days ) , water from the blood meal must be recycled through the hemolymph to the salivary glands to generate adequate saliva for prolonged feeding . This water flux is followed by a corresponding flux of ions such as Na+ , K+ and Ca2+ . In Dermacentor andersonii , Kaufman and Phillips demonstrated , by directly measuring the ion concentration , that the osmolarity changes throughout feeding [17–19] . Early studies suggest that , in Ixodes ricinus , the salivary glands function in osmoregulation and facilitate the recycling of 70% of the water from the blood meal to the salivary glands [20 , 21] . These studies suggest that B . burgdorferi should encounter osmotic conditions in the feeding tick midgut that are generated by water and ion flux necessary to produce the saliva required for successful feeding . Bacteria respond to the physiological changes associated with changes in osmolarity by a process known as osmoadaptation [22–24] . Osmoadaptation is classically associated with the synthesis or uptake of a limited set of molecules called compatible solutes [22 , 25 , 26] . There are two categories of compatible solutes: solutes that have no effect on growth , and those that do have an effect on growth ( osmoprotective molecules ) [27] . Bacteria use osmoprotective molecules to modulate their intracellular osmolarity so they can grow and divide [22 , 28] . In E . coli , the increasing K+ concentration is directly related to the increase of the environmental osmolarity [29 , 30] . Potassium is also known to activate key regulatory proteins that are involved in the regulation of intracellular pH [31 , 32] . Bacteria also accumulate , by transport or de novo synthesis , specific amino acids like proline or glutamate [15 , 22 , 27 , 33] . The major role of the glutamate is to offset the uptake of K+ , which inhibits enzyme activity [32 , 34] . In this study , we determined the effects of changes in osmolarity on the virulence and physiology of B . burgdorferi . First , we measured the osmolarity ( mOsm ) in the bloodmeal , saliva and hemolymph isolated from feeding ticks , then tested B . burgdorferi cells for their ability to grow over a range of osmolarities . Surprisingly , Borrelia was only able to grow normally between 250 mOsm and 650 mOsm , which very closely matched the range of osmolarity in the bloodmeal during tick feeding ( ~300–600 mOsm ) . The growth , morphology and motility were dramatically affected by osmolarity outside of this narrow range . Interestingly , at low osmolarity , Borrelia cells expressed increased levels of virulence factors ( OspC , DbpA ) required for successful transmission . Finally , we analyzed the osmoadaptation by following the expression of the genes putatively involved in osmoregulation ( proU , gltP , etc . ) , or virulence ( ospC , dbpA , etc . ) at various osmolarities . Mutants that did not express the putative L-glutamate transporter ( gltP ) or proline transport ( proU ) system were more sensitive to changes in osmolarity than wild-type cells , suggesting that L-glutamate and proline were osmoprotective molecules . We hypothesize that bloodmeal osmolarity may directly affect the expression of key virulence factors and may serve as a physiological signal to trigger B . burgdorferi to migrate from the midgut to the salivary glands during transmission .
Based on previous observations in I . ricinus and D . andersonii [17–20] , we hypothesized that the osmolarity of the Ixodes tick midgut changes throughout its feeding cycle . To test this hypothesis , we measured the osmolarity of the midgut contents , hemolymph and saliva of I . scapularis in feeding nymph or adult ticks ( Fig 1 ) . Initially , we harvested feeding ticks from host animals at specific times ( days ) after attachment to analyze midgut contents . However , because of a lack of consistency in feeding efficiency between individuals in a feeding cohort , we decided to use scutal index to measure feeding progress . The scutal index is the ratio of the length of the idiosoma to the maximum width of the scutum ( see methods , Fig 1A and 1B ) [35] and this proved to be an effective method to evaluate the duration of feeding . In both nymph and adult ticks , it was not possible to measure the osmolarity of midgut contents before 24 h of the initiation of feeding due to the extremely low volume of recoverable midgut material . Therefore , feeding was interrupted by detaching ticks , scutal index were measured , the midgut contents were extracted , osmolarity was measured and the data were plotted as a function of the scutal index ( Fig 1C ) [35] . In adult ticks , the osmolarity began at ~ 550 mOsm at a scutal index of 2 and then decreased to ~300 mOsm ( with the lowest value measured at 264 mOsm ) at a scutal index between 5–7 . Finally , the osmolarity increased as feeding finished , and in replete ticks , returned to an osmolarity of ~550 mOsm at a scutal index between 7–8 ( Fig 1C , Adults ) . We also measured the changes in osmolarity in the midguts of feeding nymphs . Again , we were unable to measure the osmolarity at early time points . As the scutal index reached 3 . 5–4 , the osmolarity approached ~600 mOsm and then decreased to 300 mOsm at a scutal index of 5 ( Fig 1C , Nymphs ) . Again , as observed in the adult ticks , the osmolarity rebounded to ~500–600 mOsm at the conclusion of feeding . These data showed a very similar pattern of amplitude and fluctuation in osmolarity in the midgut of feeding adults and nymphs . We also measured the osmolarity of mouse ( 335 mOsm ± 2 . 8 ) and rabbit ( 304 . 7 mOsm ± 9 . 5 ) blood confirming previously published values [1] . Additionally , hemolymph and saliva were collected from feeding ticks and the osmolarity of biological triplicate samples were measured . Hemolymph ( 311 . 4 mOsm ± 33 . 3 ) and saliva ( 323 . 5 mOsm ± 6 . 4 ) had osmolarities very similar to those measured in host blood . Taken together , these data suggest that B . burgdorferi encounters little change in osmolarity in the hemolymph , saliva or in the mammalian host but faces a variation in osmolarity ( ~275–600 mOsm ) in the tick midgut during feeding . After characterizing the osmolarity in the tick and mammalian blood , we attempted to understand if this dynamic affected B . burgdorferi growth and physiology . Considering that most bacteria and some spirochetes ( Leptospira ) tolerate a wide range of osmolarity [22 , 24 , 36] , we were skeptical that the range of osmolarity observed in the tick midgut would have much effect on B . burgdorferi . To define the range of osmotolerance of B . burgdorferi , we monitored the growth rate in BSK-II medium at various osmolarities ( 150 to 1 , 250 mOsm ) and in different concentrations of oxygen ( Fig 2A , S1 Fig ) . As a control , we also monitored the growth rate of E . coli MG1655 in LOS medium over the same osmolarity range . Unlike E . coli , which can tolerate a range of osmolarity between 50 to 1 , 050 mOsm , B . burgdorferi was only able to grow between 250 and 650 mOsm in microaerobic or anaerobic conditions , with an optimal growth rate between 250 and 550 mOsm . At osmolalities <200 mOsm and >750 mOsm , most of the cells lysed . Transferring “survivors” to fresh BSK-II media ( 450 mOsm ) indicated that these cells could not recover . Under aerobic conditions , B . burgdorferi was more sensitive to the osmolarity ( Fig 2A ) . These data indicated that B . burgdorferi could tolerate a relatively narrow range of osmolarity ( 250 and 650 mOsm ) . However , considering the range of osmolarities in the feeding tick , it seems that B . burgdorferi is well adapted to survive in the tick midgut environment . Because of these results , all subsequent experiments were done between 250 and 650 mOsm under microaerobic conditions . During evaluation of the growth rates of B . burgdorferi at different osmolarities , we observed an effect on motility in higher osmolarity . In addition , cell morphology also was affected . High and low osmolarity are known to have a global effect on the cell physiology and gene regulation in many bacteria [24] . In B . burgdorferi , cellular morphology is critical for proper motility as the cells utilize endoflagella to perform waveform motility [37] . To understand how differing osmolarities affect these aspects of B . burgdorferi physiology , we observed the morphology and motility of cells at the three physiologically relevant osmolarities: 250 , 450 and 650 mOsm using dark-field microscopy at mid-log phase of growth ( 4–5 X 107 cells/ml ) ( Fig 2B and 2C ) . At 450 mOsm , the cells displayed normal morphology , i . e . long waveform-shaped cells ( Fig 2B ) . At lower osmolarity ( 250 mOsm ) , the cells were slightly longer , with normal motility ( Fig 2C ) . At an osmolarity of 650 mOsm ( Fig 2B ) , ~80% of the cells were non-motile and ~10% had altered motility ( twitching ) ( Fig 2C ) , and were shorter ( Fig 2B ) . We confirmed by plating that non-motile cell were viable . These observations suggested that osmolarity affected both cellular morphology and motility in B . burgdorferi . The changes in cell shape could indicate an adaptation to a change in water flux . The observed effects on motility at higher osmolarity may reflect physical constraints on flagellar function or may indicate an effect on membrane potential and/or cellular energy . Because of the observed changes in osmolarity in the feeding tick midgut , we analyzed the production of key proteins involved in successful transmission at different osmolarities . The levels of virulence factors OspC , DbpA and BBA66 increased at an osmolarity of 250 mOsm while OspA increased slightly at higher osmolarity ( 650 mOsm ) ( Fig 3A , S2 Fig ) . Key regulatory proteins involved in the regulation of these virulence related proteins were also assayed . The levels of Rrp2 , Rrp1 and BosR did not change significantly at any osmolarity tested . However , RpoN and RpoS , which have been shown to regulate these and other virulence factors , increased at an osmolarity of 250 mOsm ( Fig 3A , S2 Fig ) . More importantly , we directly tested the production of virulence factors in strains B31-A3ΔrpoN and B31-A3ΔrpoS at different osmolarities ( Fig 3B and 3C , S2 Fig ) . While these mutants grew normally at all osmolarities tested compared to wild-type B31-A3 , no changes were observed in the production of OspC , DbpA or BBA66 when the mutants were grown at 250 mOsm . These data strongly suggest that the RpoN-RpoS regulatory cascade was involved in the regulation of these virulence factors at lower osmolarity . Also , immunoblots of cell lysates of B . burgdorferi grown at 250 , 450 and 650 mOsm were probed using serum from mice infected with B . burgdorferi B31-A3 by tick bite ( Fig 3D ) . Spirochetes grown at 250 mOsm , corresponding to the osmolarity measured at the midpoint in a feeding tick bloodmeal , in tick saliva or in mammalian blood , showed increased reactivity with infected serum ( Fig 3D ) . We also tested the expression of the genes encoding these proteins by qRT-PCR . At 250 mOsm , similar to osmolarity that was measured in the blood , at the mid-point of feeding and in tick saliva , expression of the sigma factors rpoN and rpoS increased 2 . 5 and 4 . 5-fold respectively ( Fig 4 ) . We measured the expression of RpoS-dependent virulence factors ( ospC , dbpA , bba66 , bb0844 ) and found that similar to the rpoN-rpoS expression pattern , the expression of these four genes increased significantly ( 6 . 9 , 5 . 9 , 5 . 7 and 6 . 8-fold , respectively ) at 250 mOsm ( Fig 4 ) . Although both rpoN and rpoS showed changes in gene expression in response to lower osmolarity , the sigma factor , rpoD , did not change in response to osmolarity ( Fig 4 ) . Regulation of rpoS and RpoS is transcriptional , translational and post-translational [7 , 8 , 38–41] . Expression analysis of the Borrelia oxidative stress regulator ( BosR ) , which is thought to directly regulate rpoS , indicated that there was no change in transcription or translation of bosR in response to changes in osmolarity ( Figs 3 and 4 ) . Additionally , the transcription and translation of rrp2 , which is required to activate the RpoN-RpoS cascade , was not affected by osmolarity . Since both BosR and Rrp2 are transcriptional activators , their regulatory effects on RpoN and RpoS might only require activation of these proteins ( e . g . , oxidation of BosR , phosphorylation of Rrp2 ) rather than an increase in the transcription or translation of the genes encoding them . Taken together , these data indicate that lower osmolarity could trigger an increase in the expression of key virulence factors in actively growing cells and this increase was directly linked to the RpoN-RpoS regulatory cascade . B . burgdorferi cells colonizing ticks are exposed to a distinct range of osmolarities during the tick lifecycle ( Fig 1C ) . Previous studies show that the Hk1-Rrp1 TCS is required for tick midgut colonization . hk1 or rrp1 mutants are unable to be acquired by ticks fed on infected mice or introduced by artificial feeding [2] . Because TCSs are known to be involved in osmoregulation ( e . g . , OmpR ) [42] , we investigated the possibility that Rrp1 might be involved in the adaptation of B . burgdorferi to tick midgut osmolarities . To test our hypothesis , we monitored the growth rate of B . burgdorferi strains 5A4 , 5A4 Δhk1 and 5A4 Δrrp1 at 250 , 450 , 650 mOsm ( Fig 5A and 5B ) . The mutant strains were not affected at low osmolarity ( 250 mOsm ) but were dramatically affected at osmolarities >550 mOsm . In fact , Δrrp1 mutant cells in BSK-II media at increased osmolarity lysed completely mimicking the phenotype that has been reported for these mutant strains in ticks ( Fig 5B ) [2] . Further , hk1 expression has been shown to increase in the bacteria during acquisition by the tick from the host [2] . In this study , hk1 expression increased 5-fold as osmolarity increased from 250 to 650 mOsm ( Fig 5C ) . In contrast , rrp1 and Rrp1 expression did not change at the osmolarities tested ( Figs 3 and 4 ) . However , because Rrp1 has been shown to have diguanylate cyclase activity , we measured the levels of cyclic di-GMP ( c-di-GMP ) in cells at different osmolarities . At 650 mOsm , the intracellular levels of c-di-GMP increased from 220 nM/mg protein ( 250 mOsm ) to 1120 nM/mg protein ( Fig 5D ) . These data suggest that: i ) Hk1 and Rrp1 are required for the transition of B . burgdorferi from the mammal ( 300 mOsm ) to the initial conditions in the tick midgut at the beginning of feeding ( 600 mOsm ) , ii ) Rrp1 enzymatic activity dramatically increases at 650 mOsm , iii ) Hk1 and Rrp1 could be sensing changes in osmolarity , and iv ) c-di-GMP could act as an effective secondary messenger for the successful acquisition and long-term survival of B . burgdorferi in ticks . Changes in osmolarity affect B . burgdorferi morphology , motility and virulence factor expression . We next sought to characterize factors demonstrated to aid in osmoadaptation in other bacteria . Osmoadaptation involves both the efflux and influx of osmolytes , as well as ions . Among all of the characterized osmolyte transporters in E . coli or B . subtilis , only the ProU system is found in the B . burgdorferi genome [43] . This system is an ATP-dependent transporter for glycine betaine , proline , and/or choline [44–46] . The ProU locus consists of proV ( ATP-binding subunit ) , proW ( integral membrane protein ) , and proX ( periplasmic glycine betaine binding protein ) and the genes are found in that order in the B . burgdorferi genome [43] . In other bacteria , the ProU system protects bacterial cells from high osmolarity by scavenging glycine betaine , proline or choline from the growth media [44–46] . To determine whether the ProU system served as an osmoprotectant system in B . burgdorferi , we first analyzed the expression of proV , proW , proX at 250 , 450 and 650 mOsm ( Fig 6A ) . proV and proX showed a significant increase in expression at 250 and 650 mOsm when compared to 450 mOsm while proW did not change at any osmolarity tested . This may suggest that these genes may be transcribed from different promoters or a full length transcript may be post-transcriptionally modified . We also measured the gene expression of the proV gene during nymph feeding ( Fig 6B ) . Surprisingly , proV expression remained unchanged at different points of nymph feeding . To determine if the ProU system played a role in the osmotolerence of B . burgdorferi , we inactivated the ProU locus by deleting proX and tested the pro mutant for survival at different osmolarities . We attempted to delete the entire locus ( proX , proW and proV ) but we were unable to do so , probably because choline is used to synthesize phosphatidylcholine ( a major phospholipid in B . burgdorferi [47] ) . We grew B31-A3 and B31-A3ΔproX at various osmolarities and the proX mutant showed a narrower range of osmotolerance than the wild-type ( Fig 6C ) . Because of the effect of the proX mutation on growth , we tested the effect of proX inactivation on virulence and strain B31-A3proX was fully virulent in mice ( S1 Table ) . While the growth rates were slower in this mutant than in the wild-type B31-A3 or the complemented strain B31-A3proX pSABG1 , B31-A3proX cells , at an osmolarity of 300 mOsm ( Fig 6C , denoted by the arrow ) , were motile and showed an increase in the expression of OspC that was characteristic of wild-type cells at lower osmolarity ( Fig 6D ) . Overall , these data suggest that the ProU locus facilitates osmoadaptation but over a very narrow range of osmolarities and glycine betaine , proline or choline do not expand the osmotolerance of B . burgdorferi cells . It seems remarkable that B . burgdorferi is so very well adapted to living within a narrow range of osmolarities that directly reflects its immediate environment during the infective cycle . L-glutamate has been described to be involved in osmoadaptation in bacteria and is readily available in mammalian blood [15 , 22 , 24 , 27 , 33 , 48] . Classically , bacteria , such as E . coli , increase the synthesis of L-glutamate to promote growth at high osmolarity ( ~1000 mOsm ) [26 , 48 , 49] . Because B . burgdoferi is unable to synthetize L-glutamate [43 , 47] , we searched for L-glutamate uptake systems in the B . burgdorferi genome and identified two putative transporters for L-glutamate: bb0729 ( gltP ) and bb0401 . Gene expression analyses revealed that only gltP , not bb0401 , was differentially regulated in response to changes in osmolarity , increasing 3-fold at 250 mOsm ( Fig 7A ) . We also measured gltP expression before , during and after feeding in nymphs . The expression of gltP increased 5 . 2-fold during feeding at a scutal index of ~4 and decreased 2-fold in replete ticks compared to unfed ticks ( Fig 7B ) . These data suggested that glutamate might function as an osmoprotectant at lower osmolarity . Because these data suggest a role for L-glutamate as an osmoprotective molecule , we tested this more directly . An insertion inactivation mutant ( B31-5A18gltP ) was obtained from the transposon mutant library [50] and this mutant was tested for growth and survival at different osmolarities . As expected from the expression ( Fig 7A ) , strain B31-5A18gltP had a slower growth rate at 300 mOsm ( blood osmolarity ) than the wild-type strain B31-5A18 or the complemented strain B31-5A18gltP pSABG2 ( Fig 7C , denoted by the arrow ) . As was observed in B31-A3proX , B31-5A18gltP cells , at 300 mOsm , showed normal motility and increased expression of OspC ( Fig 7D ) . The effect of gltP inactivation on virulence was tested and , as with B31-A3proX , B31-5A18gltP was fully virulent in mice ( S1 Table ) . These data showed that: i ) gltP expression responded to low osmolarity both in vitro and in vivo , ii ) exogenous L-glutamate played a role in osmoprotection at low osmolarity , and iii ) L-glutamate transport does not affect survival in mice . Currently , we are trying to test the role of osmoprotectants , such as glutamate , glycine betaine and proline , in ticks by ( i ) measuring the levels of these molecules in the tick bloodmeal , ( ii ) generating a B31-A3ΔgltP-ΔproX double mutant , and ( iii ) testing all mutants in mice and ticks . To investigate the role of ion transport in osmotolerance , we analyzed the gene expression profiles of ion transport systems identified in the genome of B . burgdorferi [43] ( Fig 8A ) . These included the ktrAB transport system ( potassium uptake ) , the K+/Na+/Ca2+ transport system ( bb0164 ) , the three Na+/H+ antiporter systems ( bb0447 and nhaC-1 , nhaC-2 ) and the Mg2+ uptake system ( mgtE , bb0380 ) . The expression of both nhaC-1 and nhaC-2 increased 10-fold at 250 mOsm osmolarity , suggesting an import of H+ and export of Na+ was involved in osmoadaptation ( Fig 8A ) . Expression of the K+/Na+/Ca2+ antiporter system increased 3-fold , suggesting an adaptive flux of K+ , Na+ and/or Ca2+ ( Fig 8A ) . Furthermore , the expression of the ktrAB system increased 4 . 6-fold suggesting that the flux of K+ could augment osmoadaption ( Fig 8A ) . mgtE expression was not affected by the changes in osmolarity which was expected since magnesium has never been shown to have a role in osmotolerence ( Fig 8A ) . Taken together , the gene expression data suggest that the flux of ions would promote survival at low osmolarity . To confirm that the in vitro analysis was consistent with observed in vivo expression , we analyzed the gene expression of each of the previously mentioned transporters during nymph feeding ( Fig 8B ) . The three Na+/H+ antiporters ( bb0447 , nhaC-1 , nhaC-2 ) were induced during the feeding , increasing 8 . 6-fold , 2 . 1-fold and 2 . 7-fold respectively ( Fig 8B ) . The expression of bb0447 and nhaC-1 in replete ticks returned to the initial expression level observed in unfed ticks ( Fig 8A ) . Only nhaC-2 stayed at the levels of expression observed during tick feeding ( Fig 8B ) . Taken together , these data suggest that B . burgdorferi alters the expression of its ion transport systems which may allow the bacterium to adapt to changing osmotic conditions in the tick midgut during feeding and in its mammalian hosts . It is also possible that other factors such as ion availability ( e . g . sodium ) may be affecting the regulation of these transport systems .
B . burdorferi lives in two distinctly different environments: the mammalian host and the tick vector . As the bacteria shuttles back and forth between host and vector , they encounter conditions that are distinct to each setting . For example , when B . burgdorferi are colonizing a mammalian host , they must switch their surface proteins from OspC to VlsE to evade the host immune system [51 , 52] . While the exact signal to trigger this change has not been identified , it is clear that the host immune system provides selective pressure to eliminate bacterial cells that have not made the necessary antigenic changes [52] . Surviving cells colonize immune privileged sites existing in a nutrient rich environment with stable physiological parameters ( temperature , pH , oxygen , osmolarity , etc . ) Conversely , the tick midgut is the locale where B . burgdorferi faces a different set of conditions . Physiological conditions change between flattened and feeding ticks but most would hardly be considered to be extreme . For example , temperature ( 23°–34°C ) , oxygen ( mostly anaerobic to ~2–3% O2 during feeding ) , pH ( 6 . 8 in flattened or feeding ticks ) do not vary significantly while nutrients ( nutrient rich to starvation ) and reactive oxygen ( ROS ) or reactive nitrogen ( RNS ) species may be considered more variable challenges . What is remarkable is that B . burgdorferi is very well adapted to these conditions and senses minor changes in the tick “environment” to regulate expression of key virulence factors . In this study , we characterized another physiochemical parameter , osmolarity , that changed during tick feeding and may be a signal triggering the expression of essential virulence factors ( e . g . , OspC , DbpA , etc . ) . As previously described in other species and genera of ticks , osmolarity fluctuates during acquisition of a blood meal [17–20] . This seemed to be the case for I . scapularis . Midgut contents , isolated from feeding ticks , showed an interesting , triphasic shift from ~600 mOsm to ~300 mOsm returning to ~600 mOsm during the sequential stages of feeding . The physiological reasons for this shift are certainly related to ion and water flux required to balance the effects of non-diffusible or non-transportable anionic polypeptides concentrated in the bloodmeal ( Gibbs-Donnan equilibrium ) [53] . Clearly , water and ion fluctuations are required for the recycling of water and solutes necessary to generate the amounts of saliva that are required for long-term , successful feeding of I . scapularis . Interestingly , experiments on wild-type B . burdorferi at different osmolarities indicated that the cells had a narrow range of osmotolerance ( Fig 2A ) compared to E . coli . Normal doubling times were observed over a range of ~250 to ~650 mOsm under anaerobic and microaerobic conditions , mimicking the conditions observed in the bloodmeal during feeding . The initial observations of cells by dark-field microscopy at different osmolarities indicated that motility was affected as osmolarity reached 650 mOsm . This was of particular interest because Dunham-Ems et al . reported that B . burgdorferi cells have two phases of motility in the midgut of ticks during feeding [37 , 54] . Cells were observed to have normal motility and evenly distributed throughout the bloodmeal or were nonmotile and clumped associating with the interior face of the midgut lining . Our observations of the motility of B . burgdorferi suggest that increased osmolarity may be partially responsible for altered motility observed in feeding ticks [37] . Other interesting trends occurred in B . burgdorferi cells at physiologically relevant osmolarities . Immunoblots of protein isolated from cells grown at low osmolarity , indicated that the cells increased the expression of virulence related proteins such as OspC , DbpA and BBA66 ( Fig 3A ) . It has been shown that these proteins are required for the successful transmission and survival of B . burgdorferi in mammalian hosts [6 , 7 , 14 , 55–60] . Analysis by qRT-PCR of RNA isolated from cells grown at 250 mOsm showed an increase in the transcription of ospC , dbpA and bba66 correlating with the increase in expression of these proteins in immunoblots . An increase in the expression of rpoN and rpoS were observed at low osmolarity . Additionally , immunoblot analysis of B31-A3ΔrpoN and B31-A3ΔrpoS indicated that OspC , DbpA and BBA66 were not induced in these mutants at low osmolarity . Since it has been shown that OspC , DbpA and other virulence factors are controlled by the Rrp2-RpoN-RpoS regulatory cascade , it seems very likely that low osmolarity is directly affecting this regulatory network . It is interesting to note that the increased expression of important virulence factors at low osmolarity corresponds to the osmolarity measured at the midpoint of feeding ( Fig 1B ) . It has been shown that transmission of B . burgdorferi occurs ~2 days after the initiation of the feeding , which correlates with the drop in osmolarity measured in the bloodmeal of B . burgdorferi infected ticks . Additionally , these changes were observed in actively growing ( mid-log phase ) , motile cells . It is interesting to speculate that a drop in osmolarity could also serve as a signal to trigger the migration of B . burgdorferi from the midgut to the hemolymph and ultimately to the salivary glands during feeding . However , at this time , we do not have any direct experimental evidence supporting this hypothesis . High osmolarity ( 650 mOsm ) occurs in the midgut of an unfed tick , at the initiation of feeding and after feeding is complete . Except for a slight increase in the expression of OspA , the expression of other virulence factors remained unchanged at high osmolarity ( 650 mOsm ) compared to cells grown in BSK-II ( 450 mOsm ) ( Fig 3 ) . However , high osmolarity not only affected motility but also had another very interesting effect on B . burgdorferi . Analysis of a B31-5A4Δrrp1 mutant indicated that this strain was exquisitely sensitive to osmolarities >500 mOsm compared to strain B31-5A4 and cells rapidly lysed after less than 4h of exposure to increased osmolarity . Rrp1 is the response regulator in the Hk1-Rrp1 TCS and functions as a di-guanylate cyclase [2–4 , 61] . C-di-GMP acts as a secondary messenger for signal transduction in bacteria and the levels of c-di-GMP increased dramatically at high osmolarity ( Fig 5D ) . Rrp1 has also been shown to be required for tick colonization , motility and the regulation of genes involved in glycerol metabolism [2 , 3 , 62] . In addition to its regulatory functions , Caimano et al . showed that B31-5A4Δrrp1 was virulent in mice but this mutant rapidly lysed after being acquired by ticks fed on mice infected with this strain [2] . Also , B31-5A4Δrrp1 rapidly lysed when introduced into ticks by artificial feeding . Collectively , these data suggest that at high osmolarity , Rrp1: i ) was required for survival; ii ) had increased diguanylate cyclase activity; iii ) is required for tick colonization; and iv ) could putatively regulate B . burgdorferi motility . Lastly , we investigated the osmoadaptation of B . burgdorferi . In bacteria , the response to changes in external osmolarity happens at two levels . To restore a conductive intracellular environment , cells transport ionic solutes like K+ , Na+ and compatible solutes glutamate , proline and glycine betaine [22 , 23] . At low osmolarity , ionic solutes ( primarily K+ ) accumulate while at high osmolarity , compatible solutes accrue to support a high intracellular osmotic pressure without the deleterious effects that ionic solutes have on the activity of metabolic and biosynthetic enzymes [24] . When compatible solutes are not available in the extracellular milieu , the cells will increase their intracellular concentrations by accelerating the synthesis of these important osmoprotectants . Together , these osmoadaptive systems allow bacteria like E . coli and Salmonella typhimurium to tolerate osmolarities from 50–1400 mOsm . Unlike E . coli or other spirochetes like Treponema denticola and L . interrogans [22–24 , 36] , the B . burgdorferi genome does not harbor the genes encoding proteins to synthesize osmolytes ( e . g . , proline , choline or glutamate ) . However , the genome does have three putative osmolyte transport systems: the proU system for the transport of glycine betaine , proline or choline , as well as bb0729 ( gltP ) and bb0401 both of which are annotated as glutamate transporters . Transcription of the proU system increased at low and high osmolarity in vitro suggesting that this transport system might be involved in osmoprotection . Additionally , a B31-A3proX mutant strain showed a narrower range of osmotolerance than wild-type B31–A3 . However , choline , proline and glycine betaine did not increase the range of osmotolerance of B31-A3 . These data indicate that these compatible solutes are required for the survival of B . burgdorferi within the narrow range of osmolarities encountered in the bloodmeal of feeding ticks . The results for glutamate are distinctly different from what was expected based on previously published information on the role of glutamate in protecting E . coli and S . typhimurium at high osmolarity [48] . As previously mentioned , compatible solutes ( e . g . , glutamate , proline ) protect cells at high osmolarity while ionic solutes ( e . g . , K+ ) protect cells at low osmolarity [22 , 27 , 32 , 34] . The mutant strain B31-5A18gltP was more sensitive to low osmolarity while high osmolarity had no effect on the growth and survival of this mutant compared to B31-5A18 . Predictably , the expression of the genes encoding ionic solute transport systems such as ktrAB ( K+ transport ) , bb0164 ( K+/Na+/Ca2+ ) , bb0447 , nhaC-1 and nhaC-2 increased at low osmolarity ( 250 mOsm ) . Currently we do not understand why a compatible solute like glutamate is required as an osmoprotectant for B . burgdorferi at low osmolarity but we suspect that it plays a role in the accumulation of ionic solutes in cells as they respond and adapt to low osmolarity . This may be an important function since it has been shown that >70% of the K+ is cycled into the hemolymph and saliva during the feeding of I . ricinus and D . andersonii [17–21] . Clearly , the inability of B . burgdorferi to synthesize compatible solutes has narrowed the limits of their osmotolerance but , despite this , they are finely adapted to the narrow range of osmolarities that they encounter in the tick bloodmeal/midgut and the mammalian host . While B . burgdorferi cells are well adapted to a narrow range of osmolarity , what was remarkable was that they were using these relatively small changes in osmolarity as a signal to affect at least two regulatory pathways . First , high osmolarity ( 650 mOsm ) has a dramatic effect on motility in wild-type B31-A3 , B31-5A18 and B31-5A4 . In addition , strains B31-5A4Δhk1 and B31-5A4Δrrp1 did not survive at osmolarities above 500 mOsm . As important , the levels of the secondary messenger molecule , c-di-GMP , increased dramatically at high osmolarity , most likely due to an increase in the diguanylate cyclase activity of Rrp1 [63] . These data suggest a role for Hk1 and Rrp1 in the adaptation to and survival of B . burgdorferi cells at osmolarities of 600 to 650 mOsm . Second , analyses of protein and gene expression in B31-A3 , B31-A3ΔrpoN and B31–A3ΔrpoS suggested that B . burgdorferi cells express key virulence factors , such as OspC , DbpA and BBA66 at low osmolarity and this increase in expression was dependent on RpoN and RpoS . Our current working model ( Fig 9 ) is that as B . burgdorferi cells are acquired by feeding ticks , they rapidly transition from osmolarities of ~300 mOsm in mammalian blood and tissue to ~600 mOsm in the tick midgut at the beginning of feeding . It seems very likely that Hk1-Rrp1 and c-di-GMP are essential for this transition . At the end of acquisition , in replete ticks , the osmolarity returns to ~600 mOsm and motility is impaired , potentially limiting spread of B . burgdorferi and trapping them in the midgut . Long-term survival of cells through the molt is most likely mediated by RelBbu ( RelA/SpoT homolog ) [5] . At the midpoint of the second feeding , as the osmolarity cycles from ~600 to ~250 , motility increases and the cells respond to lower osmolarity by expressing ionic solute transport systems . Most importantly , the RpoN-RpoS regulatory cascade is also stimulated by low osmolarity and triggers the expression of vertebrate virulence-related proteins . At this point , the cells are actively growing , have normal motility and are expressing proteins necessary to promote successful transmission to the next mammalian host . It seems clear that changing osmolarity can affect two different regulatory pathways , Hk1-Rrp1 and RpoN-RpoS , and is potentially a major signal sensed by B . burgdorferi during acquisition and transmission .
The strains used in this study are described in the S2 Table . B . burgdorferi strains were grown in BSK-II medium , pH 6 . 8 at 34°C [64] under microaerobic environment ( 5% O2 , 5% CO2 ) and , when indicated , under anaerobic ( 5% CO2 , 5% H2 , balance N2 ) or aerobic condition . Cell densities were determined by dark-field microscopy ( Eclipse E600 , Nikon , Melville , NY ) . The osmolarity of the BSK-II medium is 450 mOsm . To obtain high-osmolarity medium , NaCl was added . To obtain low-osmolarity medium , ddH2O was added . BSK-II medium for plating contained 0 . 6% agarose . Importantly , low-osmolarity BSK-II media was tested to ensure that essential nutrients were not too dilute to support normal growth . This was accomplished by adding NaCl to the dilute BSK-II to adjust the osmolarity to 450 mOsm . Restoring the osmolarity of diluted BSK-II to 450 mOsm restored normal growth of wild-type B31-A3 [65] . For survival assays , various wild-type and mutant strains were grown in BSK-II medium at different osmolarities starting at 1 x105 cells/ml to early stationary phase of growth . Every 24 h an aliquot of each culture was examined by dark-field microscopy and plated on BSK-II . Plates were incubated at 34°C under microaerobic conditions for 7–14 days to allow enumeration of CFU . The cell length ( 40 cells per slide , 5 slides from 5 independent cultures ) was measured using ImageJ software . E . coli strains were grown in Lysogeny broth [66] or in low osmolarity medium , called LOS ( 4 g of casein hydrolysate , 0 . 5 mg of FeSO4 , 18 mg of MgCl2 , 200 mg of ( NH4 ) 2SO4 and 175 mg of K2HPO4 per liter , pH 7 . 2 ) [67] . The LOS medium osmolarity is 70 mOsm . To obtain high-osmolarity medium , NaCl was added . Growth rates were defined during the exponential phase [68] . Briefly , the growth rate is defined by 1/doubling time and expressed in 1/h . All osmolarities were measured with a Wescor vapor pressure osmometer at 21°C ( model 5500 , Wescor , Inc . , Logan UT , USA ) and expressed as milli-osmolar ( mOsm ) . The proX::himar1-Gm was amplified by PCR from B31-5A18 NP1 proX::himar1-Gm ( proUF ACAGATGAGGTTGTAGCAGCA and proUR GCATATACAAACCTACCTGCTC ) and cloned into TopoZeroBlunt ( Invitrogen , Carlsbad , CA ) to obtain Topo0ProX::Gm vector . The resulting plasmid was transformed into low-passage B . burgdorferi B31-A3 strain as described previously [69] and gentamicin-resistant colonies were analyzed by PCR to confirm the inactivation of proX . Mutants were screened using plasmid specific primer sets [25] . Mutant strain B31-A3proX harbored all plasmids except cp9 was used for further characterization . For the complementation , the proU operon was amplified by PCR using proUF and proUR primers and cloned into PCR-XL-TOPO following the manufacturer’s recommendations ( Invitrogen , Carlsbad , CA ) . The resulting plasmid was digested with SacI-PstI and the proU fragment was cloned into the pKFSS1 [70] shuttle vector digested with the same restriction enzyme to obtain pSABG1 . The gltP gene was synthesized by Genscript , USA and cloned into the pKFSS1 shuttle vector digested with SacI-PstI to obtain pSABG2 . The resulting plasmids were transformed into low-passage B . burgdorferi mutants strains as described previously [69] and spectinomycin-resistant colonies were analyzed by PCR to confirm the construction . RNA samples were extracted from B . burgdorferi cultures using the RNeasy mini kit ( Qiagen , Valencia , CA ) according to the manufacturer’s protocol . Three independent cultures were used for each osmolarity . Total RNA from ticks was isolated from 3 pools of 7 nymphs fed on mice infected by needle inoculation with B31-A3 . RNA samples was extracted using RNeasy mini kit ( Qiagen , Valencia , CA ) . Ticks were frozen at -80°C directly and crushed . TRIzol ( Life technologies , Carlsbad , CA ) was added with chloroform . After centrifugation , the upper phase was mixed with ethanol 70% ( 1:1 ) and loaded onto the provided Qiagen column according to the manufacturer’s instructions . Digestion of the genomic DNA was performed using TURBO DNA-free DNase I ( Life Technologies , Carlsbad , CA ) . The cDNA was synthesized using the Superscript III reverse transcriptase with random primers ( Invitrogen , Carlsbad , CA ) . To determine gene expression levels , a relative quantification method was employed using the enoS gene as a reference gene ( S3 Fig ) . All samples were performed in at least three biological replicates and three technical replicates on a Roche LightCycler 480 System using Green PCR Master Mix ( Life technologies , Carlsbad , CA ) . All primers used for the study are listed in S3 Table . To determine relative gene expression , the LightCycler 480 software version 1 . 5 was used . The relative quantification was performed following the E-Method using the enoS as a housekeeping gene [71] . For analysis of cell lysates by Western-blot , bacteria were grown to mid-log phase at 34°C in microaerobic conditions . The cells were harvested by centrifugation , washed twice in HN buffer ( 50 mM HEPES pH7 . 5 , 50 mM NaCl ) , resuspended in 0 . 25M Tris-HCl pH 6 . 8 and lysed by sonication . The protein concentration was determined with Take3 micro-volume plate in a Synergy 2 Multi-Mode plate reader ( BioTek Instruments , Winooski , VT , USA ) . 40 μg of protein was loaded in a 4–20% pre-cast SDS-PAGE gel ( Bio-Rad , Hercules CA , USA ) and transferred to a nitrocellulose membrane using a Trans-Blot TurboTM blotting system ( Bio-Rad , Hercules CA , USA ) with a pre-programmed protocol ( 2 . 5A , up to 25V , 3 min ) . Western blotting was performed using standard protocols , i . e . membrane blocking 1 h in 5% nonfat milk in PBS-T ( 0 . 1% Tween 20 ) , then incubating 1hr in PBS-T with primary antibodies , washing in PBS-T and then incubating 30 min in PBS-T with Rec Protein A-HRP ( 1:4 , 000; Life technologies , Carlsbad , CA , USA ) or with the anti-IgY conjugated to HRP ( 1:50 , 000; for α-BBA66 , Aves Laboratories , Tigard , OR , USA ) . For the primary antibodies , the following dilutions were used: α-OspC 1:1 , 000 [72] , α-DbpA purified antibody 1:1 , 000 ( Rockland Immunochemicals , Gilbertsville , PA , USA ) , α-BBA66 1:4 , 000 [73] , α-RpoS 1:500 [74] , α-RpoN 1:1 , 000 , α-BosR 1:500 , α-OspA 1:2000 ( Rockland Immunochemicals , Gilbertsville , PA , USA ) , α-Rrp1 1:1 , 000 , α-Rrp2 1:2 , 000 or infected-mouse serum 1:200 ( mice infected with wild-type B31-A3 spirochetes by tick bite ) . Blots were imaged by chemiluminescent detection using Super Signal Pico chemiluminescent substrate kit ( Thermo Scientific , Rockford , IL , USA ) . Rabbit polyclonal antisera directed against Rrp2 or BosR protein was prepared according to a previously published protocol [72] . Rabbit polyclonal antisera directed against Rrp1 protein was prepared by Rockland Immunochemicals , Gilbertsville , PA , USA . I . scapularis egg masses ( Oklahoma State University ) were allowed to hatch and mature in a controlled temperature , humidity and photoperiod environment . RML mice were needle inoculated by intradermal injection with 100 μl of BSK-II containing 1 x 105 B . burgdorferi B31-A3 and after three weeks , infection confirmed by culturing ear punch biopsies . Larval ticks were fed to repletion on infected mice ( naïve mice for non-infected cohort ) , collected and allowed to molt into nymphs and cure in a controlled environment . Nymphal ticks were then fed on naïve RML mice and mechanically removed periodically during the feeding and further processed for osmolarity measurement or RNA isolation as indicated . For infected nymphs , mice were sacrificed 3–6 weeks post inoculation and tissues ( ankle joint , bladder and ear ) were cultured to verify infection of the ticks through transmission to the naïve animal . Several of the nymphs ( infected and non-infected cohorts ) were fed to repletion , collected and allowed to molt into adults . After maturation , these ticks were fed on New Zealand White rabbits . Ticks were removed during feeding and further processed for osmolarity determination or RNA extraction as indicated . c-di-GMP was quantified from B . burgdorferi cultures using the cGMP Direct Biotrak EIA ( GE Healthcare , UK ) according to the manufacturer’s protocol . Four independent culture samples were used for each condition . Protein was quantified using a Microplate with Synergy 2 plate reader ( BioTek , VT , USA ) Scutal index in feeding ticks was determined as previously described [35] . Briefly , for nymphs , the width of the scutum and length of the body ( Fig 1A ) were measured under a dissecting microscope configured with an ocular micrometer calibrated to a stage micrometer at a given magnification . For adult ticks , a similar procedure was performed except that a hand held magnifying micrometer was used . Because the width of the scutum remains constant and the length of the body increases proportionately during tick feeding , its ratio provides the most reliable and reproducible indicator of feeding progress . In triplicate , RML mice were inoculated intradermally with 1 x 105 cells in 100 μl BSK-ll with B . burgdorferi strains B31-A3 , B31-5A18 , B31-A3proX and B31-5A18gltP . Four weeks post-infection , the mice were sacrificed , tissues dissected ( ankle joint , bladder and ear ) and cultured in BSK-II to confirm the presence of spirochetes . Rocky Mountain Laboratories ( RML ) , NIAID , NIH in Hamilton , MT are accredited by the International Association for Assessment and Accreditation of Laboratory Animal Care . The blood meal from the midgut of fed Ixodes scapularis adults and nymphs was collected from interrupted and replete ticks using the following methods . Ticks were held behind the basis capituli with fine pointed forceps . With a second set of forceps , the abdomen was pierced and the contents extruded with slight downward pressure into a microfuge tube . Adults were collected individually and nymphs of similar scutal index were pooled to provide sufficient sample subsequent for analyses . Mouse infection studies were carried out in accordance with the Animal Welfare Act ( AWA 1990 ) , the guidelines of the National Institutes of Health , Public Health Service Policy on Humane Care ( PHS 2002 ) and Use of Laboratory Animals and the United States Institute of Laboratory Animal Resources , National Research Council , Guide for the Care and Use of Laboratory Animals . All animal work was done according to protocols approved by the Rocky Mountain Laboratories , NIAID , NIH Animal Care and Use Committee ( Protocol Number 2014–021 ) . The Rocky Mountain Laboratories are accredited by the International Association for Assessment and Accreditation of Laboratory Animal Care ( AAALAC ) . All efforts were made to minimize animal suffering . Prism 6 software ( v6 . 00 , GraphPad , San Diego , CA ) was used for all statistical analyses . The data were analyzed using an unpaired t test . P<0 . 05 was considered significant .
|
Borrelia burgdorferi , the Lyme disease agent , exploits a multifaceted enzootic cycle that requires a tick vector for successful transmission between mammalian hosts . Two different regulatory systems control genes that are required to complete this infective cycle . The Hk1/Rrp1 two-component system affects genes required for successful transfer between mammal and tick vector while the Rrp2-RpoN-RpoS regulatory cascade modulates genes essential for the transmission from the tick to a new vertebrate host . Data presented in this study indicate that fluctuations in osmolarity in the tick midgut directly affect these two regulatory pathways . Osmolarity in the lumen of the tick adjusts to the osmolarity of the incoming blood ( blood meal ) to promote water and ion flux into tick tissues . A positive water flux is essential to generate sufficient saliva for prolonged feeding . We propose that B . burgdorferi uses this physiological parameter as an important signal to adapt and regulate genes required for survival in the tick ( through Hk1/Rrp1 ) and transmission to a new host ( through Rrp2-RpoN-RpoS ) .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"motility",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"gene",
"regulation",
"pathogens",
"microbiology",
"developmental",
"biology",
"nymphs",
"bacteria",
"bacterial",
"pathogens",
"borrelia",
"burgdorferi",
"medical",
"microbiology",
"gene",
"expression",
"microbial",
"pathogens",
"borrelia",
"hematology",
"pathogen",
"motility",
"blood",
"anatomy",
"cell",
"biology",
"virulence",
"factors",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"metamorphosis",
"organisms"
] |
2016
|
Two Different Virulence-Related Regulatory Pathways in Borrelia burgdorferi Are Directly Affected by Osmotic Fluxes in the Blood Meal of Feeding Ixodes Ticks
|
We observe and quantify wave-like characteristics of amoeboid migration . Using the amoeba Dictyostelium discoideum , a model system for the study of chemotaxis , we demonstrate that cell shape changes in a wave-like manner . Cells have regions of high boundary curvature that propagate from the leading edge toward the back , usually along alternating sides of the cell . Curvature waves are easily seen in cells that do not adhere to a surface , such as cells that are electrostatically repelled from surfaces or cells that extend over the edge of micro-fabricated cliffs . Without surface contact , curvature waves travel from the leading edge to the back of a cell at ∼35 µm/min . Non-adherent myosin II null cells do not exhibit these curvature waves . At the leading edge of adherent cells , curvature waves are associated with protrusive activity . Like regions of high curvature , protrusive activity travels along the boundary in a wave-like manner . Upon contact with a surface , the protrusions stop moving relative to the surface , and the boundary shape thus reflects the history of protrusive motion . The wave-like character of protrusions provides a plausible mechanism for the zig-zagging of pseudopods and for the ability of cells both to swim in viscous fluids and to navigate complex three dimensional topography .
During chemotaxis , a chemical gradient directs cell migration . Chemotaxis occurs in phenomena as diverse as wound healing [1] and embryonic development [2] , and has also been implicated in a wide array of pathological processes including aberrant angiogenesis [3] and cancer metastasis [4] . Chemotaxing cells can migrate individually during immune responses and neuronal patterning , or in cooperative groups during embryogenesis , wound healing , and organ and vasculature formation . The social amoeba Dictyostelium discoideum is a model system for directed cell migration , and has been used to elucidate the regulatory processes of chemotaxis . Dictyostelium chemotaxis is of comparable speed to neutrophil chemotaxis and involves similar regulatory processes [5] . The chemotaxis of Dictyostelium is also more than an order of magnitude faster than typical epithelial cell migration and does not involve mature focal adhesion complexes [6] . Dictyostelium cells migrate directionally in gradients of cyclic adenosine monophosphate ( cAMP ) . The binding of cAMP to specific cell surface receptors leads to the activation of various effectors , including adenylyl cyclase , resulting in the production and secretion of additional cAMP and the relay of the signal to neighboring cells [7] . This signal relay guides the cells to migrate collectively in a head-to-tail fashion toward aggregation centers and to maintain a preferred migration direction over distances much larger than the characteristic diffusion lengths of molecules [8] . Cells with the aca− mutation do not produce cAMP and , in its absence , are mostly round and immotile [9] . However , when stimulated with exogenous cAMP , aca− cells polarize and begin to migrate . If stimulated with a uniform concentration of cAMP , aca− cells exhibit chemokinesis ( chemically induced random migration ) . In response to a cAMP gradient , these cells chemotax effectively [9] . The regulatory processes behind Dictyostelium chemotaxis have been investigated in great depth . Recent discoveries demonstrate the complexity of key chemotactic signaling pathways , including multiple compensatory mechanisms for sensing the direction of cAMP gradients [10] . Models of cell movement often consider migration as the final step in this process and use the output of the directional sensing machinery to assign a probability of forming a protrusion ( and subsequently migrating ) in a particular direction [11] . In most models , the location at which a new protrusion forms determines the direction of the protrusion and ultimately the direction of migration of the cell . For example , Haastert and Bosgraaf found that pseudopods grow in a direction perpendicular to the local boundary [12] . Directional migration in response to chemotactic signals is considered to be due to either the formation of new protrusions ( the compass model ) or the biased bifurcation of existing protrusions ( the bifurcation and bias model ) [13] . Quantitative studies of cell shape and motion that follow protrusions are now emerging , and indicate that protrusion dynamics are richer than anticipated [12] , [14] , [15] , [16] , [17] , [18] . Machacek and Danuser , tracking a fixed region of the boundary of slow epithelial cells with sub-pixel resolution , found three distinct states of local protrusion activity: local protrusions that grow and retract in a single location along the cell edge , waves that travel along the edge of the cell , and fast , large-scale protrusions [14] . Dobereiner et al . found that wave-like protrusion dynamics are common to a wide range of slow-moving cells [19] , especially during cell spreading . How are such complex protrusion dynamics relevant to chemotaxis and directed cell migration ? For the models of chemotaxis reviewed above , and for fast migration , a local protrusion phenotype is usually assumed in which protrusions are a ( noisy ) output in response to a chemotactic signal . This view appears to be supported by experiments in which a strong chemotactic signal is placed in close proximity to individual cells , causing a pseudopod to form in the direction of the signal [7] . However , recent experiments and modeling efforts point to a more complex protrusive machinery . Indeed , given signal strengths at physiological levels ( and even in the absence of chemotactic gradients ) , protrusions tend to form on alternating sides of the leading edge of the cell , resulting in a zig-zag migration pattern [17] . Such zig-zag motion indicates that protrusion locations are not just described by a noisy output based on a chemical signal , but are also influenced by the prior protrusion history . The alternating position of pseudopods can be explained , for example , by considering the protrusive machinery as an excitable system [20] . The conclusion that alternating pseudopods are prominent in fast migrating cells relies on thresholds to separate individual pseudopods in a consistent way . However , it is unclear whether the underlying biology of protrusions justifies such thresholding . Instead , zig-zagging and alternating of pseudopods may be the result of wave-like behavior of the protrusive machinery . In developing tissues , actin waves can be seen to propagate across groups of cells [21] . There is also direct evidence that wave-like intracellular actin polymerization processes are prominent in fast migrating cells . Some of the first indications of this effect were from Vicker and colleagues , who found that Dictyostelium pseudopod dynamics are not random [22] , but rather are associated with actin filament polymerization waves [23] that can drive locomotion [24] . Others were able to observe wave-like dynamics directly in actin polymerization and depolymerization [25] . The connection between internal waves and forces has also been elucidated [26] . Finally , the interaction between waves and surfaces is key to understanding migration: For example , Weiner found that when a neutrophil runs into another cell or an obstacle , actin waves are extinguished at the interaction site , allowing cells to change direction and avoid the obstacle [27] . This observation indicates that waves may not be observed when cells are confined between boundaries , such as e . g . in a key paper that reported zig-zagging of Dictyostelium protrusions [28] . To study the character of protrusions during fast cell migration and chemotaxis , we present new methods for the quantification of the dynamic shape of migrating cells . Using these methods , we demonstrate that protrusions in Dictyostelium have a wave-like character . As found in neutrophils , the waves appear to stop moving when touching the surface , so the wave-like character of protrusions is most evident when cells do not touch the surface , or extend over the edge of a cliff . While our study highlights that wave-like dynamics exist under many different conditions , we do not directly measure motion of the actin cytoskeleton or membrane , and thus have no conclusive evidence whether the wave-like dynamics reflect reaction-diffusion waves ( due to actin polymerization or myosin contraction dynamics ) or transport of membrane or intracellular material . These results could lead to a model that relates protrusive waves to the zig-zag-like appearance of cell tracks and the directional persistence of cell motion during chemotaxis .
To study the changing shape of migrating Dictyostelium cells , we used a snake ( also known as active contour ) algorithm . We analyzed the shapes of starved , wild-type ( WT ) cells that were self-aggregating . The extracted shape of one such migrating cell is shown in figure 1a . The boundary color represents curvature , a measure of local shape . As expected , the front and back of the cell have high curvature . Additional peaks in boundary curvature indicate other local protrusions . We find that additional protrusions first become visible near the front of the cell and , with respect to the cell , propagate towards the back as the cell migrates . At the sides of cells , these curvature bumps appear stationary relative to the surface ( see figure 1b for a representative image sequence ) . To study the motion and evolution of these propagating curvature peaks , we labeled the boundary in order to follow individual boundary points as they move . We defined a 1∶1 mapping between the boundary points in successive frames , choosing the mapping that minimized the total boundary point displacement . To create such a 1∶1 mapping , we described the boundary with the same number of points in every frame ( 400 ) , even though the length of the perimeter of the cell varies from frame to frame . With the boundary points labeled in this way , we are able to visualize and analyze how local properties of the boundary , such as curvature , vary as a function of both space and time . Figure 1c is a kymograph that depicts how curvature evolves along the boundary as a function of time . The front and back of the cell each appear as thick stripes of high curvature and the additional , lateral peaks appear as thin lines that connect the front to the back . Since the thin lines representing the curvature peaks are approximately parallel , the curvature peaks travel at equal speed with respect to the cell . The boundary curvature kymograph of an additional , developed , WT cell is shown in figure S1a ( video S3 ) . Averaging across 26 cells , we find that the peaks move relative to the cell at 10 . 9±0 . 7 µm/min , comparable to the average migration speed of Dictyostelium cells . This finding is consistent with the observation that peaks of high curvature on the side of the cell are stationary with respect to the surface . We have previously found that alternative tracking mappings yield similar results , but generally do not track the entire local boundary [29] . For instance , kymographs can be constructed with boundary points that are a constant distance from neighboring boundary points , by using the boundary point that is closest to 0° relative to the center of the cell to align the boundaries from frame to frame . Such a mapping shows similar curvature peak patterns , but these kymographs are more difficult to interpret since only part of the local boundary is tracked from frame to frame [29] . Figure 1c shows that most high curvature regions on the side of cells start at the leading edge , and thus that the shape of the cell reflects a history of activity at the leading edge . Indeed we find a connection between curvature and protrusions at the leading edge , as described below . The amplitude of each peak varies as it travels from front to back , but tends to diminish with time , a further indication that the associated local protrusions are passive or shrinking . The local protrusions appear on roughly alternating sides , resulting in a loose right-left-right-left pattern that may reflect the zig-zag pattern of pseudopod activity . The pattern is not without defects , as multiple protrusions can travel simultaneously along the same side of the cell . Migration requires not only pseudopods but also adhesion to a surface . Thus , surface contact can indicate whether bumps on the side of the cell reflect pseudopods that have successfully adhered ( and thus can contribute to motion ) or unsuccessful pseudopods that failed to adhere . We imaged fluorescently-labeled , developed , WT cells , while simultaneously using internal reflection microscopy ( IRM ) to image the region of cell-surface contact . We extracted both the boundary of the entire cell and the boundary of the surface contact region ( s ) from the images . Figure 2a shows , for a representative cell , an IRM image sequence overlaid with the boundary of the surface contact region and the boundary of the entire cell . In this image sequence , the cell extends a protrusion that is not in contact with the surface , the protrusion makes contact with the surface , and then the area of surface contact under the protrusion grows . Our data reveal that the local protrusions at the side of the cell correspond to regions of enhanced cell-surface contact , and thus likely reflect successful pseudopods . We note that some protrusions retract quickly and never contact the surface . For example , in the movie associated with this image sequence ( video S4 ) , 4 out of 14 large protrusions retract before they contact the surface . Protrusions that do contact the surface are rarely retracted , indicating that contact stabilizes the protrusions . At the sides of migrating cells , regions of high curvature are in contact with the surface and remain stationary . However , in analyzing the shapes of migrating cells it is difficult to determine if regions of high curvature are stationary at the front of cells . To analyze the behavior of high curvature regions in the absence of surface contact , we analyzed the shape of aca− cells that were electostatically repelled from the surface . Since cell membranes and glass coverslips are both negatively charged , cells do not adhere to coverslips at a low salt solution [30] . The cells remain viable – upon addition of standard buffer the cells adhere to the surface and migrate . We found that for non-adherent cells , regions of high curvature actively move in a wave-like manner from the cell front to the cell back ( video S5 , figure 2b ) . Figure 2c shows the curvature kymograph of a non-adherent cell . Here , the curvature waves move at an average speed of 36±2 µm/min ( with a standard deviation of 8 µm/min ) , which is much faster than the average cell speed in the movie of 8 µm/min ( with a standard deviation of 5 µm/min ) . In both adherent and non-adherent cells , the high curvature regions tend to alternate between the left and right sides of the cell . We also analyzed the shape of myoII− cells that were electostatically repelled from the surface ( figures 2d , e ) . Unlike aca− cells , myoII− cells do not exhibit traveling curvature waves . Instead , myoII− cells are round with localized , transient patches of protrusive activity . These protrusive patches are not confined to one region of the cell boundary , as they would be if there were a stable cell front , but rather can appear anywhere along the cell boundary . Using surfaces with-three dimensional topography , we also analyzed the shapes of cells that adhere only at their back , even in standard salt concentrations . To guide a cell to move away from the surface , we placed a point source of cAMP above and over the edge of a microfabricated ramp that terminated with a 75 µm tall cliff , such that the surface closest to the point source was the cliff edge . A schematic of our set-up is shown in figure 3a . The majority of cells that reach the edge of the cliff extend themselves over the cliff edge . Many of those cells also migrate along the cliff edge to attempt to reach the needle . We never observed a cell falling off of a cliff . Instead , cells at the cliff edge extend up to 80% of their surface area over the edge toward the cAMP source . The shape dynamics at first glance look quite distinct – the cells swing back and forth quickly over the edge ( figure 3b ) . However , curvature waves are again seen in the portion of each cell that is not adhered to the cliff ( which represents the majority of the cell ) . Figure 3c shows an image sequence of one such aca− cell , in which a curvature wave propagates from the front to the back of the cell ( video S6 ) . ( This cell appeared in our previous work [31] , although the boundary curvature was not analyzed . ) Figure 3d shows the corresponding curvature kymogragh . The average curvature wave speed for the cell shown here is 29±3 µm/min ( as calculated from 9 waves over 2 . 4 minutes ) . The average curvature wave speed of cells extended over the edge of a cliff is comparable to the curvature wave speed of cells that are electrostatically repelled from their substrate . The swinging of the cell appears correlated with curvature waves hitting the surface , which provides a simple mechanism for cellular reorientation , and hence the exploration of 3D space . To explore the onset of curvature peaks at the sides of cells , we analyzed the dynamic shape of polarizing Dictyostelium . Cells are almost always round when initially placed on a surface , although WT cells quickly polarize and begin to migrate . Aca− cells , which do not produce cAMP , are more basal ( quiescent ) than WT cells . However , WT and aca− cells migrate with comparable speed and directional persistence [8] , and exhibit indistinguishable curvature peaks ( figure S2a ) . We therefore analyzed the onset of curvature peaks in aca− cells . Even prior to polarization , cells send out small , quickly retracted protrusions . Since these transient protrusions are more visible in phase-contrast images than in images of fluorescently dyed cells , we also analyzed phase-contrast movies ( figure 4 ) . The polarization of a fluorescently dyed aca− cell is shown in figure S2a ( video S8 ) . When a cell polarizes , its shape elongates . To quantify the degree of polarization , we define the non-circularity as the ratio of the cell perimeter to its area , normalized so that the non-circularity of a circle is 1 . The non-circularity , centroid velocity , and boundary curvature of a polarizing aca− cell are shown in figure 4 . While the non-circularity is initially near one , it soon begins to increase in an oscillatory fashion . The boundary curvature , on the other hand , changes pattern abruptly at the onset of motion . While the curvature prior to polarization is slightly irregular , curvature peaks ( tilted red lines in figure 4c ) appear only after polarization . Hence , polarization and migration coincide with a change in the nature of protrusions from being localized and intermittent to being more continuous and possibly wave-like . Curvature peaks are suggestive of protrusions , because a localized protrusion is necessarily associated with a localized region of high curvature . To compare boundary curvature to motion , we developed a measure of local boundary motion . We calculated the motion of each boundary point by measuring the distance to the closest boundary point in a later frame and then smoothing over the list of mapped to boundary points . Protrusive motion was defined to be positive , while retractive motion was defined to be negative . Figure 5b shows representative local motion mapping vectors colored by the value of the local motion measure , while figure 5c shows the local boundary motion kymograph of the same developed , WT cell shown in figure 2a , ( See figures S1b and S2b for additional plots of WT and aca− cells . ) From local motion kymographs , we see that cells have two regions of activity , one associated with protrusions and one with retractions . Neither the fronts nor the backs of cells move at constant speed; rather , they start and stop intermittently . However , while the location along the boundary of retractions shifts little from retraction to retraction , protrusions tend to zig-zag . We can compare boundary curvature to boundary motion by comparing the curvature and local motion kymographs . Curvature peaks ( shown as dashed black lines ) are overlaid on figures 5a and 5c . In general , we find that the initiation of curvature peaks tends to coincide with the growth of protrusions , at the front of the cell . ( Also , figure S4 shows how these curvature peaks relate to the distance from the cellular boundary to the footprint . ) Thus protrusive motion might travel along the boundary in the same manner as the curvature peaks . Protrusive motion has often been discretized into pseudopod extension and retraction events . Here we analyze protrusive motion both with and without discretization and show that discretization may hide the wave-like nature of the protrusive process . We first analyze boundary motion under the assumption that protrusions and retractions are discrete events . The times and locations of individual protrusions and retractions along the boundary were defined as the peaks and valleys of the local motion measure , respectively . Figure 6b shows the protrusions and retractions extracted from the local motion data shown in figure 6a . Protrusions are shown as black dots , while retractions are shown as white dots . As expected based on the visualization in figure 6a , protrusions are more spread out at the front of the cell than are retractions at the back . However , some retractions do occur at the front of the cell when protrusions are retracted . We calculated temporal and spatial statistics based on data from 26 self-aggregating WT cells , including those with developmental times 1 . 5 hours longer or shorter than our normally used time . We found a total of 2219 protrusions and 2220 retractions in these data . On average , WT cells exhibit 2 . 9±0 . 2 protrusions/minute and 2 . 9±0 . 2 retractions/minute . Boosgraaf et al . found a similar frequency of 4 protrusions/minute [17] . The statistics of the extracted protrusions and retractions are further explored in [29] . While analyzing protrusions as discrete events yields results consistent with prior work [17] , two facets of the kymograph of local protrusions and retractions in figure 6a indicate that this discretization can mask continuous or wave-like characteristics of protrusive motion . First , protrusions are not well separated from each other . Second , many protrusions appear as tilted streaks in the kymograph , indicating that lateral motion occurs during the protrusion . To analyze protrusive and retractive motion as continuous boundary movement , we define the location of greatest protrusion and retraction activity for each frame as the location of the weighted average of the protrusive or retractive motion . Figure 6c shows a representative example of extracted mean protrusion and retraction locations . We measured the mean squared displacement ( MSD ) of the average protrusion location along the boundary ( figures 6d and S3 ) . Note that this is not the MSD of centroid motion , but the MSD of protrusive motion along the boundary of the cell . We find that on short time scales , the protrusive motion along the boundary is nearly ballistic . As the time scale increases , protrusive motion becomes caged to the front of the cell . The transition between the time scales for ballistic and caged motion occurs at roughly 20 seconds , which corresponds to the average frequency of 2 . 9±0 . 2 protrusions/minute derived from discrete protrusions . Ballistic motion on short time scales suggests that peak protrusive activity travels along the boundary in a wave-like manner , similar to the peaks in boundary curvature . We estimate from the MSD at 8 seconds that the speed of these waves is at least 21 µm/min . This estimate is likely to be less than the true wave speed , since local motion ( from which the mean protrusion is found ) is measured across frames that are obtained 12 seconds apart , and since the displacement is not completely ballistic from 0 to 8 seconds . Together , these findings indicate that for adherent cells , protrusive activity is continuous and constantly shifts along the leading edge of the cell in a wave-like manner similar to the dynamics of the wave-like high curvature regions observed both in suspended cells and in cells extended over cliff edges .
Dictyostelium discoideum cells , WT ( AX3 ) , adenylyl cyclase A null , and myosin II null ( both aca− and myoII− are in an AX3 background ) were prepared as described previously [8] . Unless otherwise specified , WT and aca− cells were developed for 5 hours . MyoII− cells were developed for 6 hours . Since myoII− cells do not divide in suspension culture , they were harvested directly from plate cultures . In electrostatic repulsion experiments , cells were washed and run in 10−3 diluted phosphate buffer ( see [8] ) . All cells , except in cliff experiments , were cytoplasmically dyed with CellTracker Green CDMFA ( Invitrogen ) [8] . Acrylic resin micro-cliffs were fabricated using multiphoton absorption polymerization [36] ( See text S1 for more information ) . For cliff experiments , phase-contrast images were obtained with a 10× objective every 1 . 55 seconds . For footprint and polarization experiments , fluorescence , phase contrast and internal reflection microscopy ( IRM ) images were captured with a 40× objective every 2 or 4 seconds . For electrostatic repulsion experiments , fluorescence images were obtained on a Leica TCS SP2 confocal microscope with a 100× objective every 2 seconds . For the remaining experiments , fluorescence images were obtained on the same confocal microscope with a 40× objective every 4 seconds . Image sequences that show the migration of individual cells were pre-processed using ImageJ to enhance contrast and to remove other cells . The shape of the cell was then extracted using a snake algorithm . We adapted sample code [37] to follow the shape of a cell automatically throughout an image sequence and to parametrize the shape with 400 boundary points . Note that while within each frame neighboring points are equidistant along the perimeter , the distance between neighboring points varies across frames as the length of the cell perimeter changes . With a constant number of boundary points per frame , all local boundary points can be tracked simultaneously by defining a 1∶1 mapping between the points in each frame and the successive frame . We chose the mapping that maintained the numbering sequence of boundary points , and that minimized the sum of the square distances points moved from frame to frame . An example of such a tracking mapping is shown in video S10 . ( See text S1 for more about tracking . ) At each boundary point , we calculate the boundary curvature by fitting a circle to that boundary point and the two points that are 10 boundary points away from it . The magnitude of the boundary curvature is then defined as the reciprocal of the radius of that circle . If the midpoint of the two points 10 boundary points away is inside the cell , the curvature is defined as positive , otherwise it is defined as negative . For visualization , the curvature is smoothed over 3 boundary points and 3 frames , and the color scale is cut off at a maximum curvature magnitude . To calculate local motion , we first mapped each boundary point to the closest boundary point in the frame 12 seconds later . This time interval was chosen so that boundary motion dominates over image noise . Next , we used an averaging window to smooth the mapping twice . ( The first smoothing had a window size of 19 boundary points , while the second had a size of 15 boundary points . ) Figures S5a and b show an example of the mapping before and after smoothing . We defined local motion as the distance between mapped points . Like the curvature , when visualized , the color scale for local motion is cut off at a maximum magnitude . ( See text S1 for more information . )
|
Migration of cells on surfaces and through tissues is an important part of life , from the amazingly coordinated migration of cells during development to the uncontrollable migration of metastatic cancer cells . Here we investigate the physics of cell migration with the goal of gaining new insights into how cells move and how they respond to obstacles . Through detailed quantitative analysis of time-dependent cell shapes , we demonstrate the existence of wave-like dynamic shape changes during the migration of Dictyostelium discoideum . Specifically , we observe that in migrating cells local protrusions propagate from the front toward the back along roughly alternating sides of the cell . Near the leading edge , protrusions that have not yet adhered to the surface move faster than the cell migration speed . We also investigated cells that were not in contact with a surface and cells extended over cliff edges . Such cells exhibit similar protrusion dynamics at their leading edges but , since in this case the protrusions cannot adhere to a surface , the associated boundary curvature waves continue to travel along the sides of the cells . While our study shows wave-like dynamics under many different conditions , we do not directly measure whether waves involve transport of membrane or intracellular material .
|
[
"Abstract",
"Introduction",
"Results",
"Materials",
"and",
"Methods"
] |
[
"physics",
"model",
"organisms",
"protozoan",
"models",
"biology",
"microbiology",
"biophysics"
] |
2012
|
Cell Shape Dynamics: From Waves to Migration
|
Chronic Chagas cardiomyopathy ( CCC ) , a life-threatening inflammatory dilated cardiomyopathy , affects 30% of the approximately 8 million patients infected by Trypanosoma cruzi . Even though the Th1 T cell-rich myocarditis plays a pivotal role in CCC pathogenesis , little is known about the factors controlling inflammatory cell migration to CCC myocardium . Using confocal immunofluorescence and quantitative PCR , we studied cell surface staining and gene expression of the CXCR3 , CCR4 , CCR5 , CCR7 , CCR8 receptors and their chemokine ligands in myocardial samples from end-stage CCC patients . CCR5+ , CXCR3+ , CCR4+ , CCL5+ and CXCL9+ mononuclear cells were observed in CCC myocardium . mRNA expression of the chemokines CCL5 , CXCL9 , CXCL10 , CCL17 , CCL19 and their receptors was upregulated in CCC myocardium . CXCL9 mRNA expression directly correlated with the intensity of myocarditis , as well as with mRNA expression of CXCR3 , CCR4 , CCR5 , CCR7 , CCR8 and their ligands . We also analyzed single-nucleotide polymorphisms for genes encoding the most highly expressed chemokines and receptors in a cohort of Chagas disease patients . CCC patients with ventricular dysfunction displayed reduced genotypic frequencies of CXCL9 rs10336 CC , CXCL10 rs3921 GG , and increased CCR5 rs1799988CC as compared to those without dysfunction . Significantly , myocardial samples from CCC patients carrying the CXCL9/CXCL10 genotypes associated to a lower risk displayed a 2–6 fold reduction in mRNA expression of CXCL9 , CXCL10 , and other chemokines and receptors , along with reduced intensity of myocarditis , as compared to those with other CXCL9/CXCL10 genotypes . Results may indicate that genotypes associated to reduced risk in closely linked CXCL9 and CXCL10 genes may modulate local expression of the chemokines themselves , and simultaneously affect myocardial expression of other key chemokines as well as intensity of myocarditis . Taken together our results may suggest that CXCL9 and CXCL10 are master regulators of myocardial inflammatory cell migration , perhaps affecting clinical progression to the life-threatening form of CCC .
Approximately 8 million people are infected with the protozoan parasite Trypanosoma cruzi [1] in Central and South America , with an estimated 300 , 000 cases in the USA alone . T . cruzi is a major cause of heart disease and cardiovascular-related deaths in endemic areas located in Latin America , with approximately 50 , 000 fatalities per year due to Chronic Chagas cardiomyopathy ( CCC ) [2] . The high parasite load typical of the acute infection results in a strong innate and adaptive immune response against T . cruzi , leading to the control - but not the complete elimination - of tissue and blood parasitism , establishing a low-grade chronic persistent infection [3] . CCC is an inflammatory cardiomyopathy that affects approximately 30% of infected individuals and occurs 5–30 years after acute infection , while the remaining patients develop digestive disorders ( 5–10% ) or remain asymptomatic ( ASY , 60–70% ) [4] . It has been observed that the occurrence of myocarditis is correlated with clinical severity , ASY patients having minimal inflammation , while patients with advanced CCC display frequent and intense myocarditis [5] . Approximately 1/3 of patients developing CCC present a particularly lethal form of dilated cardiomyopathy , with shorter survival than idiopathic dilated cardiomyopathy [6] . In addition , CCC is associated with a worse prognosis and survival than other cardiomyopathies of non-inflammatory etiology ( NIC ) [7] . Although currently used trypanocidal drugs are effective in the treatment of acute or recent infection in children their efficacy in halting the progression of cardiac lesions has not been established yet [8] . Data suggest that the mononuclear inflammatory infiltrate , associated with cardiomyocyte destruction and fibrosis observed only in CCC myocardium , plays a major role in the development and progression of the disease [5] , [9] . Histologically , CCC myocardium displays a diffuse myocarditis with focal aspects; a mononuclear infiltrate , intense heart fiber damage , prominent fibrosis and scarcity of T . cruzi parasites ( reviewed in [10] ) . The inflammatory infiltrate of CCC heart lesions is composed mainly by T cells and macrophages [9] , [11] . Heart-infiltrating T cells and other mononuclear cells predominantly produce TNF-α and IFN-γ [12] , [13] , [14] , and a similar increase in IFN-γ and TNF-α is seen in cardiac tissue from animals infected with T . cruzi [15] , [16] . An increased number of IFN-γ-producing cells is found in peripheral blood mononuclear cells ( PBMC ) of CCC patients [14] , [17] . Together , data suggest a predominance of Th1-type T cells in CCC myocardium and peripheral blood [18] . However , the factors that determine the migration and accumulation of the Th1-type T cell in CCC heart tissue are still obscure . The possibility that chronic myocardial inflammation and tissue damage in CCC are a consequence of recognition of T . cruzi antigen on heart tissue must be entertained . However , T . cruzi DNA , a surrogate marker of the presence of living parasites , has been equally detected in hearts of CCC and ASY patients [19] , [20] , indicating that mere parasite presence in the heart is not sufficient for inducing inflammatory tissue damage . Both T . cruzi-specific [21] or heart antigen ( cardiac-myosin ) - specific T cells - crossreactive with T . cruzi antigen [22] - have been identified in the myocardial inflammatory infiltrate . Since both cardiac proteins and T . cruzi antigen are present in hearts of both CCC and ASY patients , some other factor distinct from antigen availability must control the difference in inflammation and disease progression between Chagas disease patients . Given the important role of chemokines such as CCL3 , CCL4 , CCL5 and CXCL9 and CXCL10 in tissue accumulation of CCR5+ and/or CXCR3+ Th1-type T cells [23] , it is possible that the development and maintenance of the tissue-damaging Th1-rich mononuclear infiltrate in CCC myocardium could be a consequence of the in situ expression of chemokine ligands to those receptors . Indeed , increased mRNA levels of CC and CXC chemokines have been detected in the heart tissue of T . cruzi-infected mice ( reviewed in [24] , [25] ) . Additionally , increased numbers of CCR5+ CXCR3+ CD4+ and CD8+ T cells producing IFN-γ and/or TNF-α were found in PBMC from CCC patients , as compared with ASY patients [26] . Previous studies from our group showed that gene expression levels of IFN-γ-inducible chemokines CXCL9 and CXCL10 were significantly up-regulated in CCC myocardium [27] . Local production of CXCL9 and CXCL10 by mononuclear infiltrating and stromal cells leads to the recruitment of effector Th1 lymphocytes into inflamed tissues in delayed hypersensitivity reactions ( reviewed in [28] ) . Conversely , functional polymorphisms controlling expression , and loss-of-function deletions in protein-coding regions of genes encoding chemokines and their receptors have been associated with development of several inflammatory and autoimmune diseases [29] , [30] , [31] . This indicates that chemokine polymorphisms can control organ-specific inflammatory damage -even in the presence of similar amounts of antigen- by means of regulating inflammatory cell influx . We therefore hypothesized that an imbalance at the Th1-associated chemokine-chemokine receptor axis - perhaps of genetic origin - could play a role in the maintenance of the inflammatory infiltrate in CCC patients . Here , we analyzed cell surface staining and gene expression of the chemokine receptors CXCR3 and CCR5 , associated to the Th1 phenotype , and their ligands . Conversely , we also studied the CCR4 and CCR8 receptors , associated to the Th2 phenotype , and their chemokine ligands , as well as the chemokine receptor CCR7 , associated to memory phenotypes , and its ligands in myocardial samples form end-stage CCC and non-inflammatory cardiomyopathy patients , as well as control subjects . In addition , we analyzed single-nucleotide polymorphisms ( SNPs ) for genes encoding 8 such chemokines and receptors in a cohort of Chagas disease patients stratified according to clinical form ( ASY , CCC ) and presence of ventricular dysfunction among CCC patients .
The protocol was approved by the Institutional Review Board of the University of São Paulo School of Medicine ( Protocol number 739/2005 and 0324/2009 ) and written informed consent was obtained from the patients . In the case of samples from heart donors , written informed consent was obtained from their families . All Chagas disease patients were considered serologically positive for antibodies against T . cruzi on the basis of results of at least 2 of 3 independent tests ( EIA [Hemobio Chagas; Embrabio São Paulo] , indirect immunofluorescence assay [IFA-immunocruzi; Biolab Merieux] , and indirect hemagglutination test [Biolab Merieux] ) . All Chagas disease patients underwent standard electrocardiography and echocardiography . Echocardiography was performed in the hospital setting using an Acuson Sequoia model 512 echocardiographer with a broad-band transducer . The left ventricular dimensions and regional and global function evaluations were performed using a 2-dimension and M-mode approach , in accordance with the recommendations of the American Society of Echocardiography . Patients with CCC presented with abnormal electrocardiography findings that ranged from typical conduction abnormalities ( right bundle branch block and/or left anterior division hemiblock ) to severe arrhythmia [32] . A group of patients also presented varying degrees of ventricular dysfunction classified on the basis of left ventricular ejection fraction , with all other causes of ventricular dysfunction/heart failure excluded . All asymptomatic ( ASY ) subjects had normal echocardiograph and echocardiogram findings , chest radiographs with no evidence of cardiac enlargement , and normal findings of radiographs of the esophagus and colon . Myocardial left ventricular free wall heart samples were obtained from end-stage heart failure CCC patients ( n = 14 , positive for antibodies against T . cruzi and born in endemic areas for Chagas disease , five males and nine females , mean age 47 . 2±14 . 6 years , Table 1 ) and end-stage heart failure patients with non-inflammatory cardiomyopathies ( NIC , n = 8 , five patients with idiopathic dilated cardiomyopathy and three patients with Ischemic cardiomyopathy , all seronegative for T . cruzi , eight males , mean age 53 . 3±7 . 5 years , Table 1 ) . Control adult heart tissue from the left ventricular-free wall was obtained from nonfailing donor hearts ( N , n = 6 , males , mean age 32 . 2±12 . 8 years , Table 1 ) not used for cardiac transplantation due to size mismatch with available recipients . Hearts were explanted at the time of heart transplantation at the Heart Institute - InCor , University of São Paulo School of Medicine , São Paulo , SP , Brazil . For histological studies , fresh tissues were fixed in buffered formalin solution and paraffin-embedded; for immunofluorescence , fresh samples were maintained in a 30% sucrose solution for approximately 30 min at 4°C; then they were transferred to OCT Tissue Tek freezing medium and immediately frozen in isopentane and stored at −80°C . For mRNA extraction , samples were quickly dissected , and myocardial tissue was frozen in liquid nitrogen and stored at −80°C . For SNP genotyping , genomic DNA was obtained from myocardial left ventricular free wall heart samples using the QIAamp DNA Blood Max Kit ( Qiagen , Hilden , Germany ) and stored at −20°C . Blood samples were collected from the patients who were categorized as being ASY ( n = 151 , 72 males , mean age 50 . 8±3 . 6 years , and 79 females , mean age 54 . 1±3 . 5 years ) or as having CCC ( n = 174 , 83 males , mean age 50 . 8±0 . 5 years , and 91 females , mean age 48 . 0±6 . 6 years ) on the basis of clinical , radiological , electrocardiographic ( ECG ) , echocardiographic criteria and all of patients were serologically positive for antibodies against T . cruzi and born in endemic areas for Chagas disease . All of ASY patients had normal ECG findings and a normal left ventricular ejection fraction ( LVEF ) at the time of echocardiography , as well as normal findings for chest , esophagus and colon radiography . The CCC patients presented with typical ECG findings ( right bundle branch block and/or left anterior division hemiblock and were classified as having moderate ( LVEF>40% , n = 79 ) or severe ( LVEF≤40% ) , n = 95 ) CCC . Genomic DNA was extracted by the dodecyltrimethyl ammonium bromide/hexadecyltrimethylammonium bromide method and stored at −20°C . Individual 5-µm sections of paraffin-embedded tissue or cryosections of frozen myocardial fragments were applied to microscope slides . Slides were subjected either to hematoxylin-eosin or immunofluorescence staining . Standard hematoxylin-eosin staining was performed for evaluation of the intensity and location of the inflammatory infiltrate . Slides were evaluated and scored for the intensity of myocarditis , fibrosis and hypertrophy . Slides for immunofluorescence were washed in phosphate-buffered saline ( PBS ) and blocked with PBS-2% bovine serum albumin . Slides were incubated overnight at 4°C with the primary monoclonal antibodies anti-CD3 , anti-CD4 , anti-CD8 , anti-CCR5 , anti-CXCR3 , anti-CCR4 , anti-CCL5 or anti-CXCL9 ( R&D Systems , Minneapolis , MN , USA ) ; after washing , slides were incubated with the secondary antibody , anti-mouse IgG conjugated to AlexaFluor 633 ( Invitrogen , Carlsbad , CA , EUA ) and 4′ , 6′-diamidino-2-phenylindole ( DAPI , Sigma-Aldrich , Steinhein , Germany ) for nuclear staining . Slides were mounted using antifade mounting medium ( Hydromount , National Diagnostics , Atlanta , GA , USA ) . Fluorescent images were acquired using UV/Laser excitation on an LSM/Meta 510 Zeiss microscope , with an oil immersion objetive ( 63× , 1 . 25 numerical aperture ) . For each section , the area with most uniform infiltrate was selected for analysis . Within this inflammatory area , minimums of five fields were acquired . Image analysis and processing were performed using LSM Image Examiner software ( Carl Zeiss , Standort Göttingen , Germany ) . Omission of primary antibody and replacement with non-immune mouse IgG was used to confirm the lack of nonspecific staining . Total RNA was extracted from 5×5×5 mm myocardial samples using the Trizol® method ( Life Technologies Inc . , Grand Island , NY ) . The RNA was quantified using NanoDrop Spectrophotometry ( Thermo Scientific ) , and treated with Rnase-free DNase I ( USB , Ohio , USA ) . cDNA was obtained from 5 µg total RNA using Super-script II™ Reverse Transcriptase ( Invitrogen , Carlsbad , CA , USA ) . We designed forward and reverse primers for real-time qPCR assays using the Primer Express software ( Applied Biosystems , Foster City , CA , USA; Table S1 ) . Real-time qPCR reactions were carried out in an ABI Prism 7500 Sequence Detection system ( Applied Biosystems ) using the SYBR Green PCR Master Mix ( Applied Biosystems ) , as described [33] . For all genes , we constructed standard curves and determined the slope to calculate the PCR efficiency . All the samples were tested in triplicate with the GAPDH , previously shown to display little variance among human myocardial tissue samples [27] , as the reference gene for normalization of data , and relative expression of each mRNA was calculated with the 2−ΔΔCt method [34] , using expression in six normal donor hearts as calibrator . We searched for polymorphisms in our target genes preferentially in putatively regulatory regions such as 5′ and 3′untranslated regions ( UTR ) or intronic regions , and which had previously been studied in disease association studies . This was the case of CXCL9 rs10336 , CXCL10 rs3921 , CCL5 rs2107538 and CCR5 rs1799988 [35] , [36] , [37] . In the cases of target genes where no previous study was found - CCL4 , CCL17 , CCL19 and CXCR3 - the selected SNP's ( CCL4 rs1719153 , CCL17 rs223827 , CCL19 rs3136658 and CXCR3 rs2280964 ) were located in putatively regulatory sites , with minimum allele frequencies above 0 . 2 in the CEU ( Caucasian ) and Yoruba ( YRI ) populations at the HapMap database ( site http://hapmap . ncbi . nlm . nih . gov ) ( Table S2 ) . Genotyping was performed using the designed assay of the TaqMan allelic discrimination technique ( ABI 7500 , Applied Biosystem , Foster City , USA ) and evaluated according to the manufacturer's instructions . The following SNPs were tested: SNP CCL4 rs1719153 ( part n° . C_12120537_10 ) , CCL5 rs2107538 ( part n° . C__15874407_10 ) , CCR5 rs1799988 ( part n° . C__11988170_10 ) , CXCL9 rs10336 ( part n° . C_486222_10 ) , CXCL10 rs3921 ( part n° . C_497062_10 ) , CXCR3 rs2280964 ( part n° . C_15874773_10 ) , CCL17 rs223827 ( part n° . C_2845028_10 ) and CCL19 rs3136658 ( part n° . C__25981926_10 ) CCL4 , CXCL9 and CXCL10 are mapped to the 3′ untranslated region ( UTR ) ; CCL5 and CCR5 are mapped to the 5′ UTR; CCL17 , CCL19 and CXCR3 are mapped to intronic regions . Values of the relative expression of each mRNA in the CCC and NIC groups were compared with the Mann-Whitney U test . Correlation analysis was performed by Spearman's nonparametric correlation test with SPSS version 14 . 0 software ( SPSS , Chicago , III ) . Associations between patient groups and genotypes or a specific allele were analyzed by the x2 statistical test , along with the relevant odds ratio ( OR ) and 95% confidence interval ( CI ) . Fisher's exact test was used when at least one value in the contingency table was <5 . We considered differences as significant when the p value was less than 0 . 05 and the CI did not cross 1 . Hardy-Weinberg equilibrium ( HWE ) was determined by comparing the observed number of different genotypes with the expected number under the HWE for the estimated allele frequency .
While myocardial sections from both CCC and NIC groups displayed cardiomyocyte hypertrophy and fibrosis upon histopathological analysis , lymphocytic myocarditis was only observed among samples from CCC patients ( Figure 1A , Table 1 ) . Although all CCC patients presented clinically similar end-stage heart disease , variable degrees of lymphocytic myocarditis were observed . No significant differences were found in age , ejection fraction ( EF ) or left ventricular diastolic diameter ( LVDD ) between the two groups , and mRNA expression of the natriuretic peptides ANP and BNP was substantially upregulated in comparison to control myocardium samples ( Table 1 ) , indicative of activation of the embryonic/hypertrophic gene expression pattern consistent with advanced heart failure . Using immunofuorescence confocal microscopy , we assessed the surface phenotype of mononuclear cells in the myocardial inflammatory infiltrate of CCC samples ( Figure 1B and Figure S1 ) . CD3+ , CD4+ and CD8+ T cells were abundant in CCC myocardium , very scarce in NIC samples and essentially absent in control samples ( N ) , in line with histopathological data ( Figure S1 ) . We observed mononuclear cells staining for CXCR3 and CCR5-typical of Th1 T cells - as well as CCR4 – more commonly described as a marker of Th2 T cells in CCC myocardium ( Figure 1B ) . CXCR3 staining was also observed in non-mononuclear , stromal cells in CCC heart tissue ( Figure 1B , Figure S1 ) . In addition , we observed in situ expression of the chemokines CCL5 and CXCL9 , ligands of the CCR5 and CXCR3 receptors , in mononuclear cells infiltrating the myocardium of CCC patients ( Figure 1B ) . We could also observe CCL5 staining in endothelial cells lining what seems to be a blood vessel in CCC myocardium ( Figure 1B , Figure S1 ) . Specificity of the tested antibody staining is shown at Figure 1B . We assessed mRNA expression of CCR5 and CXCR3 chemokine receptors associated with inflammatory , Th1-type T cells , as well as the chemokine receptors CCR4 and CCR8 , which are associated to a Th2 phenotype , and CCR7 , a chemokine receptor associated to memory phenotypes , in myocardium tissue from CCC , NIC and control samples using real-time qPCR . Figure 2A shows that mRNA expression levels of CCR5 , CXCR3 and CCR4 were significantly higher in CCC myocardial tissue that of NIC patients and normal donors , which corroborated with the immunofluorescence detection data ( Figure 1B ) . The expression of CCR8 mRNA did not differ significantly between the CCC and NIC groups ( Figure 2A ) . The expression of CCL1 ( CCR8 ligand ) was undetectable in CCC , NIC and control samples . The expression of mRNA encoding CCR5 ligands ( CCL3 , CCL4 and CCL5 ) , CXCR3 ligands ( CXCL9 and CXCL10 ) , the CCR4 ligand CCL17 and the CCR7 ligand CCL19 was significantly higher in CCC than in NIC myocardium ( Figure 2B ) . Of note , median CCL5 expression in CCC myocardium was over 100-fold higher than normal controls . The expression of CXCL9 , CXCL10 , CCL17 and CCL19 in CCC samples was between 10 and 100-fold higher than normal control samples . Of interest , mRNA levels of CCL21 , another CCR7 ligand , were 5-fold higher in samples from both CCC and NIC heart tissue than in samples from control individuals ( Figure 2B ) . As the expression of CCR7 was undetectable in myocardium samples from control individuals , the relative expression of this receptor could not be subjected to relative quantification using the ΔΔCt method . We found that the average ΔCt value of CCR7 in heart tissue from the CCC patients was significantly lower than in NIC myocardium ( 14 . 52±2 . 9 versus 18 . 34±2 . 5; p = 0 . 01 , data not shown ) , indicating that this gene also displayed increased expression in CCC as compared to NIC myocardium . We also analyzed whether chemokine expression was associated with histopathological parameters in CCC heart tissue . We found a positive correlation between mRNA expression of CXCL9 and the histopathologically assessed intensity of myocarditis ( Figure 3 ) . CXCL9 also showed a positive correlation with chemokine receptors CXCR3 , CCR4 and CCR5 , as well as chemokines CCL5 , CXCL10 and CCL17 . In addition , positive correlations were also found between CXCR3 and CCR5; the chemokine receptors and several of their ligands; and between several CC and CXC chemokines ( Figure S2 ) . Given the focal nature of the myocarditis and fibrosis in CCC myocardium , we assessed the gene expression profile in different sites of the left ventricular free wall of three patients . This profile was found to be essentially concordant in samples from different sites of the myocardium of the same patients ( Table S3 ) . We also found a positive correlation of global gene expression between almost all CCC samples ( Figure S3 ) , indicating that most CCC patients share a similar pattern of gene expression , with individual-specific variation in global magnitudes of expression . These global magnitudes of gene expression seem thus to be maintained in distinct regions of the myocardium ( Table S3 ) . In order to investigate whether the observed variation in global magnitudes of gene expression was due to genetic polymorphisms , we evaluated single nucleotide polymorphisms in the most highly expressed genes in CCC heart tissue ( fold expression >10 as compared to normal donor samples ) . We compared genotypic and allele frequencies in the CCC and ASY groups of a sample of 325 Brazilian chronic Chagas disease patients , and also stratified CCC patients according to their LVEF status , comparing the two CCC groups according to severity of ventricular dysfunction: those patients classified as severe ( LVEF≤40% ) or moderate ( LVEF>40% ) CCC . All the SNPs were in HWE in both CCC and ASY subjects . There are no significant differences in genotype and allele frequencies of SNP CXCL9 rs10336 between CCC and ASY ( Table 2 ) . However , the CXCL9 rs10336 CC genotype was significantly less frequent among the severe CCC than in the moderate CCC group ( OR , 0 . 47 [95% CI , 0 . 25–0 . 87] Table 2 ) . Accordingly , the CXCL9 rs10336 C allele was significantly less frequent among the severe CCC than in the moderate CCC group . Thus , both the C allele and the CC genotype are associated to a lower risk for severe CCC . There are no significant differences in GG and CG frequencies of SNP CXCL10 rs3921 polymorphism in the patients with CCC when compared with ASY patients ( Table 3 ) . However , the CXCL10 rs3921 GG genotype was significantly less frequent among the severe CCC than in the moderate CCC group ( OR , 0 . 41 [95% CI , 0 . 22–0 . 79] Table 3 ) . Accordingly , the CXCL10 rs3921 G allele was significantly less frequent among the severe CCC than in the moderate CCC group . Therefore , both the G allele and the GG genotype apparently are associated to a lower risk for severe CCC . Conversely , the CXCL10 rs3921 C allele was significantly more frequent between the severe CCC than in the moderate CCC group , thus the CC genotype is associated to a higher risk for severe CCC . The genotype and allele frequencies of the CCR5 rs1799988 CT polymorphism are given in Table 4 . The genotype and allele distribution of this polymorphism in the CCC patients was not significantly different from that found among ASY patients . The CCR5 rs1799988 CC genotype was significantly more frequent among the severe CCC than in the moderate CCC group ( OR , 2 . 31 [95% CI , 1 . 14–4 . 67] Table 4 ) . Accordingly , the CCR5 rs1799988 C allele was significantly more frequent among the severe CCC than in the moderate CCC group . Results indicate that both the C allele and the CC genotype of CCR5 rs1799988 are associated to a higher risk for severe CCC . No differences in genotypic and allelic frequencies were found in the studied SNPs in CCL4 , CCL5 , CXCR3 , CCL17 and CCL19 between CCC and ASY patients ( Tables S4 , S5 , S6 , S7 and S8 ) . We next analyzed whether CXCL9 and CXCL10 polymorphisms associated to decreased or increased risk of CCC development conveyed different levels of chemokine/chemokine receptor expression in CCC heart samples . The study could not be done for CCR5 because there was only one TT patient . We found that myocardial expression of CXCL9 mRNA among samples from patients bearing the protective SNP CXCL9 rs10336 genotype CC ( n = 5 ) was 5-fold lower than that of patients bearing the TT or CT genotypes . We also found that CCL3 , CCL5 , CXCL10 , CXCR3 , CCL17 and CCR4 were 2–6-fold less expressed in SNP CXCL9 rs10336 genotype CC as compared with samples from patients bearing the TT or CT genotypes ( Figure 4A ) . Likewise , we found that myocardial CXCL10 expression among samples from patients bearing the protective SNP CXCL10 rs3921 genotype GG ( n = 4 ) was 3-fold less intense than that of patients bearing the CC or CG genotypes . We also found that CCL3 , CXCL9 and CCR4 were 2–3 fold less expressed in samples from patients carrying SNP CXCL10 rs3921 genotype GG as compared with those from patients bearing the CC or CG ( Figure 4B ) .
Although the pathogenic role of myocarditis in CCC is well established , the factors that lead to the maintenance of the myocardial infiltrate have remained unclear . In this paper , we report that mononuclear cells expressing CD3 , CD4 , CD8 , CCR5 , CXCR3 , CCR4 , CXCL9 and CCL5 are found in CCC myocardial tissue . We also found increased mRNA expression of CXCL9 , CCL5 , the above mentioned chemokine receptors and several of their other chemokine ligands . CXCL9 expression correlated with the intensity of the myocardial infiltrate , as well as with the mRNA expression of CXCR3 , CCR5 , CCR4 , CCR7 and several of their ligands . The closely linked CXCL9 rs10336 CC and CXCL10 rs3921 GG genotypes were associated with protection from severe CCC , while CCR5 rs1799988 CC is associated with increased risk for development of severe CCC . Myocardium samples carrying the protective CXCL9 rs10336 CC genotype expressed significantly less mRNA encoding CXCL9 itself , along with CCL3 , CCL5 , CXCL10 , CXCR3 , CCL17 and CCR4 , than samples carrying other CXCL9 rs10336 genotypes . Similarly , samples carrying the protective CXCL10 rs3921 GG genotype expressed less mRNA encoding CCL3 , CXCL9 , CXCL10 and CCR4 than samples carrying other CXCL10 rs3921 genotypes . These results suggest that CXCL9 ( and possibly CXCL10 ) may play an important role modulating the CXCR3/CCR5 chemokine axis and the intensity of mononuclear cell migration to CCC myocardium . The identification of CCR5+ and CXCR3+ mononuclear cells , corroborated by the increased mRNA expression of the same chemokine receptors ( Figure 1 and 2 ) in CCC myocardium , is in line with the prominent finding of mononuclear cells secreting IFN-γ and TNF-α in CCC heart tissue [13] , [14] , [38] , and together are consistent with the influx of Th1 cells expressing these receptors . CCR5 is known to be expressed by macrophages , T and B cells , while CXCR3 is expressed by T and B cells; such cells may thus have been attracted by the respective chemokine ligands [39] . Indeed , CCC patients display increased numbers of CD4+ and CD8+ T cells coexpressing CXCR3/CCR5 and IFN-γ/TNF-α in peripheral blood when compared to T . cruzi-seropositive ASY/IND individuals [26] . In addition , CXCR3 expression was also observed in stromal cells from CCC myocardium ( Figure 1B ) , in line with reports of CXCR3 expression in non-lymphoid cells [40] , which may have contributed to the increased CXCR3 mRNA expression in CCC heart tissue . The high expression of CCR4 and CCL17 ( Figure 2 ) is consistent with the presence of Th2 T cells [41] . However , it is unlikely that such cells are functional Th2 T cells , since preliminary observations indicate that IL-4 , IL-5 , and IL-13 are not expressed in CCC heart tissue ( unpublished results ) . It is possible that other cell types , such as endothelial cells or regulatory T cells , may be responsible for CCR4 and CCL17 expression in CCC myocardium [42] , [43] . The finding that CCR7 and its ligands CCL19 and CCL21 were upregulated in CCC myocardium is consistent with a role in the recruitment of CCR7+ cells from the periphery ( Figure 2B ) . CCR7 can be expressed by naïve or central memory T cells , along with endothelial cells , fibroblasts and B cells [44] . Endothelial expression of CCR7 ligands is a known chemoattractant for CCR7+ central memory and even naïve T cells for migration into inflamed tissue [44] . CCL21 , whose mRNA expression is upregulated in myocardium samples of both CCC and NIC patients , is also a chemoattractant for CCR7+ fibrocytes [45] , inducing expression of type I collagen in vitro [46] and thus promotes fibrosis . CCL21 expression may thus be one of the stimuli driving myocardial fibrosis and pathological remodeling both in CCC and NIC . The identification of CCL5+ and CXCL9+ mononuclear cells ( Figure 1B ) , as well as the mRNA expression of these and other CCR5 and CXCR3-binding chemokines ( Figure 2B ) is in line with the observed accumulation of CCR5+ and CXCR3+ mononuclear cells in CCC myocardium . CXCR3 and CCR5-binding chemokines have been postulated to play a role in recruiting polarized Th1 T cells to sites of chronic inflammation [41] , and most of them are also IFN-γ-inducible chemokines [47] . Interestingly , discrete staining for CCL5 was also observed in structures consistent with microvascular endothelium ( Figure 1B ) , suggesting a role for this chemokine , and possibly other CCR5 ligands , in the migration and accumulation of Th1 cells in CCC heart tissue . CCL3 , CCL4 , CCL5 , CXCL9 and CXCL10 are produced by stimulated T lymphocytes , monocytes/macrophages , as well as fibroblasts , endothelial cells and cardiomyocytes [39] , [48] , [49] , [50] and the myocardial tissue of T . cruzi-infected mice [51] , [52] . Indeed , treatment of T . cruzi-infected mice with Met-Rantes , a CCL5-based CCR5 antagonist , reduced the intensity of the inflammatory heart infiltrate , with little effect on parasitism [52] , suggesting that CCL5-induced migration of CCR5 inflammatory cells may play a direct role in the genesis of acute T . cruzi-induced myocarditis . A recent study has shown that Beagle dogs presenting the cardiac form of T . cruzi infection presented higher myocardial mRNA expression of CXCL9 and CCL5 than uninfected animals , during the acute and chronic phases , respectively [53] . Moreover , microarray analysis have identified genes related to inflammation , such as chemokines , to be upregulated in myocardium of mice chronically infected with the Colombian strain of T . cruzi and in primary murine cardiomyocytes infected with T . cruzi [54] , [55] , The fact that CXCL9 mRNA expression was the only chemokine found to display significant correlation with the intensity of the myocardial infiltrate suggests CXCL9 plays a role in mononuclear cell migration to CCC myocardium . The correlation between CXCL9 expression and that of other chemokines , as well as their CXCR3/CCR5/CCR4/CCR7 receptors ( Figure 3 and Supplemental data ) may be partially related to the finding that CXCL9 has immunomodulatory properties , being able to directly increase expression of CCL2 , CCL3 and CCL4 among other inflammatory genes such as TNF-α [56] . Expression of all three chemokines has been shown to be upregulated in CCC myocardial tissue [27] , suggesting that CXCL9 may multiply its T-cell chemoattractant properties with the aid of CCR5 ligands in CCC heart tissue . It can thus be hypothesized that IFN-γ secreting type 1 T cells in CCC myocardium induce local production of CXCL9 ( and perhaps CXCL10 ) , which in turn further recruits IFN-γ-secreting CXCR3+/CCR5+ type 1 T cells and other inflammatory cells directly and indirectly , perpetuating the influx of pathogenic Th1-type T cells and inflammation . The polymorphisms CXCL9 rs10336 CC and CXCL10 rs3921 GG , associated with protection from progression to severe CCC , are located in the 3′UTR region , and the polymorphism CCR5 rs1799988 CC , associated with increased risk for development of severe CCC , is located at the 5′UTR region , where they may influence binding of gene expression control regions to regulatory elements . The findings are in line with previous associations of chemokine or chemokine receptor polymorphisms in CCC ( CCL2 [57] , CCL5 [58] , CCR5 [29] , [58] . However , to our knowledge , is the first study to find gene polymorphisms associated to the transition to CCC with ventricular dysfunction , which is the most clinically relevant presentation of disease , the one associated with significant morbidity and mortality . Previous gene polymorphism association studies performed in Chagas disease only disclosed polymorphisms relevant for the transition between the ASY and the CCC forms , irrespective of clinical severity ( reviewed in [10] , [59] , [60] , [61] . Associations among polymorphic variants in the CCR5 , CXCL9 and CXCL10 genes and disease phenotypes have previously been reported . A sequence variant in the promoter region of the CCR5 gene , which is associated to reduced expression of CCR5 , is associated with protection to AIDS progression [62] and is also present at a higher frequency in ASY compared with CCC patients [29] . Although Bruck et al [35] failed to observe any association between the CXCL9 rs10336 CT polymorphism and type 1 diabetes , the CXCL9 rs2276886 GA polymorphic variant was associated with a reduced risk for pediatric Crohn's disease; significantly , CXCL9 was found to be overexpressed in Crohn's disease gut tissue [63] . The CXCL10 -201 G/A polymorphism is associated with increased susceptibility to chronic hepatitis B disease progression in males . The CXCL10 -201 G/A is a functional polymorphism , where allelic variants modulate CXCL10 expression by differential interaction with CXCL10 transcription factors [64] . Still another CXCL10 polymorphic variant , CXCL10 rs8878 TT , was found to reduce the risk of type 1 diabetes [65] . Since all our CCC myocardium samples came from clinically similar end-stage patients submitted to transplantation , it could be argued that possessing CXCL9 and/or CXCL10 genotypes associated to reduced myocardial chemokine expression , or even displaying a less significant inflammatory infiltrate by itself – may not be relevant for the progression of CCC . However , CCC is not a monogenic disease , and it is likely that the progression to overt inflammatory dilated cardiomyopathy may result from the combined effect and inadequate counterregulation of relevant genes . Polymorphisms in multiple innate immunity/inflammatory genes , like IL1β , TNF-α , IL10 , IL12 , lymphotoxin-α , BAT1 , NFKBIL1 , MAL/TIRAP , MIF have been found to associate with risk for developing CCC ( reviewed in [10] , [59] ) . In addition to interference by other genes , differential myocardial resilience , including responses to hypertrophic/fibrogenic factors occurring in CCC heart tissue ( IL1β , TNF-α , IFN-γ , IL18 , CCL21 ) , could explain why these patients carrying the CXCL9/CXCL10 protective genotypes progressed to end-stage cardiomyopathy . In the Syrian hamster model of chronic Chagas disease cardiomyopathy , although the intensity of chronic inflammation correlated with ventricular dilation , intensity of myocarditis was similar in hamsters dying from chronic T . cruzi-induced dilated cardiomyopathy and survivors euthanized 11 months post-infection [10] , suggesting the existence of additional factors related to death from CCC . The finding that the protective CXCL9 rs10336 CC and CXCL10 rs3921 GG genotypes were associated with lower myocardial expression of the respective chemokines indicate either that the polymorphisms are functional themselves or closely linked to functional polymorphisms . These polymorphisms are strongly linked in the caucasian and negroid populations in the HapMap and test populations ( data not shown ) . Indeed , the CXCL9 and CXCL10 genes are contained in a CXC chemokine “mini-cluster’ at chromosome position 4q21 . 2 [61] , [66] and positioned less than 20 Kb apart . The proximity and linkage of the loci renders it difficult to ascertain which of the two genes is actually responsible for the immunological effect . In fact , we found over 60 additional SNPs tightly linked to CXCL9 rs10336 and CXCL10 rs3921 in the reference HapMap/CEPH CEU ( Northern European Caucasian ) and YRI ( Yoruba , African ) populations in the vicinity of the CXCL9-CXCL10-CXCL11 chemokine minicluster at Chromosome 4 . Some of them are at the 3′ UTR or 5′UTR regions of either CXCL9 or CXCL10 , with possible transcriptional regulatory activity , and could , in theory , be responsible for the observed functional association ( Table S9 ) . The fact that samples carrying genotypes associated to lower risk also displayed reduced expression of several other chemokines and receptors , as well as with reduced intensity of the myocardial infiltrate , may suggest that the CXCL9 and CXCL10 genotypes control the intensity of myocarditis through modulation of CXCL9 and/or CXCL10 production and its ensuing effects on chemokine ligands of the CXCR3/CCR5/CCR4 receptors . Together with the finding of a correlation between expression of CXCL9 and intensity of myocarditis , these results suggest that CXCL9 may be a master regulator of chemokine-driven inflammatory cell migration to CCC heart tissue . Since the intensity of the infiltrate seems to be a major pathogenic factor , inflammatory cell recruitment into the myocardium driven by CXCL9 and other CC and CXC chemokines may play an important role in the clinical progression to severe , life-threatening CCC . Several chemokine receptor antagonists are either on late clinical trials or have been licensed [67] , [68] . CCR5 and CXCR3 blockers have been shown to suppress migration of inflammatory cells in acute allograft rejection [69] . Treatment with Met-Rantes ameliorated acute Chagas disease myocarditis in murine models [52] . Together with these reports , our results suggest that the CCR5/CXCR3 receptors may be therapeutic targets in CCC , where receptor antagonists could dampen inflammatory T cell migration to myocardium , thus controlling CCC myocarditis and , possibly , ventricular dysfunction and death .
|
Chronic Chagas cardiomyopathy ( CCC ) is an inflammatory heart disease that affects millions in Latin America , and in growing numbers in USA and Europe . Survival among CCC patients is shorter than among patients with cardiomyopathy of non-inflammatory etiology . This suggests that the inflammatory cell influx plays an important pathogenic role in CCC . However , little is known about the factors that maintain this myocardial inflammation . We hypothesized that Th1 T cell-attracting chemokines , involved in driving leukocyte migration , could play a role in myocardial inflammation . Herein , we have analyzed expression of several chemokines and receptors in heart tissue from patients with CCC and controls . We found inflammatory cells expressing chemokines and receptors consistent with Th1 T cell influx into CCC myocardium . mRNA expression levels of the chemokine CXCL9 correlated with inflammation . We also studied whether genetic variations in these genes could be associated to CCC development . Polymorphisms in CXCL9 , CXCL10 and CCR5 were associated to differential risk of progression to the more severe form of CCC . Polymorphisms of CXCL9 and CXCL10 were also associated to the intensity of myocardial inflammation and chemokine expression . These results suggest that such chemokines may be master regulators of myocardial inflammatory cell migration , perhaps affecting clinical progression to severe CCC .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"gene",
"expression",
"genetics",
"immunology",
"biology",
"genetics",
"and",
"genomics"
] |
2012
|
Myocardial Chemokine Expression and Intensity of Myocarditis in Chagas Cardiomyopathy Are Controlled by Polymorphisms in CXCL9 and CXCL10
|
Loxoscelism is the designation given to clinical symptoms evoked by Loxosceles spider's bites . Clinical manifestations include skin necrosis with gravitational spreading and systemic disturbs . The venom contains several enzymatic toxins . Herein , we describe the cloning , expression , refolding and biological evaluation of a novel brown spider protein characterized as a hyaluronidase . Employing a venom gland cDNA library , we cloned a hyaluronidase ( 1200 bp cDNA ) that encodes for a signal peptide and a mature protein . Amino acid alignment revealed a structural relationship with members of hyaluronidase family , such as scorpion and snake species . Recombinant hyaluronidase was expressed as N-terminal His-tag fusion protein ( ∼45 kDa ) in inclusion bodies and activity was achieved using refolding . Immunoblot analysis showed that antibodies that recognize the recombinant protein cross-reacted with hyaluronidase from whole venom as well as an anti-venom serum reacted with recombinant protein . Recombinant hyaluronidase was able to degrade purified hyaluronic acid ( HA ) and chondroitin sulfate ( CS ) , while dermatan sulfate ( DS ) and heparan sulfate ( HS ) were not affected . Zymograph experiments resulted in ∼45 kDa lytic zones in hyaluronic acid ( HA ) and chondroitin sulfate ( CS ) substrates . Through in vivo experiments of dermonecrosis using rabbit skin , the recombinant hyaluronidase was shown to increase the dermonecrotic effect produced by recombinant dermonecrotic toxin from L . intermedia venom ( LiRecDT1 ) . These data support the hypothesis that hyaluronidase is a “spreading factor” . Recombinant hyaluronidase provides a useful tool for biotechnological ends . We propose the name Dietrich's Hyaluronidase for this enzyme , in honor of Professor Carl Peter von Dietrich , who dedicated his life to studying proteoglycans and glycosaminoglycans .
Bites involving brown spiders are characterized by skin injuries at the venom inoculation site , including swelling , erythema , hemorrhage , dermonecrosis , and the hallmark of loxoscelism: gravitational spreading of cutaneous lesions [1] , [2] . Systemic involvement has also been reported including fever , malaise , weakness , nausea , vomiting and in severe cases , intravascular coagulation , hemolysis and acute renal disturbance [1] , [2] , [3] , [4] , [5] . The gravitational spread of skin lesions is a distinct characteristic of loxoscelism , described after experimental venom exposure in the skin of rabbits and in real cases . It appears hours or days after venom inoculation . Macroscopically , the development of lesions disperses in a gravitational direction with erythema , swelling , dark blue-violet color , and an eschar . Histologically , the lesion is reported as a collection of inflammatory cells in and around the blood vessels and diffusely distributed in the dermis . It is also possible to observe degeneration of blood vessel walls , disorganization of collagen fibers with edema , hemorrhage into the dermis , necrosis of cells , and destruction of tissue structures . Pathologically , the wound is described as aseptic coagulative necrosis [1] , [2] , [6] , [7] , [8] . The molecular mechanism by which brown spider venom induces gravitational spreading of skin lesions and systemic involvement is not fully understood . A fundamental requirement for venom to induce local spreading of lesions and systemic involvement is the presence of venom components that are able to degrade tissue barriers . The delivery of venom toxins to neighboring bite sites and into systemic circulation is assisted by molecules that degrade extracellular matrix constituents such as proteases and hyaluronidases [9] , [10] , [11] , [12] . The venom is a mixture of proteins enriched in molecules with low molecular mass in the range of 5–40 kDa . Toxins including hyaluronidase , proteases , low molecular mass insecticidal peptides , Translationally Controlled Tumor Protein ( TCTP ) and phospholipases-D have been identified [1] , [2] , [13] , [14] , [15] , [16] , [17] . The existence of hyaluronidases in Loxosceles venoms comes from a previous report by Wright et al . ( 1973 ) [18] , which reported hyaluronidase activity in the venom of L . reclusa . Additionally , hyaluronidase activity was described in the venom of L . rufescens [19] and several other Loxosceles venoms , including L . deserta , L . gaucho , L . intermedia , and L . laeta [11] , suggesting biological conservation and significance of these enzymes . Da Silveira and colleagues ( 2007 ) [20] identified the hyaluronidases in Loxosceles intermedia venom as endo-β-N-acetyl-D-hexosaminidases that degrade hyaluronic acid and chondroitin sulfate . The idea that brown spider venom hyaluronidases play a role in the gravitational spreading of dermonecrosis and/or systemic diffusion of venom toxins , and then act as “spreading factors” comes from its degradative activity on hyaluronic acid and other glycosaminoglycans that mediate tissue integrity and stability . By degrading glycosaminoglycans , the hyaluronidase reduces the viscosity of hyaluronic acid and renders the extracellular matrix less rigid . This change makes the matrix more permeable to other toxins and facilitates the spread of other venom constituents and inflammatory cell mediators [21] , [22] . Although hyaluronidase activity has been described in several venoms including that of snakes [10] , [23] , scorpions [24] spiders [22] , bees [25] , caterpillars [26] , wasps [27] , cone snails [28] and fish [29] , the biochemical and biological characterization of these enzymes is restricted to a few examples [10] , [21] , [30] . In the case of the brown spider , the major technical problem is the minute volume of venom obtained per animal . The total volume harvested after an electric shock on the cephalothorax of the spider is limited to one or two microliters , and contains a few tens of micrograms of protein [31] . Purification of this native glycosidase to homogeneity for additional biochemical determination has not yet described for brown spider venom , and would be a very difficult task . In recent years , using molecular biology techniques , scientists have obtained sufficient amounts of various recombinant toxins from brown spider venom to bring deeper insight into the molecular action of these toxins . Various venom phospholipase D isoforms and an Astacin-like metalloprotease have been reported for various species of spider venom [20] , [32] , [33] , [34] , [35] , [36] . Here , by using a cDNA library of L . intermedia venom glands [15] , we described the cloning , heterologous expression and functional evaluation of a novel hyaluronidase . The results bring insight into loxoscelism , opening up possibilities for biotechnological applications for this recombinant enzyme as a research tool . This recombinant hyaluronidase would also be useful in future structural and functional studies of this class of enzymes .
Salts and organic acids were purchased from Merck ( Darmstadt , Germany ) . Agar , β mercaptoethanol , molecular mass markers and purified hyaluronic acid were purchased from Sigma ( St . Louis , USA ) . Ethidium bromide , a Wizard Plus SV miniprep kit and pGEM-T vector were acquired from Promega ( Madison , USA ) . Agarose , IPTG and Trizol were purchased from Invitrogen ( Carlsbad , USA ) . We acquired DNA molecular mass standards , X-Gal , Taq DNA polymerase , Pfu DNA polymerase , T4 DNA ligase , restriction enzymes , dNTPs and CIAP from Fermentas ( Burlington , Canada ) . For bacterial culture , we used tryptone , yeast extract and agar purchased from HiMedia ( Mumbai , India ) . The antibiotics were purchased from USB ( Cleveland , USA ) . The bacterial strains used in this study and the ImPromII Reverse Transcription System kit were acquired from Invitrogen . We purchased the pET-14b expression plasmid from Novagen ( Novagen , Madison , USA ) . The glycosaminoglycan standards used were heparan sulfate from bovine pancreas [37] , chondroitin sulfate from bovine cartilage and dermatan sulfate from pig skin ( Seikagaku , Kogyo Co . , Tokyo , Japan ) . The venom gland cDNA library construction was performed as described by Gremski et al . ( 2010 ) [15] . Briefly , processed sequences were compared to GenBank sequences using the Basic Local Alignment Search Tools blastx , blastn ( E values<1e-05 ) and tblastx ( E values<1e-10 ) algorithms . Afterwards , ESTs were manually inspected for functional classification . One toxin-coding messenger similar to a hyaluronidase from Rattus novergicus ( gb|EDL77243 . 1 ) was found in this cDNA library and was used as a base sequence . To obtain the complete 5′ end of hyaluronidase cDNA , a 5′RACE ( Rapid Amplification of 5′cDNA Ends ) protocol was performed following Sambrook and Russel ( 2001 ) [38] with minor modifications . Briefly , 1 µg total RNA from L . intermedia venom glands was used as a template . The first-strand cDNA was synthesized using the gene-specific reverse primer R1 ( 5′-GTTGCAGGGTAGACAACATCCACG-3′ ) and the Improm-II Reverse Transcriptase ( Promega ) , according to the manufacturer's instructions . The cDNA was recovered by ethanol precipitation in the presence of ammonium acetate . The cDNA was poly ( A ) tailed with terminal deoxynucleotidyl transferase ( Fermentas ) , as recommended by the supplier . The modified cDNA was amplified using PCR with a ( dT ) 17 adaptor primer ( 5′-CGGTACCATGGATCCTCGAGTTTTTTTTTTTTTTTTTV-3′ ) and the nested gene-specific reverse primer R2 ( 5′-CTCCATGCTTCCCAGTCGATGATGC-3′ ) using a Pfu DNA polymerase ( Fermentas ) . Finally , the PCR product was purified from the gel using Illustra GFX PCR DNA and a Gel Band Purification kit ( GE ) following the manufacturer's instructions and sequenced on both strands using MegaBace DNA Analysis Systems ( Amersham Bioscience ) . To achieve the complete sequence , a 3′RACE ( Rapid Amplification of 3′cDNA End ) protocol was modified from Sambrook and Russel [38] . Briefly , 1 µg total RNA from L . intermedia venom glands was used to synthesize the first-strand cDNA using the gene-specific forward primer F1 ( 5′-CGAATCAATCAACGGTGGCATCCCTC-3′ ) and the Improm-II Reverse Transcriptase ( Promega ) . The cDNA was recovered as previously described . The cDNA was amplified with F2 ( 5′ -CCGCATTGGTTTTAGCCGCATTC-3′ ) , ( dT ) 17 adaptor primer ( 5′ -CGGTACCATGGATCCTCGAGTTTTTTTTTTTTTTTTTV-3′ ) and Pfu DNA polymerase ( Fermentas ) . Purification from the gel with a Gel Band Purification kit ( GE ) was performed according to manufacturer's instructions , and the amplicon was sequenced on both strands . The cDNAs encoding the putative mature hyaluronidase , which we named Dietrich's Hyaluronidase , were amplified with PCR using primers designed to contain Nde I restriction sites at the 5′ ends ( 5′-GGAATTCCATATGGACGTCTTCTGGAACG-3′ ) and BamH I sites ( 5′-CGGGATCCCTCACTTTGTTTTCTGCTC-3′ ) . The PCR product was digested with Nde I and BamH I restriction enzymes . Subcloning was performed with a pET-14b plasmid ( Novagen ) digested with the same enzymes . The recombinant construct for mature protein was expressed as a fusion protein with a 6× His-Tag at the N-terminus . The expression construct was inserted into E . coli BL21 ( DE3 ) pLysS cells and plated on LB-agar plates containing 100 µg/mL ampicillin and 34 µg/mL chloramphenicol . Single colonies of the construct were inoculated into LB broth ( 100 µg/mL ampicillin and 34 µg/mL chloramphenicol ) and grown overnight at 37°C . These cultures were diluted 1∶100 into 1 L fresh LB broth/ampicillin/chloramphenicol and incubated at 37°C until the OD550 nm = 0 . 4–0 . 6 . Recombinant protein expression was induced by the addition of 0 . 1 mM IPTG ( isopropyl β-D-thiogalactoside ) and cells were incubated for 3 . 5 h at 30°C in a shaker . Cells were harvested by centrifugation ( 4 , 000×g , 7 minutes , 4°C ) , suspended in 20 mL of extraction buffer ( 50 mM sodium phosphate pH 8 . 0 , 500 mM NaCl , 10 mM imidazole , 1 mg/mL lysozyme ) and frozen at −20°C overnight . Cells were thawed and disrupted with 8 cycles of sonication at medium intensity for 20 seconds using a 500-W ultrasonic cell disruptor . Lysed materials were centrifuged ( 20 , 000×g , 30 minutes , 4°C ) , and the pellet was washed with a denaturing buffer ( 100 mM Tris-HCl pH 10 . 0 , 2 M urea , 1% Triton X-100 ) and sonication at low intensity . After centrifugation at 6 , 000×g for 10 minutes the resulting pellet was solubilized in 100 mM Tris-HCl pH 10 . 0 with 8 M urea and 100 mM DTT . The solution containing denatured and reduced recombinant protein was adjusted to a concentration of ∼5 mg/mL and was added dropwise ( 1∶10 ratio ) to a refolding buffer ( 100 mM Tris-HCl pH 10 . 0 , 3 mM reduced glutathione , 0 . 3 mM oxidized glutathione , 0 . 4 M L-arginine , 0 . 2 mg bovine serum albumin ) by stirring over 16 h at 4°C . The protocol used was based on the works of Burgess et al . and Hofinger and co-workers [39] , [40] . Dialysis was against phosphate buffered saline , and recombinant hyaluronidase was concentrated using filter devices ( MWCO 30 , 000 Millipore , Schwalbach , Germany ) . Polyclonal antibodies against L . intermedia whole venom and against recombinant hyaluronidase were produced in rabbits as described by Harlow and Lane ( 1988 ) [41] , with minor modifications [42] . Protein concentration was determined using the Coomassie Blue method [43] or ultraviolet measurement ( 280 nm ) . For protein analysis , 12 . 5% SDS-PAGE was performed under reducing conditions and gels were stained with Coomassie Blue . For immunoblotting , proteins were transferred onto a nitrocellulose membrane following Towbin et al . ( 1979 ) [44] and immunodetected using hyperimmune antisera , which reacts with hyaluronidase or venom . Agarose gel electrophoresis was developed in 50 mM Tris–acetate buffer at pH 8 . 0 to evaluate hyaluronic acid degradation . Electrophoresis was performed in 50 mM 1 , 3-diaminopropane acetate buffer , pH 9 . 0 ( Aldrich , Milwaukee , USA ) to evaluate the cleavage activity of glycosidase on chondroitin sulfate , dermatan sulfate and heparan sulfate . After electrophoresis , compounds were precipitated in the gel using 0 . 1% cetavlon ( cetylammonium bromide ) for 2 h at room temperature . The gels were dried and stained with Toluidine Blue . The glycosaminoglycan standards used were as previously described [20] , [45] The Ethics Animal Experiment Commitee of the Setor de Ciências Biológicas of the Federal University of Parana , established by the decree 787/03-BL from May 9th 2003 , and upon the internal regiment , certifies that the procedures using animal in this work are in agreement with the Ethical Principals established by Experimental Animal Brazilian Council ( COBEA ) , and with the requirement of the “Guide for the Care and Use of Experimental Animals ( Canadian Council on Animal Care ) ” . Processes numbers: 23075 . 052088/2008-32 Approved: April 7 , 2009 and 23075 . 087106/2011-01 Approved: August 9 , 2011 . Adult rabbits ( ∼3 kg ) from the Central Animal House of the Catholic University of Parana were used for in vivo experiments with whole venom and recombinant enzymes ( the product of antibodies and dermonecrosis studies ) . All procedures involving animals were carried out in accordance with Brazilian Federal Law , following the Ethical Subcommittee on Research Animal Care from the Federal University of Parana . To evaluate the potential gravitational spreading effect of brown spider hyaluronidase , 10 µg of a recombinant dermonecrotic toxin ( LiRecDT1 ) diluted in PBS was injected intradermally into a shaved area of rabbit skin with or without recombinant hyaluronidase ( 10 µg ) . For the same purpose 10 µg of Dietrich's Hyaluronidase alone ( diluted in PBS with 0 . 2 mg/mL BSA ) was also injected . We used two negative controls: one was PBS with 0 . 2 mg/mL BSA , to assure that BSA used to refold recombinant hyaluronidase did not induce changes . The other was a recombinant protein with similar molecular mass obtained under the same conditions as hyaluronidase , but without hyaluronidase activity . The last assay guarantees that potential bacterial contaminants did not influence the results . Ten micrograms of venom and dermonecrotic toxin were used as positive controls . Rabbits were used in dermonecrosis experiments because this model reproduces skin lesions very close to those observed in accidents with humans [46] . Experiments were repeated with 4 animals and the development of experimentally induced dermonecrosis was observed 3 h , 6 h and 24 h after the injection . Rabbit skin pieces from animals that intradermally received recombinant proteins were collected following anesthetization with ketamine ( Agribands , Campinas , Brazil ) and acepromazine ( Univet , São Paulo , Brazil ) . Collected tissue samples were fixed in “ALFAC” ( ethanol 85% , formaldehyde 10% and glacial acetic acid 5% ) for 16 h at room temperature . After fixation , samples were dehydrated in a graded series of ethanol before paraffin embedding ( for 2 h at 58°C ) . Thin sections of 4 µm thickness were processed for histological procedures [42] . Tissue sections were stained with Masson's trichrome ( TM ) as described [47] . To compare new sequences generated against the GenBank database , we used the BLAST site ( http://blast . ncbi . nlm . nih . gov/Blast . cgi ) . The parameters for blastn were refined by selecting the “others” search set with E values <1e-05 . For blastx , we chose standard genetic code 1 and non-redundant protein sequences with E values <1e-05 parameter . Deduced amino acid sequence was found using the Open Reading Frame Finder site ( http://www . ncbi . nlm . nih . gov/projects/gorf/ ) and an analysis of protein parameters and glycosylation modifications was conducted using the ExPASy Bioinformatics Research Portal ( http://web . expasy . org/protparam/ ) . Disulfide bond prediction was made using the DiANNA 1 . 1 web server http://clavius . bc . edu/~clotelab/DiANNA/ . CLUSTAL W ( http://www . ebi . ac . uk/Tools/msa/clustalw2/ ) was used for alignment and cladogram production . Finally , to search any possible signature recognition within Dietrich's Hyaluronidase sequence , we used the InterProScan site ( http://www . ebi . ac . uk/Tools/pfa/iprscan/ ) . The Dietrich's Hyaluronidase cDNA sequence has been submitted to the Genbank database under accession number JX402631 .
By screening clones of a cDNA library of the L . intermedia venom gland , a cDNA encoding for a hyaluronidase was obtained through a blastx search [15] . The putative protein product from this cDNA was designated Dietrich's Hyaluronidase ( in honor and memoriam of Professor Carl Peter von Dietrich , born in 1936 , deceased in 2005 ) . The complete cDNA sequence of Dietrich's Hyaluronidase consist of 1200 bp with a single open reading frame ( ORF ) coding for 400 amino acids and a putative N terminal endoplasmic reticulum import signal of 19 residues . At least two types of post translational modification were observed: 4 putative N-glycosylation sites and 3 possible disulfide bonds ( Fig . 1 ) . The predicted molecular mass for mature Dietrich's Hyaluronidase protein was approximately 44 . 8 kDa , and its pI was 8 . 75 . To explore the structural and evolutionary relationships among the newly identified glycosidase and other members of the hyaluronidase family , a BLAST GenBank database search through alignment of cDNA-deduced amino acid sequences revealed that Dietrich's Hyaluronidase has structural similarity to other hyaluronidase family members ( Fig . 2A ) . The overall identity of hyaluronidase from L . intermedia venom is approximately 46% compared with the scorpion venom hyaluronidase of Mesobuthus martensii ( gb|ACY69673 . 1 ) . Dietrich's Hyaluronidase is also similar to snake venom hyaluronidases , sharing 33% sequence identity with Echis pyramidium ( gb|ABI33941 . 1 ) and Cerastes cerastes ( gb|ABI33939 . 1 ) . Interestingly , the enzyme from an arthropod venom showed approximately 30% amino acid identity with hyaluronidases from mammal species such as Bos taurus ( gb|AAP55713 . 1 ) and Rattus novergicus ( gb|EDL77243 . 1 ) ( Fig . 2B ) . An InterProScan search matches Dietrich's Hyaluronidase with more than 500 proteins belonging to the glycoside hydrolase 56 family . A typical signature domain for hyaluronidases has not yet been demonstrated , and there are few studies about of the residues involved in catalysis [30] , [48] . However , the hyaluronidase from L . intermedia venom has 3 conserved amino acids ( indicated in Fig . 2A at arrows: D118 , E120 , E259 ) that appear to be important for the catalytic activity of some hyaluronidases [30] . Recombinant hyaluronidase from L . intermedia venom was expressed in pET 14b in E . coli BL21 ( DE3 ) pLysS cells . The expression of recombinant protein was optimized when induced for 3 . 5–5 h with 0 . 1 mM of IPTG . The recombinant protein was detected only in the insoluble fraction of cell lysates and was purified under denaturing conditions by washing inclusion bodies until there were no bacterial contaminants visible in SDS-PAGE gels stained with Coomassie blue . The SDS-PAGE mobility of recombinant protein reduced by β-mercaptoethanol treatment was ∼45 kDa , consistent with the calculated molecular mass ( Fig . 3 ) . After testing several refolding protocols in vitro ( data not shown ) , the protocol described in the methods section , was effective in solubilizing recombinant hyaluronidase from inclusion bodies , as shown in lane 4 ( Fig . 3 ) . Immunoblot analysis using antibodies against recombinant hyaluronidase ( Dietrich's Hyaluronidase ) and antibodies against whole venom toxins established the epitope relationship of native hyaluronidase with recombinant enzyme ( Fig . 4 ) . Antibodies against Dietrich's Hyaluronidase recognize hyaluronidase in the whole venom , as well as anti-whole venom serum reacted with recombinant glycosidase . To evaluate the activity of Dietrich's Hyaluronidase after refolding , purified HA was incubated at a ratio of 1∶1 with recombinant hyaluronidase for 3 h , 6 h and 16 h at 37°C . The enzymatic activity was analyzed through agarose gel electrophoresis stained with toluidine blue . HA was completely degraded within 3 h ( Fig . 5A , lane 3 ) showing that the in vitro refolding method was effective in adjusting hyaluronidase in an active conformation . Through analysis with CS , DS and HS at a substrate-to-enzyme ratio of 1∶1 , we noted that recombinant hyaluronidase degraded CS ( Fig . 5B , lane 3 ) and did not degrade DS and HS ( data not shown ) , similar to the native hyaluronidases from venom . To verify the stability of refolding in vitro achieved with recombinant hyaluronidase , we performed a zymogram assay using gels containing copolymerized purified HA and CS as substrates . Hydrolytic activity was found as a specific band between the 29–45 kDa regions in both substrates and strengthened the results described above ( Fig . 5C ) . We sought to develop dermonecrosis experiments using rabbit skin to determine the in vivo involvement of brown spider venom hyaluronidase in envenomation . Moreover , we aimed to observe the relationship of hyaluronidases in the skin deleterious activities of whole venom . Injections of L . intermedia crude venom , recombinant dermonecrotic toxin [33] ( positive controls ) , and dermonecrotic toxin plus Dietrich's Hyaluronidase all induced macroscopic erythema , ecchymosis and dermonecrosis in rabbit skin , which was followed by 3 , 6 and 24 h of observation ( Fig . 6A and Fig . 6B ) . On the other hand , the injection of recombinant hyaluronidase alone , PBS/BSA and a recombinant protein not related to hyaluronic acid hydrolysis ( negative controls ) did not show macroscopic erythema , ecchymosis or dermonecrosis ( Fig . 6A and Fig . 6B ) . The length of dermonecrosis lesions in the rabbit skin of Dietrich's Hyaluronidase concomitantly injected with dermonecrotic toxin was substantially larger than for dermonecrotic toxin alone after 24 h ( Fig . 6B ) . The macroscopic gravitational spreading and edema induced by the two recombinant enzymes injected together were very similar to whole venom ( Fig . 6C ) . Light microscopic analysis of rabbit skin biopsies 24 h after of the experimentally induced dermonecrosis experiments performed above suggested that recombinant hyaluronidase was able to disorganize the extracellular matrix from rabbit skin dermis ( Fig . 7B-4 ) . Dermonecrotic toxin alone triggered typical inflammatory cell accumulation ( Fig . 7B-2 and 7B-5 ) and edema signals ( Fig . 7A , 7B-5 ) , as well as collagen fiber disorganization ( Fig . 7B-5 ) and presence of fibrin in connective tissue ( Fig . 7B-5 ) . Nevertheless , we noticed that these inflammatory events , developed by dermonecrotic toxin , were intensified when combined with Dietrich's Hyaluronidase ( Fig . 7A; 7B-3 and 7B-6 ) , supporting an extensive edema and spreading of lesion . By comparison , using a panoramic image under the same conditions , deleterious effects induced by dermonecrotic toxin alone were lower than those evoked by mixture of this toxin and Dietrich's Hyaluronidase ( Fig . 7A ) . These microscopic findings corroborate the information obtained by macroscopic analysis of rabbit skin lesions .
Hyaluronidases are present in various tissues and are involved in important biological events , including embryonic development , inflammation , fertilization , tumor cell metastasis , bacterial pathogenesis and aging [10] , [14] . These enzymes were described many years ago in the venom of different Loxosceles species , featuring a conservative event and suggesting biological significance in the life cycle of these spiders [1] , [2] , [11] . A more refined understanding of brown spider venom hyaluronidase is hampered by native molecules present in low amounts in the venom and by extremely minute volume of venom obtained from spiders [2] , [15] , [49] . Here , we described the cloning , heterologous expression , purification , refolding and characterization of a novel hyaluronidase from a cDNA library of the L . intermedia venom gland . These results corroborate previous biochemical data that have described these enzymes in brown spider venom from different species [11] , [20] . The recombinant protein identified herein is designated Dietrich's Hyaluronidase ( GenBank Accession no . JX402631 ) . The primary sequence of Dietrich's Hyaluronidase includes a hydrophobic signal sequence of 19 residues and a mature protein . The hydrophobic signal sequence probably directs its expression to the endoplasmic reticulum in venom gland epithelial cells . The molecular mass and isoelectric point calculated from the deduced amino acid sequence of mature protein were 44 . 8 kDa and pI 8 . 75 , respectively . The theoretical molecular mass is within the expected range , considering the size of brown spider venom native hyaluronidases [11] , [20] , and the isoelectric point was similar to the calculated pI of hyaluronidases from other venoms . A protein sequence analysis of brown spider venom hyaluronidase compared to hyaluronidases from different sources showed that this venom enzyme has conserved amino acids ( D118 , E120 , E259 ) that seem to be important for catalytic hydrolysis [30] , [48] . Dietrich's Hyaluronidase was assigned to the Glycosidase 56 family; which compose a wide group of O-glycosyl-hydrolases [50] . The highest similarity to Dietrich's hyaluronidase was found in the sequences of scorpion hyaluronidase from Mesobuthus martensii venom ( 46% amino acid identity ) and the snake venom of Echis pyramidium and Cerastes cerastes ( 33% ) . As might be expected , the hyaluronidase from L . intermedia venom was particularly similar to hyaluronidases from other venoms . However , we highlight Dietrich's Hyaluronidase's significant identity with mammalian species such as Bos taurus and Rattus novergicus . This fact , coupled with its activity at physiological pH , would be interesting for pharmaceutical use . Dietrich's Hyaluronidase was heterogeneously expressed as a mature protein with an N-fusion His-tag using the E . coli BL21 ( DE3 ) pLysS strain . Isolation of the recombinant glycosidase was performed by washing inclusion bodies with denaturing buffer . The purification of recombinant proteins from inclusion bodies might be advantageous , because it is protease-resistant and is close to functional native structure [51] , [52] , [53] . Characterization of the antigenic cross-reactivity of recombinant hyaluronidase and L . intermedia venom showed that the venom contains proteins that have antigenic identity ( sequence epitopes ) with recombinant hyaluronidase because antisera raised against recombinant enzyme was able to react with native venom . Two protein bands of the whole venom were recognized by hyaluronidase antisera , corroborating with previous results that showed two lytic zones in zymography experiments using whole venom in a gel co-polimerized with hyaluronic acid . Those bands probably correspond to two isoforms of hyaluronidase present in L . intermedia whole venom [11] , [20] . Additionally , anti-venom cross-reacted with the recombinant hyaluronidase , suggesting that the recombinant glycosidase retains linear antigenic determinants from native hyaluronidases . In this way , the antisera to Dietrich's Hyaluronidase or purified antibodies from it could be a possible biological tool in loxoscelism therapy or for research purposes . The functionality of refolded recombinant hyaluronidase was demonstrated through its activity upon purified hyaluronic acid and chondroitin sulfate . Dietrich's Hyaluronidase was able to directly degrade hyaluronan and chondroitin sulfate . These results are in agreement with previous data reported for native hyaluronidases from whole venom , which degrade both glycosaminoglycans [20] . The hydrolytic activity found at 29–45 kDa in zymogram assays copolymerized with hyaluronic acid and chondroitin sulfate corroborates with the degrading activity of glycosaminoglycans viewed in agarose gels , suggesting certain stability in the active folding of Dietrich's Hyaluronidase acquired after refolding in vitro . Again , results show that recombinant hyaluronidase can also be considered a chondroitinase , as previously reported for native hyaluronidases [20] . Regarding in vitro refolding it is worth mentioning that buffers with different pHs , redox agents and additives were tested [39] , [54] until the solubility and activity of Dietrich's Hyaluronidase was maximized ( data not shown ) . When bovine albumin was removed from the buffer , the hyaluronidase activity was null . This result is consistent with the literature that describes how BSA may compete with hyaluronidases to form inactive electrostatic complexes with hyaluronic acid . This competition induces free hyaluronidase resulting in a large increase in the hydrolysis rate [55] . Several research groups have reported that serum proteins are able to enhance hyaluronidase activity and are sometimes required to detect the presence of hyaluronidases [55] , [56] , [57] . But what is the role of brown spider venom hyaluronidase on bite pathology ? The involvement of hyaluronidase on the activity of brown spider venom is supported by the conservative phenomenon of this enzyme , which indeed is found in different species of Loxosceles venoms [11] . Hyaluronidases have been described for several venoms [26] , [55] , [58] and act as “spreading factors” by degrading hyaluronic acid and chondroitin sulfate . In this way , venom hyaluronidases may render surrounding regions at the bite site more permeable . Furthermore , these enzymes may facilitate the diffusion of other venoms constituents through the bodies of victims [10] , [58] . A hyaluronidase isolated from funnel web spider venom was able to enhance the potency of a myotoxin and a hemorrhagic toxin . This work is in accordance with the hypothesis that hyaluronidase mediates enhanced toxicity of whole venom during envenomation [22] . A typical symptom of loxoscelism is the gravitational spreading of skin lesions that appear a few hours after bites or experimental envenomation of animal models [1] , [2] . The mechanism by which Loxosceles venom causes gravitational spreading of dermonecrosis is not fully understood . Experimental inoculation of purified recombinant dermonecrotic toxins ( phospholipases-D , non-proteolytic or hyaluronidase activities ) , evokes gravitational spreading of skin lesion on rabbits [34] . Macroscopically , Dietrich's Hyaluronidase increases the erythema , ecchymosis and dermonecrotic lesion area induced by recombinant phospholipase-D along exposure times observed . Histopathological findings for the dermonecrotic toxin exposed tissue samples revealed , as expected , the presence of a neutrophilic infiltrate , collagen fiber disorganization and signs of edema [36] , [46] . However , when we analyzed tissue samples treated with phospholipase-D toxin plus Dietrich's Hyaluronidase , we observed that these inflammatory evidences were much more intense , as if recombinant hyaluronidase was allowed a free diffusion of phospholipase D by rendering the connective tissue more permeable . For the first time in the literature , experimental results strongly indicate that the hyaluronidase from Loxosceles venom is in fact a “spreading factor” for this venom . For brown spider venom hyaluronidase , based on the spreading property , we extrapolated that this molecule would be primarily responsible for spreading other venom toxins from the bite site into the systemic circulation [1] , [2] . Literature data have reported signs of systemic intoxication including fever , vomiting , hemolytic anemia , thrombocytopenia , disseminated intravascular coagulation , and nephrotoxicity following brown spider bites [1] , [2] , [4] , [31] . Moreover , brown spider venom hyaluronidase is also suggested to play a role in the cutaneous lesions following bites , as described for other venom hyaluronidases [23] . By disturbing the extracellular matrix structure , venom hyaluronidase may influence the stability of blood vessel walls . These events may increase the spread of other venom toxins , which in turn can cause the cutaneous lesions that may follow after brown spider bites . Corroborating previous ideas , our work shows that Dietrich's Hyaluronidase is able to degrade hyaluronic acid and chondroitin sulfate in vitro and has increased the erythema and ecchymosis cause by phospholipase-D injected into rabbit skin . We suggest that this enzyme likely degrades both glycosaminoglycans in vivo . It is known that low molecular weight HA fragments are pro-inflammatory , immunostimulatory and angiogenic . Besides , HA oligosaccharides formed due to plasma HA degradation may result in hemostatic disturbances . Hence , exploring the in vivo fragmentation of HA and the effects on pathophysiology of envenomation might be a topic for further researches [59] . Another important feature that has been described for hyaluronidases of venoms , such as from scorpions , bees , hornets and wasps , is their classification as major allergens that can induce anaphylaxis and sometimes death [21] , [60] . Based on its sequence identity with other venom hyaluronidases , it is possible that the brown spider venom hyaluronidase may act as a potential allergen in susceptible individuals . This notion is strengthened by evidence observed in the course of loxoscelism , which includes itch , morbilliform erythema , cutaneous rash and petechial eruption [1] , [2] , [31] , [61] , [62] . It might also suggest the involvement of the immune system [63] , [64] . Cutaneous rashes respond to treatment with systemic steroids [31] , [63] . Thus , the hyaluronidase from Loxosceles genus could be further investigated as a molecule capable of inducing allergy reactions . In the same way , whether allergy's molecular mechanism is due to the enzyme molecule itself or even the HA fragments resulting from its glycosidase activity . Besides a novel understanding of the pathogenesis of loxoscelism , Dietrich's Hyaluronidase could be an important tool for future biotechnological purposes [14] , [16] . Hyaluronidases are known to be involved in physiological and pathological processes such as bacterial pathogenesis , spread of toxins and venoms , fertilization , and cancer progression [65] , [66] , [67] , [68] . For example , hyaluronidase recombinant molecules may be developed as tools for in vitro fertilization [10] . On the other hand , hyaluronidase inhibitors may serve as contraceptives , because they are involved in the fertilization of eggs by mammalian sperm and could thus be used to block fertilization . Other possible applications of hyaluronidase inhibitors are as anti-tumor [58] , [66] , anti-bacterial [67] and anti-venom/toxin agents [16] , [20] . Interestingly , a cloned Buthus martensi hyaluronidase ( BmHYA1 ) down-regulated CD44 ( a hyaluronic acid-ligand transmembrane glycoprotein involved in cell–matrix connections ) in a cancer cell line , suggesting that a variant of cancer cells can be modulated by external venom hyaluronidase treatment [24] . In summary , for the first time in the literature , we have described a recombinant hyaluronidase from the venom glands of Loxosceles sp . We cloned , expressed , purified and refolded this enzyme , which showed degradative activity on hyaluronic acid and chondroitin sulfate . This recombinant glycosidase increased the area of dermonecrosis , gravitational spreading and edema induced by a recombinant dermonecrotic toxin , mimicking the profile of whole venom in an animal model for skin loxoscelism in vivo . Finally , results showed the role of this enzyme as a spreading factor in the mechanism of spreading of necrotic lesions followed by brown spider envenomation . Together , these results provide insights into loxoscelism and contribute to a further understanding of venom mechanisms and will perhaps unveil novel treatment protocols for envenomation or biotechnological applications for this venom protein . We propose naming this novel brown spider hyaluronidase described herein as Dietrich's Hyaluronidase , in memoriam and honor of Professor Carl Peter Von Dietrich . Professor Dietrich was born in 1936 in Rio de Janeiro , Brazil , graduated in medicine from the Universidade do Estado do Rio de Janeiro , 1963 , Rio de Janeiro , Brazil , specialized in biochemistry at the Instituto de Investigaciones Bioquimicas , 1963 , Buenos Aires , Argentine , and earned his doctorate at the University of Saskatchewan , 1970 , Saskatchewan , Canada . He was a professor and researcher in biochemistry and molecular biology at the Universidade Federal de São Paulo , São Paulo , Brazil , where he worked with macromolecules such as proteoglycans , glycosaminoglycans and heparins . Professor Dietrich died in 2005 in São Paulo , Brazil .
|
Accidents involving brown spiders ( Loxosceles genus ) are reported throughout the world . South and Southeast of Brazil are endemic areas for this spider . Loxosceles bites commonly trigger local signs as swelling , erythema , hemorrhage and the hallmark symptom: a dermonecrotic lesion with gravitational spreading . Systemic effects are less common; however , are implicated in more severe cases . Hyaluronidases are referred in several venoms as “spreading factors” due to their enzymatic activity upon extracellular components . This activity facilitates the permeation of other toxins through the victim's body . In fact , a previous study identified the activity of L . intermedia venom upon glycosaminoglycans which are abundant components in the extracellular matrix of many tissues . Disclosing a little more about the role of hyaluronidases within this venom , we investigated the activities of a recombinant hyaluronidase from L . intermedia venom . Dietrich's hyaluronidase , as it was designated , was produced as a recombinant protein . By performing a rabbit skin dermonecrosis assay using Dietrich's Hyaluronidase and a dermonecrotic toxin , we showed that Dietrich's Hyaluronidase increased the dermonecrotic area induced by the dermonecrotic toxin . Our results confirm that hyaluronidases are a “spreading factor” of L . intermedia venom .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"sequencing",
"medicine",
"dermatology",
"inflammatory",
"diseases",
"enzymes",
"immunology",
"toxicology",
"toxic",
"agents",
"glycobiology",
"histology",
"proteoglycans",
"glycoproteins",
"sequence",
"analysis",
"proteins",
"extracellular",
"matrix",
"connective",
"tissue",
"biology",
"recombinant",
"proteins",
"proteomics",
"immune",
"response",
"immunopathology",
"biochemistry",
"molecular",
"cell",
"biology",
"protein",
"engineering"
] |
2013
|
A Novel Hyaluronidase from Brown Spider (Loxosceles intermedia) Venom (Dietrich's Hyaluronidase): From Cloning to Functional Characterization
|
Unlike most bacteria , Vibrio cholerae harbors two distinct , nonhomologous circular chromosomes ( chromosome I and II ) . Many features of chromosome II are plasmid-like , which raised questions concerning its chromosomal nature . Plasmid replication and segregation are generally not coordinated with the bacterial cell cycle , further calling into question the mechanisms ensuring the synchronous management of chromosome I and II . Maintenance of circular replicons requires the resolution of dimers created by homologous recombination events . In Escherichia coli , chromosome dimers are resolved by the addition of a crossover at a specific site , dif , by two tyrosine recombinases , XerC and XerD . The process is coordinated with cell division through the activity of a DNA translocase , FtsK . Many E . coli plasmids also use XerCD for dimer resolution . However , the process is FtsK-independent . The two chromosomes of the V . cholerae N16961 strain carry divergent dimer resolution sites , dif1 and dif2 . Here , we show that V . cholerae FtsK controls the addition of a crossover at dif1 and dif2 by a common pair of Xer recombinases . In addition , we show that specific DNA motifs dictate its orientation of translocation , the distribution of these motifs on chromosome I and chromosome II supporting the idea that FtsK translocation serves to bring together the resolution sites carried by a dimer at the time of cell division . Taken together , these results suggest that the same FtsK-dependent mechanism coordinates dimer resolution with cell division for each of the two V . cholerae chromosomes . Chromosome II dimer resolution thus stands as a bona fide chromosomal process .
Vibrio cholerae , the causative agent of cholera , harbors two non-homologous circular chromosomes [1] . The majority of genes believed to be necessary for the basic life processes of V . cholerae are carried on the 2 . 96 Mbp chromosome I , whereas the 1 . 07 Mbp chromosome II only harbors a few essential genes [1] . The preferential transcription of genes from chromosome II during colon colonization [2] suggests that this genomic organization is important for pathogenicity . Likewise , other bacteria with multiple chromosomes can adopt several different life cycles [3] , which led to the idea that multipartite genomes offer a selective advantage for the adaptation to very different environmental conditions . Nevertheless , most bacteria harbor a single chromosome . In contrast , there is no apparent limit to the size and numbers of chromosomes harbored by eukaryotic cells . An important difference between bacteria and eukaryotes is that specific machineries appear to exist for the coordinated maintenance of each chromosome of a given bacterium , whereas eukaryotic cells possess a single global system for all chromosomes [4]–[12] . For instance , the two V . cholerae chromosomes harbor different partition systems [4] , [7] and initiation of their replication is governed by different mechanisms [6] , [8] , [9] . In addition , many features of V . cholerae chromosome II , such as its partition system , are plasmid-like , which raised questions concerning its chromosomal nature [1] , [10] , [13] . Plasmid replication and segregation are generally not coordinated with the bacterial cell cycle [14] , further raising questions on the mechanisms ensuring the synchronous management of chromosome I and II . A second major difference between bacteria and eukaryotes is intrinsic to the structure of chromosomes: in bacteria , chromosomes are generally covalently closed circular DNA molecules while they are linear in eukaryotes . DNA circularity can result in the formation of chromosome dimers by homologous recombination [15] , which poses a barrier to the segregation of genetic information if they are not resolved before cell division ( Figure 1A ) . Indeed , inactivation of chromosome dimer resolution ( CDR ) in Escherichia coli results in ∼15% cell death per generation under laboratory growth conditions [16] , which corresponds to the estimated rate of chromosome dimers formed at each cell generation [17] . This prompted us to study how dimer resolution is achieved on each of the two V . cholerae chromosomes . The mechanism of CDR was originally elucidated in E . coli . In this organism , it depends on the addition of a crossover at dif , a 28bp site located at the opposite of the origin of replication on the chromosome , by two related tyrosine recombinases , XerC and XerD ( Figure 1A; see [18] for a review ) . In addition , CDR depends on two activities of a cell division protein , FtsK . First , FtsK functions as a DNA pump anchored in the septum [19] , [20] . It loads on DNA trapped within the division septum due to dimer formation ( Figure 1A ) . FtsK loading is oriented by specific DNA motifs , the KOPS , which dictates the orientation of translocation ( Figure 1A; [21] ) . KOPS are skewed on the two replichores of the chromosome with dif located at the junction of their polarity [22] , [23] . Thus , dif sites carried by a dimer are brought together by FtsK translocation ( Figure 1A ) . Second , FtsK serves to activate recombination at dif via a direct interaction with XerD [24] , [25] . dif contains two 11bp binding sites for XerC and XerD , separated by a central region at the outer boundary of which recombination occurs . The interaction between XerD and FtsK allows XerD to perform a first pair of strand exchanges [20] , resulting in the formation of a Holliday junction ( HJ ) . This HJ is converted to a crossover by a second pair of strand exchanges , which is catalyzed by XerC independently of FtsK ( Figure 1B , chromosomal pathway ) . Thus , in E . coli , the requirement for FtsK to bring dif sites together and to activate the catalytic activity of XerD permits coordination of CDR with the last stage of cell division [26] . The E . coli pathway of CDR is not universal . For instance , Streptococci and Lactococci possess only a single tyrosine recombinase , XerS , for CDR [27] . Plasmid and viruses have also adopted different site-specific recombination systems to avoid multimerization of their genome . In E . coli , some of them depend on their own recombinases , such as phage P1 , which encodes the Cre tyrosine recombinase [28] , while others use the two Xer recombinases of their host [29] , [30] . In the later case , XerC-catalysis initiates recombination independently of FtsK ( Figure 1B , plasmid pathway , [18] ) . In this case , however , recombination requires ∼200bp of accessory sequences flanking the plasmid sites and which are bound by accessory proteins . Orthologues of E . coli xerC , xerD and ftsK are readily identified on the larger chromosome of the V . cholerae strain N16961 ( xerCVc , xerDVc , ftsKVc , Figure S1 ) whereas its second chromosome does not encode any site-specific recombination system that could be implicated in CDR apart from the superintegron integrase ( IntIA , Figure S1 ) . N16961 chromosome I and II both carry dif-like sequences , dif1 and dif2 , which were originally identified as integration sites for the Cholera Toxin phage , CTXφ [31] . The weak filamentous phenotype of V . cholerae cells deleted for xerC or dif1 fits with a defect in CDR [31] . However , two features of the V . cholerae Xer recombination system , which could be linked to the co-existence of distinct , non-homologous chromosomes inside the same bacterium , were intriguing . First , dif2 differs from the dif consensus of γ-Proteobacteria by 5 bases , four of which belong to the central region ( Figure 1B ) . Such a divergence is only found on plasmid sites , which , coupled with the other plasmid-like features of chromosome II , suggested that chromosome II dimer resolution might follow a plasmid pathway . Second , it was reported that the position of cleavage of XerDVc on dif1 might differ from the one of its E . coli orthologue on dif [32] , even if dif1 differs from the dif consensus of γ-Proteobacteria by only 2 bases ( Figure 1B ) , further raising questions on the exact mechanisms coordinating CDR of chromosome I and II with the cell cycle . Here , we present the first formal study of CDR in V . cholerae and measure the rate of chromosome dimer formation on its two chromosomes under laboratory growth conditions . We show that the cell division protein FtsKVc is required for recombination by XerCVc and XerDVc at dif1 and dif2 . In addition , we show that the activity of FtsKVc is directed by specific DNA motifs , which display the same skewed distribution on the two chromosomes , dif1 and dif2 being located at the junction of their polarity . Taken together , these results suggest that the same FtsK-dependent mechanism coordinates dimer resolution on each of the two V . cholerae chromosomes with cell division . Chromosome II dimer resolution thus stands as a bona fide chromosomal process .
The growth of V . cholerae strains deficient in CDR was directly compared to the growth of their parental strain in competition experiments in rich media ( Figure 2 ) . These experiments revealed a defect of 5 . 8% and 3% per cell per generation for Δdif1 and Δdif2 cells , respectively , compared to their wild type counterparts . Since these growth defects were entirely suppressed in a recA background ( Figure 2 ) , they directly reflect the rates of dimer formation on chromosome I and II , fdimerChr1 and fdimerChr2 ( See Material and Methods ) . The 8 . 6% growth defect of xerCVc cells , which was also suppressed in a recA background , reflects the total rate of chromosome dimer formation in V . cholerae , fdimerChr1+2 ( Figure 2 ) . Interestingly , fdimerChr1+2 equals 1− ( 1−fdimerChr1 ) ( 1−fdimerChr2 ) , indicating that dimer formation on the two V . cholerae chromosomes is independent . Recombinase-mediated strand cleavage can be assayed in vitro using suicide substrates that contain a nick opposite of the position of cleavage ( Figure 3A ) . Cleavage of the continuous strand of a suicide substrate generates a double strand break that prevents re-ligation ( Figure 3B ) . This leads to ( i ) the accumulation of covalent protein/DNA complexes between the attacking recombinase and the 5′-end fragment of the continuous strand and ( ii ) the accumulation of free 3′-end fragments of the continuous strand ( Figure 3B ) . XerCEc and XerDEc each cleave a specific strand on difEc . The strand cleaved by XerCEc is termed Top strand . The strand cleaved by XerDEc is termed Bottom strand . Following this convention , suicide substrates in which the continuous strand is expected to be cleaved by XerCVc are called Top strand suicide substrates and suicide substrates in which the continuous strand is expected to be cleaved by XerDVc are called Bottom strand suicide substrates ( Figure 3A ) . Labeling the 5′-end of the continuous strand of suicide substrates allows the detection of covalent recombinase/DNA complexes ( Figure 3C ) . The molecular weight of XerCVc and XerDVc being very similar , we used a maltose binding protein fusion of XerCVc ( MBPXerCVc ) in conjunction with XerDVc to avoid any confusion between the two possible covalent complexes . For both dif1 and dif2 , MBPXerCVc-DNA covalent complexes accumulated when Top strand suicide substrates were used ( Figure 3C , T1 and T2 , respectively ) , indicating that XerCVc cleaves the Top strands of dif1 and dif2 . Furthermore , XerDVc-DNA covalent complexes accumulated when Bottom strand suicide substrates were used ( Figure 3C , B1 and B2 ) , indicating that XerD cleaves the bottom strands of dif1 and dif2 . The position of cleavage of XerCVc and XerDVc were then determined by comparing of the length of the free DNA fragments liberated by recombinase cleavage to a ladder obtained by chemical cleavage at purine bases of the suicide substrates ( Figure 3D ) . To this aim , the continuous strands of the suicide substrates were labeled on their 3′ end . Cleavage by tyrosine recombinases generates a 5′OH DNA extremity whereas chemical cleavage leaves a 5′ phosphate . As a consequence , the free DNA fragments had to be first phosphorylated by kinase treatment ( Figure 3D , PNK ) in order to be compared with the chemical cleavage ladder ( Figure 3D , G+A ) . We thus found that XerCVc and XerDVc cleave DNA at the junction between their respective binding site and the central region of dif1 and dif2 ( Figure 3D , black arrows ) . Analysis of the DNA sequence immediately upstream and downstream of dif1 and dif2 in different Vibrio species did not reveal any conserved motifs that could serve to bind accessory proteins ( data not shown ) . FtsKVc was thus left as the most likely candidate for activation of Xer recombination at both sites . To test this possibility , we reconstituted the V . cholerae Xer system in E . coli cells deleted for their natural FtsK/XerCD system . We used a xerC and xerD E . coli strain , which was also ftsKC− . This strain produces only the N-terminal domain of FtsKEc , essential for viability [33] , but lacks production of the C-terminal domain of FtsKEc , which is necessary for recombination at difEc [34] . XerCVc was expressed in conjunction with XerDVc from the chromosomal E . coli xerC promoter . The production of FtsKVc was controlled by placing the full length ftsKVc ORF under an arabinose-inducible promoter on a high-copy number plasmid . A low-copy plasmid carrying two recombination sites in direct repeats was used as a reporter . Recombination between the two repeated sites results in the excision of the intervening DNA , which can be monitored by agarose gel electrophoresis . For both dif1 and dif2 , the amount of recombination correlated with the amount of arabinose used for induction , indicating that Xer recombination at dif1 and dif2 depends on FtsKVc ( Figure 4A ) . To determine the order of the strand exchanges in the recombination reactions , we monitored plasmid recombination in a set of four strains encoding either wild-type XerCVc and XerDVc or the XerCYFVc and XerDYFVc mutants , in which the catalytic tyrosine is replaced by a phenylalanine ( Figure 4B ) . For both dif1 and dif2 , no resolution product or HJ intermediate were detected in XerDYFVc cells ( Figure 4B , lane 2 , 4 , 6 and 8 ) . In contrast , we could detect the accumulation of a HJ intermediate in XerCYFVc XerDVc cells ( Figure 4B , lane 3 and 7 ) , indicating that XerDVc mediates the first pair of strand exchanges during both dif1 and dif2-recombination . Recombination products were likely still observed in XerCYFVc XerDVc cells since other cellular processes than Xer recombination are capable of resolving HJs [18] . However , the amount of product was considerably decreased , indicating that intermediate HJs are preferentially resolved to crossovers by the action of XerCVc . All together , these results indicate that FtsKVc activates recombination at dif1 and dif2 by promoting the exchange of a first pair of strands by XerDVc . Several residues implicated in the interaction between E . coli XerD and FtsK have been mapped [24] , [25] . These residues are not entirely conserved between the V . cholerae and E . coli proteins ( Figure 5A ) , suggesting that the interactions between the translocase and the recombinases might be specific in these two species . Nevertheless , both FtsKEc and FtsKVc could activate recombination by XerCDEc and XerCDVc at difEc , dif1 and dif2 ( Figure 5B ) . However , the efficiency of recombination varied for each site and for each pairing of translocase/recombinases . XerCDEc-recombination at difEc and dif1 reached 80% of efficiency whether FtsKEc or FtsKVc were produced ( Figure 5B , XerCDEc , dif1 and difEc ) . In contrast , XerCDEc-recombination at dif2 was more efficient when activated by FtsKEc than FtsKVc ( Figure 5B , XerCDEc , dif2 ) . In addition , it did not reach 80% efficiency , even in the presence of the cognate partner translocase , FtsKEc . XerCDVc-recombination at difEc , dif1 and dif2 reached 80% of efficiency ( Figure 5B , XerCDVc ) . However , this required the presence of FtsKVc . XerCDVc-recombination at dif2 even fell below 20% when activated by FtsKEc . Thus , the effect of species-specificity is more pronounced on dif2 than on dif1 . We noticed that the V . cholerae recombinases could promote recombination between dif1 sites in the absence of FtsK production , albeit to a very low level ( Figure 5B , XerCDVc , dif1 , No FtsK ) . This was further exemplified on difEc substrates , in which 53% of recombination was observed without FtsK expression ( Figure 5B , XerCDVc , difEc , No FtsK ) . Resolution products were detected in the absence of XerC catalysis ( Figure 5C , XerCYFVc strains ) but not in the absence of XerD catalysis ( Figure 5C , XerDYFVc strains ) , signifying that XerDVc catalysis initiated recombination . dif1 differs from the γ-Proteobacteria consensus by only 2 bp , the substitution of A17 by G and the substitution of A10 by T ( Figure 5D ) . We therefore analyzed FtsK-independent XerCDVc recombination at hybrid sites between dif , dif1 and dif2 to identify residues important for the above observation ( Figure 5D ) . A site carrying the single [G-A]17 substitution promoted a much higher level of FtsK-independent recombination ( dif12 ) , while recombination at sites carrying the [T-A]10 and [G-A]1 substitutions was not altered ( dif13 and dif14 ) . However , the cumulative substitutions of [G-A]17 and [T-A]10 increased FtsK-independent recombination to a level equivalent to difEc-recombination ( dif15 ) . In addition , when T10 was altered to A in dif2 , we observed a faint recombination product ( dif23 ) , which was significant since FtsK-independent recombination was never observed at dif2 . Thus , G17 in the central region of dif1 and T10 in the XerC-binding site of dif1 and dif2 appear to have an important role in maintaining Xer recombination under the tight control of FtsK in V . cholerae . We next investigated if FtsKVc could serve to bring together the CDR sites carried by dimers of chromosome I or by dimers of chromosome II . Several key residues implicated in KOPS recognition have been identified in the γ domain of FtsKEc ( Figure 5A; N1296; R1300; E1303; [35] ) . The conservation of these residues in FtsKVc suggested that it could recognize the same motifs ( Figure 5A; N926; R930; E933 ) . If this was indeed the case , replacing the C-terminal domain of FtsKEc with the one of FtsKVc should completely rescue CDR in E . coli cells since FtsKVc fully activates recombination by XerCDEc at difEc ( Figure 5B , XerCDEc , difEc , FtsKVc ) . Indeed , the fitness of such cells equaled the fitness of wild-type E . coli cells in growth competition experiments ( Figure 6A , NLCVc and NLCEc ) , in contrast to cells only expressing the N-terminal domain of FtsKEc or a fusion with the C-terminal domain of H . influenzae FtsK ( Figure 6A , N and NLCHi ) . To test for the ability of FtsKVc to specifically recognize one of the E . coli KOPS motifs , we compared the efficiency with which it activates E . coli Xer recombination between plasmid-borne difEc sites flanked or not by the 5′-GGGCAGGG-3′ motif in an orientation that should prevent it from translocating towards difEc ( Figure 6B , KOPS-2 and KOPS-0 , respectively ) . Here we observed that the efficiency of recombination dropped significantly on KOPS-2 when FtsKEc or FtsKVc were used as activators ( Figure 6B , FtsKEc and FtsKVc ) . We then engineered an allele of ftsKVc carrying identical mutations to the one shown to abrogate KOPS recognition in FtsKEc [35] . No difference in recombination efficiency was noticeable between KOPS-0 and KOPS-2 when using this allele or its E . coli homologue ( Figure 6B; FtsK50CEc[N1296A; R1300A; E1303A] and FtsKVc[N926A; R930A; E933A] ) . We conclude that FtsKVc directly recognizes the GGGCAGGG motif and that recognition engages amino acids N926 , R930 and E933 . We decided therefore to analyze the skew and frequency of the GGGCAGGG motif on chromosome I and II . GGGCAGGG is highly polarized on both chromosomes with statistically significant skews ( Figure 6C ) . On chromosome I , the skew switches precisely at dif1 whereas on chromosome II one motif is present on the reverse orientation a few kb before dif2 ( Figure 6C ) . However , it has been shown in E . coli that a single non-permissive KOPS motif in the vicinity of dif is not sufficient to impair recombination [23] . The frequency of GGGCAGGG is low on both V . cholerae chromosomes ( Figure 6C ) , suggesting that this motif is not sufficient by itself to provide polar orientation of FtsKVc . We therefore analyzed the distribution of all octamers motif families with one degenerated position on both chromosomes . We ranked potential candidates according to their skew significance keeping only families that had a skew of at least 80% and a frequency of at least once every 30 kb . Only one family ( GGGNAGGG ) was among the 10 best candidates of both chromosomes . This family is highly skewed , frequent ( Figure 6C ) and contains the experimentally active GGGCAGGG motif . Taken together , these results suggest that the GGGNAGGG motifs might function as KOPS in V . cholerae .
The strand exchanges catalyzed by XerCVc and XerDVc occur at the junction between their respective binding site and the central region of dif1 and dif2 , as previously reported for the E . coli recombinases on dif ( Figure 3 ) . FtsKVc promotes recombination at both sites by activating a first pair of strand exchanges mediated by XerDVc ( Figure 4 ) , thanks to a species-specific interaction with the recombinases ( Figure 5 ) . In addition , GGGNAGGG motifs seem to function as FtsKVc-Orienting Polar Sequences , their frequency and distribution on the two V . cholerae chromosomes suggesting that the FtsKVc-translocase activity helps bring CDR sites together when dimers are formed on chromosome I or on chromosome II ( Figure 6 ) . We conclude that the same FtsK-dependent mechanism controls dimer resolution on each of the two V . cholerae chromosomes . We have previously shown in E . coli that the requirement for FtsK to activate Xer recombination delays CDR to the time of septum closure [26] , which is likely to also hold true in V . cholerae . Thus , the study of CDR provides the first example of a cell cycle coordination mechanism shared by the two V . cholerae chromosomes , which is similar to the way chromosomal maintenance processes are coordinated within the cell cycle of eukaryotes . Many bacteria harbor multiple chromosomes , which seems an important determinant of their individual life styles . A few bacterial species harbor linear replicons in addition to circular , such as Agrobacterium tumefaciens and the Borrelia species [3] . In the vast majority of cases , however , the multiple chromosomes harbored within a bacterium are circular . Maintenance of circular replicons requires the resolution of dimers created by homologous recombination events . In V . cholerae , 5 . 8% of dimers per cell per generation are formed on the 2 . 96 Mbp chromosome I and 3% of dimers are created on the 1 . 07 Mbp chromosome II ( Figure 2 ) . Under similar growth conditions , 15 . 6% of dimers are generated on the 4 . 6Mbp E . coli chromosome ( Figure 6 ) . These results suggest that dimer formation increases with replicon size , possibly reaching a theoretical upper limit of 50% for very large replicons . In addition , the rate of dimer formation seems to vary exponentially with replicon size for small replicons . Based on this hypothesis , the frequency of chromosome dimer formation in V . cholerae would be 11% per cell generation if it carried a single circular chromosome of 4 . 03Mbp . Instead , we measured a total rate of 8 . 6% for the two chromosomes ( Figure 2 ) . Thus , the particular genomic organization of the Vibrios seems to minimize chances for chromosome dimer formation , which is theoretically beneficial . Putative dif sites were readily identified on each of the two chromosomes harbored by 7 additional γ-Proteobacteria ( Figure 7 and Figure S2 ) . To determine dif sites in β- and α-Proteobacteria , we generated a profile Hidden Markov Model ( HMM ) based on the alignment of the putative CDR sites found in the larger chromosome of 27 γ-Proteobacteria using the program HHMER . We then compared each sequence by hand to ensure the proper 6 bp spacing between the putative XerC and XerD binding sites . Putative dif sites were thus identified on each of the multiple chromosomes harbored by 10 β-Proteobacteria species and 5 α-Proteobacteria species ( Figure 7 and Figure S3 and S4 ) . A single pair of recombinases orthologous to XerC and XerD was found in each of the 22 additional γ- , β- and α-Proteobacteria harboring multiple chromosomes , suggesting that a single pair of recombinases ensures dimer resolution of each of their non-homologous chromosomes . FtsK orthologues were also found . In addition , putative dif sites fell within 10 kb of the GC-skew inflection point ( data not shown ) , suggesting that dimer resolution is under the control of an FtsK-like homologue in all these species . Thus , the adoption of an FtsK-dependent dimer resolution system could be a key evolutionary step in the maintenance of large circular replicons . The sequence of Xer target sites , and especially of their central region , is a crucial determinant in the outcome of recombination [36] , [37] . Indeed , the central region of dif sites found in Proteobacteria with a single chromosome showed a high degree of conservation , most β- and γ-Proteobacteria harboring a ‘canonical’ 5′-TGTATA-3′ motif ( Figure 7 and Figure S2 and S3 ) , suggesting that there is a selective pressure on the sequence of the dif central region . This is further illustrated by the lower recombination efficiency of the E . coli system on dif2 compared to dif1 ( Figure 5 ) . In this regard , the V . cholerae Xer recombination system is remarkable since identical recombination efficiencies were obtained with the same pair of recombinases on dif1 and dif2 ( Figure 4 and 5 ) . However , XerCDVc-mediated recombination at dif2 required a tighter interaction between the recombinases and their partner translocase than at dif1 , since FtsKEc promoted 50% of recombination at dif1 but less than 20% at dif2 ( Figure 5B , XerCDVc , FtsKEc ) . In addition , a few alterations in the sequence of dif1 and dif2 decreased the stringency of the control exerted by FtsKVc ( Figure 5D ) , highlighting the extremely fine tuning of the different components of the V . cholerae CDR system . We observed that in Proteobacteria with multiple chromosomes , the central regions of dif sites from non-homologous chromosomes are divergent , as in V . cholerae ( Figure 7 and Figure S2 , S3 , S4 ) . A single exception was found in Burkholderia xenovarans , in which two of the three chromosomes of the bacterium harbor a resolution site with an identical central region . We reasoned therefore that some selective pressure imposes the divergence of the central regions of CDR sites carried by the different , non-homologous chromosomes of bacteria with multipartite genomes , which competes with the selective pressure for dif central regions to adopt the preferential 5′-TGTATA-3′ motif . Indeed , the presence of dif sites with identical central regions on two non-homologous chromosomes could lead to the formation of chromosome fusions by Xer recombination , which would disrupt the selective advantage brought by the multipartite genomic organization . In support of this hypothesis , preliminary experiments indicate that harmonization of the two V . cholerae dif sites leads to chromosomal fusions ( Val and Barre , unpublished observations ) . We are currently investigating how these fusions are formed and the consequences of harboring identical dif sites on separate chromosomes .
All growth experiments were done in LB-Lennox . Strains and plasmids are listed in Text S1 . Briefly , V . cholerae strains were derived from N16961 [1] by allele exchange using pDS132 derivatives [38] and E . coli β2163 as a donor strain [39] . E . coli strains used for in vivo plasmid resolution assays and for growth competition were engineered as previously described in [25] , [40] . Mutations were confirmed by PCR and sequencing . For growth competitions , E . coli cells were grown at 37°C with a 1000× dilution in fresh media every 12h [40] . Because of their higher growth rate , V . cholerae cells were grown at 30°C with a 10000× dilution every 12h . The numbers of CFU of mutated and parental cells in the cultures were determined by plating on cognate antibiotic plates every 12 or 24h , depending on the mutant growth defect . These numbers were used to calculate the number of generation of the parent cells between each time points and the CFU ratio of mutated versus parent cells at each time point . This ratio varies exponentially with the number of generations . The proportion of cells that the mutant strain fails to produce at each doubling time of its parent is deduced from the coefficient of this exponential . This ratio is a good estimation of the rate of dimer formation ( Text S1 ) . V . cholerae MBP-XerD and MBP-XerC recombinases were purified using nickel , amylose and heparin columns . The MBP tag was removed by thrombin digestion . dif1 and dif2 synthetic suicide substrates ( Text S1 ) were obtained by annealing synthetic oligonucleotides purified by PAGE . 5′-end labeling of oligonucleotides was performed using T4 DNA polynucleotide kinase and [32P] γ-ATP and 3′end labeling using terminal transferase and [32P] α-ddATP . Reactions were performed in 20 mM Tris-HCl ( pH 7 . 5 ) , 50 mM NaCl , 0 . 1mM EDTA , 1 μg/ml of BSA , 40% glycerol and 0 . 2 pmol of radiolabeled probe for 2 hours at 37°C . Covalent complexes were analyzed by 12% SDS-PAGE and cleavage sites by 12% urea-PAGE . Radioactivity was detected on a STORM ( GE Healthcare ) . E . coli cells were transformed with the FtsK expression vector and then with the Xer recombination reporter plasmid , as described in [25] . 10 transformant colonies were pooled in 1 ml of LB , diluted 100× in LB and grown to 0 . 6 OD at 37°C . Cells were then grown for an extra 2 hours at 37°C in the presence of 0 . 5% arabinose to induce FtsK production , unless otherwise indicated . Plasmid DNA was hydrolyzed with NdeI ( single cutter ) . Recombination efficiency was computed as the amount of replicative product over the sum of the amount of substrate and of replicative product , which were separated by agarose gel electrophoresis and detected with SybrGreen staining using a LAS-3000 ( Fuji Life Science ) . Leading strands were defined as the DNA strand reported in Genbank files downstream of the replication origin up to the terminus and the reverse complement strand from the terminus to the origin . The terminus position was chosen as the first nucleotide of the CDR site . Skew statistical significance was assessed by calculating the probability that the observed skew occurred by chance taking into account the fact that G-rich motifs are likely to be more frequent on the leading strand because of GC skew , as previously described [41] . Analysis on chromosome II was performed on a chimeric chromosome where the superintegron has been removed because this element carries more than 100 repetitions of the attC integration site , which hides the signal provided by octamer motifs .
|
During proliferation , DNA synthesis , chromosome segregation , and cell division must be coordinated to ensure the stable inheritance of the genetic material . In eukaryotes , this is achieved by checkpoint mechanisms that delay certain steps until others are completed . No such temporal separation exists in bacteria , which can undergo overlapping replication cycles . The eukaryotic cell cycle is particularly well suited to the management of multiple chromosomes , with the same replication initiation and segregation machineries operating on all the chromosomes , while the bacterial cell cycle is linked to genomes of less complexity , most bacteria harboring a single chromosome . The discovery of bacteria harboring multiple circular chromosomes , such as V . cholerae , raised therefore a considerable interest for the mechanisms ensuring the synchronous management of different replicons . Here , we took advantage of our knowledge of chromosome dimer resolution , the only bacterial segregation process for which coordination with cell division is well understood , to investigate one of the mechanisms ensuring the synchronous management of the smaller , plasmid-like , and larger , chromosome-like , replicons of V . cholerae .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"molecular",
"biology/dna",
"repair",
"molecular",
"biology/recombination",
"microbiology/microbial",
"evolution",
"and",
"genomics",
"microbiology"
] |
2008
|
FtsK-Dependent Dimer Resolution on Multiple Chromosomes in the Pathogen Vibrio cholerae
|
Protein-coding genes in eukaryotes are interrupted by introns , but intron densities widely differ between eukaryotic lineages . Vertebrates , some invertebrates and green plants have intron-rich genes , with 6–7 introns per kilobase of coding sequence , whereas most of the other eukaryotes have intron-poor genes . We reconstructed the history of intron gain and loss using a probabilistic Markov model ( Markov Chain Monte Carlo , MCMC ) on 245 orthologous genes from 99 genomes representing the three of the five supergroups of eukaryotes for which multiple genome sequences are available . Intron-rich ancestors are confidently reconstructed for each major group , with 53 to 74% of the human intron density inferred with 95% confidence for the Last Eukaryotic Common Ancestor ( LECA ) . The results of the MCMC reconstruction are compared with the reconstructions obtained using Maximum Likelihood ( ML ) and Dollo parsimony methods . An excellent agreement between the MCMC and ML inferences is demonstrated whereas Dollo parsimony introduces a noticeable bias in the estimations , typically yielding lower ancestral intron densities than MCMC and ML . Evolution of eukaryotic genes was dominated by intron loss , with substantial gain only at the bases of several major branches including plants and animals . The highest intron density , 120 to 130% of the human value , is inferred for the last common ancestor of animals . The reconstruction shows that the entire line of descent from LECA to mammals was intron-rich , a state conducive to the evolution of alternative splicing .
Spliceosomal introns that interrupt most of the protein-coding genes and the concurrent splicing machinery that mediates intron excision and exon splicing are defining features of gene architecture and expression in eukaryotes [1] , [2] . To date , eukaryote genomes including the compact genomes of parasitic protists , previously suspected to be intronless , have been shown to possess at least a few introns [3] , [4] , [5] and a ( nearly ) full complement of spliceosomal proteins [6] . However , eukaryotes dramatically differ in their intron densities , ranging from only a few introns per genome in many unicellular forms to over 8 introns per gene in vertebrates as well as some invertebrates like the sea anemone [7] , [8] . Despite the ubiquity of introns in eukaryotic genomes , their biological status is poorly understood . To what extent introns are “junk DNA” as opposed to being functional parts of the genome , remains an open question and the answers are bound to be complicated and multifaceted . There are many reports on the contribution of introns to the regulation of gene expression [9] , [10] , and in vertebrates introns encode a variety of non-coding RNAs with established or predicted regulatory functions [11] . However , it remains unclear how general such functional roles of introns are . In addition to these specific functions , numerous introns are essential for alternative splicing which involves the great majority of genes in multicellular eukaryotes and is one of the principal mechanisms of proteome diversification [12] , [13] , [14] . Given that most unicellular eukaryotes are intron-poor whereas complex , multicellular organisms are intron-rich , it would seem intuitively plausible that introns accumulated in the course of evolution of eukaryotes . However , comparative analysis of the exon-intron structures of orthologous genes of plants and animals revealed a high level of intron position conservation , with the implication that the common ancestor of these organisms was relatively intron-rich [15] , [16] , [17] , [18] , [19] . Moreover , reconstructions of the evolution of gene architecture that were performed using maximum likelihood ( ML ) approaches suggested intron-rich ancestors for several major groups of eukaryotes [19] , [20] , [21] including even the Chromalveolata , a eukaryotic supergroup that consists entirely of unicellular organisms [22] . These results imply that evolution of eukaryotes involved at least as much intron loss as intron gain , and that intron loss was the main process in the majority of eukaryotic lineages whereas intron gain was only episodic [19] , [21] . However , all these reconstructions provided relatively coarse resolution and involved substantial uncertainty with respect to the inference of intron density in deep ancestors , especially , the Last Eukaryotic Common Ancestor ( LECA ) . The uncertainty was caused by the sparseness of the genomic data sets employed for the reconstruction and by the difficulty of assigning confidence intervals to inferences of ancestral state . As a result , depending on the features of the ML models employed and the data sets analyzed , some of the reconstructions yielded evolutionary scenarios with an excess of intron gain over intron loss [23] . Here we employ a probabilistic Monte Carlo model combined with a Markov Chain Monte Carlo ( MCMC ) method for the inference of ancestral states including robust estimation of confidence intervals to analyze a representative data set of 99 eukaryotic genomes which extensively covered the three supergroups of eukaryotes , Unikonta , Archaeaplastida ( Plantae ) , and Chromalveolata , for which multiple genome sequences are available . The results clearly show that ancestral eukaryote forms were intron-rich , with LECA having a high intron density , on the order of two-thirds of the introns density in human genes . The subsequent evolution was heavily dominated by intron loss , with several episodes of massive intron gain associated with the emergence of some of the major eukaryote groups , in particular , animals .
The present analysis of gene structure evolution included an extensive data set of sequenced and annotated genomes from the Unikonta ( the Opisthokont group that combines animals and fungi , together with Amoebozoa ) , the Archaeplastida ( green algae and land plants ) , and Chromalveolata ( Heterokonta and Alveolata ) . Of the five supergroups of eukaryotes [24] , [25] , [26] , only these three are currently represented by multiple genomes with broad ranges of intron densities . There are no sequenced genomes for the supergroup of Rhizaria . The fifth supergroup , Excavata , includes mostly parasitic forms with very few introns and only one sequenced genome of a free-living organism , Naegleria gruberi , with a moderate intron density [27] , which renders ancestral reconstruction moot within this supergroup . Thus , our data set effectively covers the entire available diversity of eukaryotic genomes . The evolutionary relationships between the supergroups remain uncertain [26] , [28] , so they are represented as a trifurcation in the schematic evolutionary tree shown in Figure 1 . We identified large orthologous protein-coding gene sets that are represented in a substantial majority of the analyzed genomes using a procedure that combined ortholog clustering and gene-species tree reconciliation techniques ( see Methods and Supporting Text S1 for details ) . The encoded protein sequences from each of the orthologous gene sets were aligned and projected onto the coding nucleotide sequences , annotated with the exon-intron structures . The data set was further filtered to exclude aligned positions with significant ambiguity ( see Methods and Text S1 for details ) . The final data set contained 8403 intron presence-absence profiles from 245 sets of orthologous genes . Intron loss and gain were modeled using a probabilistic Markov model encompassing lineage-specific loss and gain rates , as well as rate variation across sites . The Markov Chain Monte Carlo ( MCMC ) method [29] was employed to sample model parameters and ancestral reconstructions by their posterior distributions , and to infer ancestral states along with the respective Bayesian confidence intervals ( see Methods and Supporting Text S1 for details ) . Experiments with various rate variation models across sites showed that only the loss rate variation had a significant impact on the model fit ( Figure 9 in Supporting Text S1 ) . Thus , it appears that , when uniform site preferences that apply across all eukaryotes are considered , introns in certain positions are prone to be lost significantly more often than others whereas no sites are significantly more prone to intron gain . This reconstruction provides a thorough view of the evolution of gene structure across three eukaryotic supergroups and reveal several general trends ( Figure 1 and Supporting Figure S1 ) . Most lineages show net intron loss that can be substantial as in alveolates , some lineages of fungi , green algae and insects , or well-balanced by concomitant intron gains as in land plants [30] , most animal lineages , and some fungi [31] . Massive intron gains were inferred only for several deep branches , most conspicuously , the stem of the Metazoa , and to a lesser extent , the stems of Mamiellales ( a branch of green algae ) , Viridiplantae , Opisthokonta , and Metazoa together with Choanoflagellata ( Figure 1 ) . These findings vindicate , on a much larger data set and with greater confidence , the previous conclusions that intron gain was rare during evolution of eukaryotes compared to intron loss . Episodes of substantial intron gain seem to coincide with the emergence of major new groups of organisms with novel biological characteristics such as Metazoa [19] . Several previous studies , performed on much smaller data sets and with less robust reconstruction methods , have suggested that at least some eukaryotic ancestral forms could have possessed intron-rich genes [19] , [20] , [31] . In particular , we found previously that the last common ancestors of Chromalveolata and particularly Alveolata could possess high intron densities despite the fact that all extant genomes available for in these groups are intron-poor [22] . The present analysis reinforces these conclusions by inferring high intron densities for the ancestors of each major group of eukaryotes within each of the three supergroups ( Figures 1 , 2 , and Supporting Figure S1 ) . The implication is that , whenever an extant eukaryotic genome shows a low intron density , this intron-poor state is a result of extensive , lineage-specific intron loss . Inspection of individual intron site histories revealed the same trends ( see Figure 3 and Supporting Video S1 ) . For example , Figure 3 shows the reconstructed history of intron loss and gain in the gene that encodes the membrane protease prohibitin . For this gene , a relatively high intron content was reconstructed for LECA , with four or five introns most likely present in the ancestral gene . The subsequent evolution of this gene involved multiple , parallel loss of introns in most of the eukaryotic lineages . Substantial intron gain is inferred only for Metazoa , one lineage of fungi , and one lineage of green algae . Notably , the intron content in mammals is the same as the inferred intron content of LECA ( five introns ) , and there is no intron-poor stage on the path from LECA to mammals ( Figure 3 ) . In addition to the Bayesian MCMC estimates , we inferred ancestral densities by using Dollo parsimony [32] , and by the posterior distributions in the maximum-likelihood ( ML ) model derived during the MCMC sampling . More precisely , the posterior reconstruction uses a fixed parameter set ( the ML model ) and infers a “plausible” history by computing the posterior probability of intron presence for every site at each ancestral node . Posterior probabilities are summed across sites to yield expected values [33] which can be interpreted as a parsimonious reconstruction weighed by the inferred lineage- and site-specific predispositions for loss and gain . The results of the comparison between the reconstructions obtained with the three methods indicate that parsimony reconstructions introduce a noticeable bias . The Dollo and ML estimates show a picture of intron-rich eukaryotic ancestors that is qualitatively similar to the MCMC results . Quantitatively , similarly to the case of ancestral molecular sequence reconstruction [34] , the Bayesian estimates often disagree with the parsimony reconstruction . Specifically , the MCMC sampling showed the tendency to infer higher ancestral densities ( 15–17% higher at intron-rich ancestors; see Figure 11 in Text S1 ) than Dollo parsimony , with the exception of the ancestors along the lineage from LECA to protostomes , for which Dollo parsimony yields up to 45% higher densities ( see Figure 11 in Text S1 ) . The differences highlight the idiosyncrasies of ancestral reconstruction methods and the pitfalls of disregarding model uncertainties . Dollo parsimony places the origin of introns at the most recent common ancestor of intron-bearing terminal taxa at each site , thereby systematically underestimating intron age and parallel gains . In contrast , ML infers similar ancestral reconstructions as MCMC ( Figure 11 in Text S1 ) , and the ML model parameters are not very different from the sampled model parameters ( 93% of the ML parameters fall within the 95% confidence intervals; see Figure S12 in Supporting Text S1 ) . The MCMC sampling procedure provides robust statistical estimates of ancestral states through Bayesian confidence intervals . The 95% confidence intervals are fairly tight around most estimates , even for such deep ancestors as those of alveolates ( 3 . 7–6 . 3 introns/kilobase ) , Dikarya ( “higher” fungi: 3 . 7–4 . 7 introns/kilobase ) , opisthokonts ( 4 . 7–5 . 5 introns/kilobase ) and , most importantly , LECA ( see below ) . The uncertainty is larger in ancestors with subsequent turbulent history in the descendants . A case in point is the amoebozoan ancestor . There was extensive intron loss along the branch leading from the intron-rich unikont ancestor to the extant Amoebozoa . It is unclear , however , whether the losses occurred in parallel in multiple descendant lineages , or prior to the split between Dictyostelium and Entamoeba ( see Figure 4 in Text S1 ) . Even more problematic is the reconstruction of the gene structure evolution in chromalveolates , because of the extensive intron turnover in many lineages within this supergroup . Indeed , there was no detectable intron conservation across haptophytes ( E . huxleyi ) , pelagophytes ( A . anophagefferens ) , diatoms , and other eukaryotes within or outside chromalveolates ( see Table 6 in Text S1 ) . For instance , the diatom T . pseudonana shares only 25% of introns with other diatoms in the data set , and only 3–6% with other eukaryotes . For comparison , human intron positions show 75–80% conservation with other Metazoa and 25–30% conservation with plants . Introns of Phytophthora and alveolates are also often conserved across large evolutionary distances . Accordingly , the reconstruction is fairly certain for the alveolate , Phytophthora and diatom ancestors and their descendants , and even for the chromalveolate ancestor , but many equally plausible scenarios are apparent for haptophyte ancestors ( see Figure 5 in Supporting Text S1 ) . Exploration of alternative phylogenies for the major chromalveolate groups yielded neither a better model fit , nor more precise estimates ( data not shown ) . These examples demonstrate the inherent uncertainties in ancestral reconstruction . Conceivably , the extensive intron turnover in chromalveolate algae , and the massive loss in Amoebozoa all but effaced any clues as to the ancestral gene structures , illustrating the fundamental limits of the reconstruction [35] . The gene architecture of LECA is of special interest . Previous estimates of intron density for LECA were very uncertain due to methodological problems with maximum likelihood inference [19] . The present reconstruction yielded the median value of 4 . 3 introns/kilobase , with the 95% confidence interval of 3 . 7–5 . 1 introns/kilobase ( Figure 2 ) , i . e . , 53–74% of the human intron density with a 95% confidence . Different resolutions of the trifurcating plant-unikont-chromalveolate root did not significantly affect the model fit ( see Figure 9 in Text S1 ) . Our analysis of the gene structure in the only sequenced genome of a free-living excavate ( a member of a fourth supergroup of eukayotes ) , Naegleria gruberi [27] , identified a high fraction ( 30–50% ) of intron positions shared with other supergroups ( see Table 14 in Supporting Text S1 ) , an observation that is compatible with an intron-rich LECA and with a moderate intron turnover within the line of descent leading from the LECA to Naegleria . Strikingly , the greatest intron density among all ancestral and extant eukaryotes was inferred for the last common ancestor of the Metazoa , at 120–130% of the human density , with a 95% confidence ( Figures 1 and 2 ) . We validated the inference procedures by simulating the evolution of intron sites ( see Figure 13 in Supporting Text S1 ) . The MCMC and ML methods infer the ancestral intron densities with no obvious bias , concurring on simulated data to a similar extent as on the main data set . In a sharp contrast , Dollo parsimony is significantly biased towards overestimation at many intron-rich ancestors . The variance of the probabilistic estimators at different ancestral nodes recalls the spread of Bayesian confidence intervals: fairly small variance was observed for almost all nodes including the LECA but the inferences for the amoebozoan and heterokont ancestors were unreliable . Additional simulation experiments ( see Figure 13 in Supporting Text S1 ) showed that the probabilistic models performed robustly even in the presence of missing orthologs , or heterotachious model violations . In all eukaryotes , with the interesting exception of the tunicate Oikopleura dioca [36] , introns show a non-uniform phase distribution , i . e . , an excess of introns that are inserted between codons ( phase 0 ) compared to introns between codon positions 1 and 2 , and 2 and 3 ( phases 1 and 2 , respectively ) [16] , [37] . We compared the inferred phase distributions for the gained , lost and ancestral introns ( or , in other words , derived the phase-specific gain and loss rates , and ancestral states ) . In most animals , including the ancestral forms , and in LECA , the ratios of the three phases remained nearly constant at 2∶1∶1 ( twice as many introns of phase 0 as there were introns of phase 1 or 2 ) . In some of the fungi and chromalveolates , the excess of phase 0 introns was less pronounced , whereas in plants , there was a greater than average excess of phase 0 and a paucity of phase 1 introns ( see Figure 7 and Table 8 in Text S1 ) . These findings indicate that the excess of phase 0 was a ( nearly ) universal feature of intron evolution throughout the history of eukaryotes but also reveal significant deviations from this pattern in some lineages . The mechanistic basis of both the ancestral excess of phase 0 and the lineage-specific variations remains to be identified . The results of this study reveal three principal modalities of evolution of the eukaryote gene structure: The choice between these routes of evolution in a particular lineage could depend primarily on the intensity of purifying selection that is linked to the effective population size [38] , [39] . Periods of large effective population size entail strong purifying selection and create a ratchet effect whereby lost introns are unlikely to be regained . Remarkably , the line of descent from LECA to mammals seems to have never gone through a strong selection stage , so the intron density remained continuously high , the only major perturbation being the gain of many introns at the onset of animal evolution followed by subsequent gradual loss ( Figure 1 ) .
The results of this work , thanks to the extensive data set of analyzed genomes and the robust reconstruction method that yields inferences of ancestral states with minimal uncertainty , seem to close the debate on the gene architecture of ancestors of extant eukaryotes including LECA . It is now clear that the genes of ancestral eukaryotes possessed high intron density , close to the densities in the most intron-rich modern genomes , those of mammals . This finding has substantial implications for understanding the evolution of eukaryotes . It has been noticed that intron-poor genomes typically possess strong , highly efficient splice signals , whereas intron-rich genomes contain mostly weak , error-prone splice signals [40] , an effect that appears to be due primarily to weak purifying selection that precludes both purging of introns and tightening of the junctions ( splice signals ) [41] . In intron-rich ancestral genomes , frequent errors of splicing yielding aberrant transcripts were inevitable . The abundance of such transcripts was the driving force behind , first , the evolution of defense systems that attack immature mRNAs and prevent their translation , like the nonsense-mediated decay ( NMD ) system that also contributes to expression regulation [42] , [43] , and second , the recruitment of aberrant transcripts to produce variants of proteins , the trend that in animals gave rise to the pervasive alternative splicing , one of the principal mechanisms of diversity generation and protein function regulation [12] , [14] , [44] . Remarkably , the present results indicate that the entire line of descent from LECA to mammals was a continuous intron-rich state ( Figure 1 ) that provided for uninterrupted evolution of the growing repertoire of functional alternative spliced forms . The unprecedented intron gain at the onset of animal evolution could further contribute to the expansion of alternative forms . This spurt of intron gain might have resulted from a combination of a population bottleneck that led to weak purifying selection with increased transposon activity that could activate double-strand break repair , a likely major mechanism of intron gain [45] .
Orthologous genes were identified using a modification of the previously described procedure [22] . The groups of putative orthologs from eukaryotes from the eggNog database [46] were employed as “seeds” to which members from the 99 selected genomes were added . The resulting candidate sets of orthologs were further filtered by verifying their phylogenetic relationships . In particular , a non-negative log-likelihood ratio between the neighbor- joining tree and the known species phylogeny , computed by PhyML ( Guindon and Gascuel , 2003 ) was required . The adopted phylogeny reflects known evolutionary relationships between major taxonomic groups [24] , [26] . Sequences of Naegleria gruberi were selected using the same procedure , but the large evolutionary distance precluded identification of a sufficient number of orthologs and unambiguous alignment of splice sites . Therefore , sequences from N . gruberi were not included them from the ancestral inference . The intron positions were mapped onto gene sequences using a previously developed computational pipeline [22] . The resulting data set is a table of intron absence and presence in which each column corresponds to a splice site projected onto an unambiguous alignment column ( retaining intron phase information ) , and each row corresponds to one of the 99 species . Table entries may be 1 ( splice site is present ) , 0 ( no splice site ) , or “*” ( ambiguous ) for a missing ortholog or an uncertain alignment portion . The final table was produced using the Malin software [47] and contained all columns with at most 24 ambiguous entries ( and at least one entry of 1 ) . Gene structure evolution was modeled mathematically by assuming that the table columns are independent and identically distributed random vectors . The distribution itself incorporates variable intron gain and loss parameters across lineages and splice sites ( 16 , 40 ) . For a formal treatment , define T as the known phylogeny for the terminal taxon set S , i . e . , a rooted tree with n leaves that are bijectively labeled by taxa from S . Internal tree nodes correspond to common ancestors . The history of a potential splice site is modeled as a binary labeling of all tree nodes: ξ = ( ξ[u]∈{0 , 1}: u∈T ) . In a Markov model , the labeling is randomly drawn from a distribution for which the parent-child relationships in the phylogeny define conditional independencies . The distribution of ξ at a site is fully determined by the presence probability at the root π = Pr{ξ[root] = 1} , and edge-specific rates . On the edge uv , labels change with probabilities Conversely , . The rates are set on each edge uv as where γ , ν are site-specific rate multipliers , and are lineage-specific average rates . The site-specific rate multipliers are drawn independently from discretized Gamma distributions [48] with the mean of 1 . The model is thus completely specified by the vector , where the hyperparameters α specify the shape of the Gamma distribution for the site-specific rate multipliers , and the edges are parametrized by their length and rate ratio , respectively . An input table column is a vector , where the character * denotes ambiguity . Accordingly , equivalence between resolved and unresolved labelings is defined bywhere ξ[S] is a random leaf set labeling . The model defines the likelihood for each table column . The likelihood for the complete data set , defined ascan be computed efficiently for a given model parametrization θ , and numerically optimized to find the maximum-likelihood parameters θ* [22] , [33] . The condition in the denominator accounts for the lack of columns with no splice site ( entry 1 ) at any terminal taxon . Ancestral intron counts were inferred using three methods . Intron count estimates were converted into densities by the formula density = intron count·6 . 946 kbp−1/875 . The conversion formula uses human as a reference: 6 . 946 is the mean number of human introns per 1000 base pairs ( kbp ) in the coding sequences of the analyzed genes , and 875 is the number of human introns in the data set . The posterior distribution for ancestral intron counts for a given model parametrization is computable without much difficulty [33] , and was used to infer the ancestral densities in conjunction with the maximum-likelihood model found during MCMC sampling , as implemented in the Malin software [47] . The ancestral intron positions were also inferred by using the Dollo parsimony principle , as implemented in Malin [47] . In order to estimate ancestral intron densities and lineage-specific changes in a Bayesian setting , we adapted mutation mapping techniques commonly employed with molecular sequence evolution models [34] . The Metropolis- Hastings algorithm [49] was used to estimate the posterior distributions for ancestral reconstructions and model parameters in a Markov-chain Monte Carlo framework [29] . The SAMPLING algorithm ( Box 1 ) generates a random walk by a Markov chain over the parameter space and ancestral reconstructions . In Line S4 , the acceptance probability includes the likelihoods L ( θ ) at different model parameters , the prior distribution P ( θ ) of parameters , and a proposed model distribution Q ( θ→θ′ ) . In Line S5 , random ancestral labelings are drawn at each column j by using the so-called conditional likelihoods for labeling node u with x = 0 , 1 , given the ( possibly unresolved ) labelings at the terminal taxa Su within the subtree rooted at u: The conditional likelihoods are calculated by dynamic programming in a postorder traversal by adapting the pruning algorithm of Felsenstein [50] ( LABELING algorithm , Box 2 ) . In Line L1 , the rate multipliers are drawn from the posterior distribution for the different discretized rate categories using the shape parameters of the respective Gamma distributions . The SAMPLING algorithm generates a Markov chain for pairs of model parameters and ancestral reconstructions . The equilibrium distribution for the chain is the posterior distribution In addition to sampling histories of profiles from the input data , we also generated “all-absent” profiles with introns missing at every terminal taxon [33] . The history of all-absent profiles was randomly sampled with the same procedure , and the number of such profiles was set as a negative binomial random variable with parameters , where is the probability of an all-absent profile . Ancestral intron counts were computed by tallying across all j , and adding the analogous sum for the sampled histories of all-absent profiles . Intron gains and losses on branches were estimated with a similar calculation . The prior distribution P ( θ ) was uniform for every parameter ( and thus absent from the formula in Line S4 ) : over the range [0 , 10] for shape parameters and edge lengths , and over the range [0 , 1] for π and the rate ratios . In a typical MCMC proposal , a subset of model parameters was chosen , and then multiplied by a random value; see Text S1for the details of the proposal distributions Q ( θ→θ′ ) . The convergence and the mixing efficiency were assessed by running 100 chains in parallel ( see Figures 1–3 in Text S1 ) . Estimates were computed using 50 , 000 independent samples from the joint posterior distribution q of parameters and ancestral intron densities . Individual intron site histories were reconstructed using the Malin software [47] with the median parameter values taken from the MCMC sampling . Simulations were performed by generating 100 random data sets of a comparable size to the input data set using the MCMC median model parameters , coupled with an erasure procedure simulating missing orthologs , or randomly generated multipliers for simulating heterotachy ( lognormal multipliers for rate parameters , exponential multipliers for edge lengths ) : see Figure 13 in Supporting Text S1 .
|
In eukaryotes , protein-coding genes are interrupted by non-coding introns . The intron densities widely differ , from 6–7 introns per kilobase of coding sequence in vertebrates , some invertebrates and plants , to only a few introns across the entire genome in many unicellular forms . We applied a robust statistical methodology , Markov Chain Monte Carlo , to reconstruct the history of intron gain and loss throughout the evolution of eukaryotes using a set of 245 homologous genes from 99 genomes that represent the diversity of eukaryotes . Intron-rich ancestors were confidently inferred for each major eukaryotic group including 53% to 74% of the human intron density for the last eukaryotic common ancestor , and 120% to 130% of the human value for the last common ancestor of animals . Evolution of eukaryotic genes involved primarily intron loss , with substantial gain only at the bases of several major branches including plants and animals . Thus , the common ancestor of all extant eukaryotes was a complex organism with a gene architecture resembling those in multicellular organisms . The line of descent from the last common ancestor to mammals was an uninterrupted intron-rich state that , given the error-prone splicing in intron-rich organisms , was conducive to the elaboration of functional alternative splicing .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"computer",
"science",
"biology"
] |
2011
|
A Detailed History of Intron-rich Eukaryotic Ancestors Inferred from a Global Survey of 100 Complete Genomes
|
Olfactory sensory neurons choose to express a single odorant receptor ( OR ) from a large gene repertoire and extend axons to reproducible , OR-specific locations within the olfactory bulb . This developmental process produces a topographically organized map of odorant experience in the brain . The axon guidance mechanisms that generate this pattern of connectivity , as well as those that coordinate OR choice and axonal guidance receptor expression , are incompletely understood . We applied the powerful approach of single-cell RNA-seq on newly born olfactory sensory neurons ( OSNs ) in young zebrafish larvae to address these issues . Expression profiles were generated for 56 individual Olfactory Marker Protein ( OMP ) positive sensory neurons by single-cell ( SC ) RNA-seq . We show that just as in mouse OSNs , mature zebrafish OSNs typically express a single predominant OR transcript . Our previous work suggests that OSN targeting is related to the OR clade from which a sensory neuron chooses to express its odorant receptor . We categorized each of the mature cells based on the clade of their predominantly expressed OR . Transcripts expressed at higher levels in each of three clade-related categories were identified using Penalized Linear Discriminant Analysis ( PLDA ) . A genome-wide approach was used to identify membrane-associated proteins that are most likely to have guidance-related activity . We found that OSNs that choose to express an OR from a particular clade also express specific subsets of potential axon guidance genes and transcription factors . We validated our identification of candidate axon guidance genes for one clade of OSNs using bulk RNA-seq from a subset of transgene-labeled neurons that project to a single protoglomerulus . The differential expression patterns of selected candidate guidance genes were confirmed using fluorescent in situ hybridization . Most importantly , we observed axonal mistargeting in knockouts of three candidate axonal guidance genes identified in this analysis: nrp1a , nrp1b , and robo2 . In each case , targeting errors were detected in the subset of axons that normally express these transcripts at high levels , and not in the axons that express them at low levels . Our findings demonstrate that specific , functional , axonal guidance related genes are expressed in subsets of OSNs that that can be categorized by their patterns of OR expression .
The development of the nervous system requires that cell fate specification be coordinated with axonal pathfinding , dendrite formation , and synapse formation . Since specific neuronal populations connect with particular regions in the brain , cellular identity is tightly linked with the expression of distinct axon guidance molecules that regulate target selection . In the developing olfactory system , the cellular identity of olfactory sensory neurons ( OSNs ) is manifested by the expression of particular odorant receptors ( ORs ) . Each OSN expresses only one OR out of a large repertoire available in the genome [1–4] . In the mature nervous system , OSNs project axons to reproducibly positioned , OR-specific glomeruli in the olfactory bulb [5–7] . Olfactory experience is thus projected onto the brain as a topographic map of glomerular activity [8 , 9] . The formation of olfactory circuitry therefore requires a close co-ordination between the expression of particular guidance molecules and particular ORs . What are some of the molecular mechanisms that lead to the establishment of this pattern of connectivity ? In the fly , semaphorin expression in the ventral antenal lobe determines whether entering sensory axons extend into dorsal or ventral regions of the lobe [10] . In the mouse , ventrally expressed Slits induce early arriving , Robo2 expressing OSN axons to target dorsal regions of the olfactory bulb [11 , 12] . Further , Sema3F expression in early arriving OSN axons is proposed to restrict later arriving , Nrp2 expressing mouse OSN axons to increasingly more ventral target locations [13] . Segregation of OSN axons into OR specific glomeruli is dependent upon the more granular and mosaic expression of a variety of signaling and adhesion molecules . In the fly , tenurins and tol2-related molecules facilitate matching between specific OSN axons and their appropriate postsynaptic targets [14–15] . In the mouse , the combined activity of adhesion and signaling molecules including kirrels , BIG-2 , clustered protocadherins , and ephrinAs are thought to drive OSN axon segregation and glomerulus formation [16–19] . Finally , odorant stimulated activity of OSNs leads to a pruning of mistargeted axons and refinement of glomeruli [20] . In both the fly and the mouse , OR choice and glomerular position are tightly coordinated . In the fly , ORs are dispensable for glomerular targeting and a network of transcription factors are likely to determine both cell identity and guidance receptor expression [21–24] . In contrast , OR expression affects glomerular targeting in the mouse [7 , 25] . This has led to the development of a model in which OR activity influences the expression of guidance receptors and adhesion molecules including NRP1 and the kirrels [16 , 26] . In spite of this progress in understanding the development of olfactory circuitry , there is still much to learn about how olfactory axons navigate to their appropriate targets , particularly in vertebrate systems . RNA-seq is a valuable approach for the identification of new candidate guidance-related genes . Several studies have examined the transcriptomes of OSNs as revealed by bulk RNA-seq of pooled olfactory sensory neurons [27–32] . Typically the pooled cells were selected using broad markers such as the Olfactory Marker Protein ( OMP ) . This approach has identified axon guidance molecules that are expressed at various stages of neuronal maturation [28] . Single cell RNA-seq studies have revealed that OSNs express multiple OR transcripts early before settling on a single OR [31–34] . However , little attention has been paid to differentially expressed transcripts that could regulate specific axonal targeting of particular OSNs . There is a dearth of markers that label smaller OSN subpopulations within the broader OMP population , making it difficult to identify axon guidance mechanisms through which OSNs achieve differential targeting . It is equally challenging to study the correlation between OR-specific neuronal identity and guidance molecule expression , a fundamental feature of olfactory circuit development . We used single-cell RNA-seq in the olfactory system of larval zebrafish to address these issues . OSNs are born , differentiate , and extend axons to the bulb starting around 30 hpf [35] . By 72 hpf a crude topographic map is already established in the olfactory bulb with OSN axons targeting 11 discrete , individually identifiable neuropilar regions called protoglomeruli [35] . Our previous work showed that OSNs choosing to express the most related ORs innervate a common protoglomerulus , thus demonstrating an early link between neuronal identity and axon guidance [36] . OSNs choosing to express ORs from OR clades A or B project to the CZ protoglomerulus , while those expressing ORs from OR clade C project to the DZ . What are the molecular mechanisms that enable these nearly identical OMP expressing axons to specifically target either the CZ or the DZ protoglomeruli ? We hypothesize that there are distinct , target specific sets of guidance receptors expressed by each category of neuron . Further , since axons that express related ORs converge onto the same protoglomerulus , guidance molecules should be differentially expressed depending upon the OR they express . For these reasons , we set out to identify guidance related molecules that are differentially expressed depending upon the OR an OSN chooses to express . In this study we show that differentially expressed transcription factors and guidance related molecules can be identified in specific subsets of OSNs expressing related ORs . In this way we identify candidate axonal guidance molecules that have the potential to guide distinct sub-classes of OSNs to specific targets in the bulb . We show that different subsets of axon guidance transcripts show strong correlation with specific subsets of transcription factors , potentially identifying transcriptional pathways that could coordinate neuronal identity with axon guidance . Finally , we provide evidence that three of the candidate guidance genes we identify: nrp1a , nrp1b , and robo2; are functionally required for targeting particular subsets of axons to distinct protoglomeruli . Our findings are consistent with a model of olfactory map formation in which subsets of OSNs express common guidance molecules leading them to target larger protoglomeruluar regions that are subsequently subdivided into smaller OR-specific glomeruli later in development .
OSN axons first reach the bulb 24 hours post fertilization ( 24 hpf ) and their initial targets , the protoglomeruli , are well formed by 72 hpf . OSNs from 48 hpf embryos were collected for analysis to maximize the chances of capturing transcripts relevant to axon guidance . We selected neurons expressing OMP , a marker of mature , axon bearing OSNs . Fish containing the zebrafish OMP promoter driven RFP transgene ( zOMP:RFP; [37] ) were intercrossed to obtain OMP:RFP expressing embryos ( Fig 1A ) . RFP positive cells were isolated from dissociated olfactory epithelia by FACS . A cell suspension of the selected cells was loaded onto a Fluidigm C1 chip . Cells were sorted into individual wells , inspected visually to ensure that wells contained single cells , and cDNA was generated and amplified in each well . Fifty-eight single-cell cDNA samples had high yield and an even distribution of transcripts between 800bp-4kb as assessed using a Bioanalyzer . These samples were used for multiplex library preparation . Each sample was deep sequenced to a depth of 50 million reads to maximize the chances of capturing transcript differences in less highly expressed genes that may drive subpopulation fate or affect axon guidance . Reads were mapped to zebrafish genome assembly GRCz10 and to a spike-in RNA set . For most cells , approximately 50% of the reads mapped to the zebrafish genome while the remaining reads mapped to spike-in RNAs . In 8 cells less than 15% of the reads mapped to spike-ins . These were likely to be cell clumps that escaped our visual check and were excluded from further analysis . The 50 remaining single-cell transcriptomes typically had an average of 22 . 6 million mapped reads to the zebrafish genome per single-cell transcriptome . 2 cells had less than 3 million reads mapped to the zebrafish genome and were excluded from the dataset . A third cell was excluded because it had a disproportionate number of reads to a small number of genes . It was first determined if the remaining isolated cells varied in their maturity . Unsupervised hierarchical clustering was conducted based on the expression of OMP and other molecular markers for mature OSNs: gng13b , gnal , ompb , cnga2 , cnga4 , ano2 , and adcy3b [31 , 33]; and markers for immature OSNs or their precursors: gng8 [33] and gap43 [31 , 33] . Five cells showed high expression of the immature markers , gng8 and gap43 , along with low expression of mature cell markers ( Fig 1B: topmost bar , black shading ) . 42 cells showed minimal expression of early stage markers and strong expression of mature cell markers like gng13b , gnal , ompb and cnga4 . Thirteen of these mature cells did not express appreciable ORs ( Fig 1B: topmost bar , grey shading ) . They may have yet to turn on their expression of OR genes , or instead , they could be non-OR expressing OMP neurons . None of these cells were found to express another major category of olfactory receptors , the TAAR receptors . They were excluded from further analysis . The remaining 29 mature cells expressed OR transcripts , typically with a single one predominating ( Fig 1B: topmost bar , magenta shading ) , and these were the cells used in all subsequent analyses . None of the immature cells expressed appreciable levels of OR transcripts . Thus , a single OR transcript is chosen very early in the development of zebrafish OSNs as the axons are making crucial targeting decisions . OSNs that express closely related ORs have been shown to project to the same protoglomerulus [36] . The molecular homology between zebrafish ORs suggests a phylogenetic organization into three major clades that we have designated clades A , B , and C ( Fig 2A , [4] ) . Clade A is comprised of subfamilies or101-114 , clade B of subfamilies or115-128 , and clade C of subfamilies or129-137 . Experiments with BAC reporter constructs for 16 ORs from clades A and B labeled axons that projected to the CZ protoglomerulus , while those for 3 ORs from clade C projected to the DZ protoglomerulus [36] . We therefore classified single OSN transcriptomes by the expression of their predominant OR and assigned each to one of three OR homology clades: A , B , or C ( Table 1 ) . Nine cells were assigned to the clade A category , 10 cells to clade B category , and 10 to clade C category ( Table 1 ) . Cell s20599 ( marked with * , Table 1 ) had less than a two-fold expression difference between or106-8 and or111-2 . Since both these ORs belong to clade A , s20599 was assigned to clade A . Similarly , s23558 ( marked with * , Table 1 ) had comparable expression levels for 2 OR transcripts , or133-2 and or129-1 , both of which belong to homology clade C . Hence , it was assigned to the clade C category . It has been proposed that particular OR genes are first expressed at specific time-points during development [38–40] . Only OSNs that expressed markers indicative of mature OSNs were included in our analysis , and ORs from all three clades were equally represented in our dataset . It is possible that OSNs expressing particular ORs differentiated before or after others , but it is unlikely that the differences we see among OSNs categorized by clade can be ascribed to differing levels of maturity . For all subsequent analyses , we compared the expression of transcripts among cells categorized by clade ( OR clades A , B , or C ) , with the underlying assumption that OR clade choice and axon targeting are closely related . We first sought to identify OR clade-specific gene expression patterns in order to identify potential marker genes that could be used in the future to label each subpopulation of neurons . A total of 29 , 033 distinct mapped transcripts were filtered to identify those detected in 3 or more mature cells at a level of 100 or more raw counts . 7 , 027 transcripts met these criteria . Rather than computing gene-by-gene differential expression , we considered multi-variate combinations that might arise from coordinated differential expression of gene sets . We utilized Penalized Linear Discriminant Analysis ( PLDA ) as implemented in the Penalized LDA package [41] , with a lasso penalty on the discriminant vector to induce feature scarcity . Applying this approach to high-dimensional gene expression covariation matrices effectively constrains the solutions when the number of variables is much higher than the number of classes [41] . We extracted two discriminant axes that separated the cells between the three OR clades ( Fig 2B ) . The PLDA functions of these two axes contain 2886 genes . The absolute values of loading coefficients for each gene are displayed in rank order for the top 500 genes in each discriminant ( Fig 2C and 2D ) . The 20 highest values for each discriminant axis are colored in red . To mitigate over-fitting the model , 40 genes , representing the cumulative top 20 loadings from each discriminant axis were used to re-compute the penalized LDA function . Cross-validation accuracy of this reduced discriminant function was 98 . 6% . The utility of this sparse feature set of 40 genes in capturing clade-specific structure within single cell OSNs transcriptomic landscape can be seen in Fig 2E . Genes whose expression is high in each clade related group are boxed . Genes gclm , abcf2a , pgbd4 , DNAJA4 , ube2e3 , asna1 , zgc:86598; gen1 , ildr1a , ier5l , fxyd6l atf3 , rbm4 . 3 , psmc5 , hadhab , and phax are strongly expressed in clade A cells . Genes sulf2b , ost4 , naa25 , and sgtb show elevated expression in clade B cells; while clade C cells show elevated expression for si:ch211-239j9 . 1 , pcdh11 , snap25b , anxa1a , lin7a smarcd1 , dlp2cb , eif2ak3 , wu:fb59d01 , sphk2 , CAMKK2 , zbtb11 , or129-1 , p4ha3 and lrrc4bb . A majority of these genes encode intracellular metabolic enzymes , a smaller number are cell surface proteins , and one is a transcription factor . Some of these genes may serve as useful markers for clade-specific subsets of OSNs . Four of these genes are cell surface receptors and three of them are candidates for guidance genes: ildr1a , pcdh11 , and lrrc4b . Interestingly , clade B cells may have a transcriptional profile that is distinct from that of clade A . This raises the possibility that OSNs expressing ORs from clade A as compared to clade B are distinct , even though they all project to the CZ protoglomerulus . The remaining genes are expressed in quantitatively but not qualitatively useful patterns , and are unlikely to be useful in classifying OSN subsets . Overall , our findings suggest that PLDA can identify genes that are expressed in OSNs that express ORs from particular clades . Our primary focus was to identify candidate guidance-related genes that could account for the differential targeting of OSN axons to different protoglomeruli in the bulb . Target-P [42] was used to identify genes with N-terminal signal sequences . Signal sequences mark genes destined for intracellular compartments , the cell surface , or secretion . Since axon guidance receptors are associated with the cell surface and axon guidance cues are either associated with the cell surface , the ECM , or are secreted , this approach provides us with an unbiased gene set potentially involved in the guidance of olfactory axons . Target-P identified 6696 genes in the zebrafish transcriptome with an N-terminal signal sequence , and of these , 6197 were expressed in our single cell transcriptomes . Of these , 1156 genes were expressed in three or more mature cells at minimum raw counts of 100 . OR genes were removed from the list . Using PLDA , we extracted two discriminant axes separating cells of the three OR-clades . The PLDA functions contain 343 genes with signal sequences ( Fig 3A ) . The genes loading to the V1 and V2 discriminant axes were plotted in descending order of their absolute value ( Fig 3B and 3C ) . Sparse discriminant functions ( see methods ) from this feature set had cross-validation accuracy of 100% . Of the approximately top 50 loadings for each discriminant axis ( 105 total ) , we selected genes most likely to be secreted or localized to the cell surface in order to enrich for potential axon guidance-related molecules ( Fig 3B and 3C , red dots ) . A total of 47 of 105 genes were selected as having the potential to be guidance related based upon annotations in UNIPROT [43] . Three of these are secreted and the rest are localized to the plasma membrane . Visualizing the expression of these 47 candidate ‘guidance genes’ in a heatmap shows that 40 of them are expressed most highly in just one of the three clade-related categories ( Fig 3D ) . Some of the cell surface genes are canonical guidance receptors such as robo2 , nrp1a , and nrp1b . lrp5 also plays a role in axon guidance [44] while other genes like nlgn2a , nlgn3b , and lrrc4b have been shown to be important for synapse formation [45 , 46] . These genes are novel candidates for playing a role in the guidance of OSNs . An interesting observation is that of the 47 guidance candidates , many belong to common families such as nrp , lrp , fxyd , and fzd . There are multiple instances where genes belonging to the same family show higher expression in different classes of neurons . For example , nrp1a has higher expression in clade A cells while nrp1b is has higher expression in clade C cells . lrp11 and lrp5 show higher expression in clade A cells while lrp1ab is highly expressed in clade C cells . On the other hand , fzd3a and fzd8a both show higher expression in clade A OSNs . These findings are consistent with the idea that a large number of signaling systems potentially contribute to OSN axon targeting , and that particular family members may guide single subsets of OSNs to their targets . One mechanism that could explain the coordination of OR choice and OSN targeting would be that their expression is regulated by some of the same transcription factors . If true , there should be distinct transcription factor profiles associated with each clade-specific category of OSN . Of the 2345 annotated transcription factors in zebrafish ( AnimalTFDB 2 . 0 [47] ) , 1610 were detected in our single cell dataset , 283 of which met the 3 cells with a minimum of 100 raw counts expression threshold criteria . PLDA using these transcription factors revealed 218 genes whose expression discriminated between the 3 clade-based categories of OSNs ( Fig 4A , 4B and 4C ) . The top 20 genes based on the absolute value of the loadings in each discriminant axis were used to re-compute a penalized LDA function consisting of a sparse 40-gene feature set . Cross-validation accuracy of this sparse PLDA function was 95 . 9% . An analysis of these 40 genes contributing to the V1 and V2 discriminant axes identified 31 transcription factors that show higher expression in one of the 3 clade related categories of cells ( Fig 4D ) . We next examined if the transcription factors and guidance genes identified through PLDA showed co-expression patterns correlated with the OR clades in which they were expressed . We computed Pearson’s correlation coefficients ( Fig 5 ) between all pairs of a combined set of the top guidance related ( from Fig 3D ) and transcription related ( from Fig 4C ) transcripts that are expressed in only one clade-related category of OSNs . Genes are annotated for the clade-specific category in which they showed higher expression and ordered by unsupervised clustering using average linkage to produce the heat map in Fig 5 ( clade A: red circles , clade B: blue circles , clade C: green circles ) . This correlation plot shows 3 easily identifiable blocks of higher correlation coefficients . The blocks correspond to OSNs expressing ORs from clades A , B , or C . These findings demonstrate the correlated co-expression of distinct ORs , transcription factors , and potential guidance related genes within our dataset . We next wanted to evaluate the biological significance of the gene expression patterns identified through PLDA . We took advantage of the Tg ( or111-7:IRES:GAL4 ) line of fish that carries a transgene in which or111-7 and GAL4 expression are driven by the combined action of an enhancer element found near the or111 subfamily gene cluster and the promoter for or111-7 . The resulting expression of GAL4 labels a subset of OSNs whose axons project to the CZ and LG1 protoglomeruli ( [48] Fig 6A , purple and grey axons ) . or111-7 is a member of the clade A category of ORs and the projection to CZ is expected while the projection to LG1 is aberrant [48] . We generated fish larvae containing or111-7:IRES:GAL4; zOMP:RFP; and UAS:Citrine transgenes . In these larvae OSNs expressing both Citrine and RFP project exclusively to the CZ protoglomerulus ( Fig 6A , purple axons ) . 48 hpf olfactory epithelia were dissociated before performing a two-way FACS to select OSNs expressing both RFP and Citrine ( Fig 6B ) . Approximately 1000 cells were collected for each of three replicates , total RNA was isolated , mRNAs were amplified by reverse transcription using oligo-dT fused to a T7 promoter sequence [49] , libraries were constructed and sequenced . We hypothesized that the gene expression profile of or111-7:IRES:GAL4; UAS:Citrine; zOMP:RFP expressing OSNs should best match that of clade A single cells . Unsupervised clustering was performed with Principal Component Analysis using only the 40 class specific genes we identified ( in Fig 2E ) that most strongly discriminate between the clade A , B , and C categories of OSNs . The results are plotted for each of the 29 single cell transcriptomes together with the three replicates of the CZ projecting bulk transcriptomes ( Fig 6C ) . First note that the single-cell transcriptomes organize into three distinct , clade-specific clusters on the two principal component axes . Second , the three CZ projecting OSN transcriptomes are most closely associated with the clade A cluster . The same result was obtained when only the 47 candidate guidance related genes identified by PLDA ( in Fig 3D ) were used to perform an unsupervised clustering of the single cell and CZ projecting transcriptomes ( Fig 6D ) . This suggests that the guidance receptor transcriptome profile of the CZ projecting OSNs is most similar to clade A category OSNs . It provides additional support for the idea that the candidate guidance genes we identified are associated with CZ protoglomerular targeting . In situ hybridizations were performed to verify that the expression of selected guidance-related genes is associated with specific sub-families of ORs . Double fluorescent ISH for each one of the four different guidance receptors nrp1a , nrp1b , robo2 , or pcdh11; were performed with sub-family ( sf ) specific cocktails of probes for the or111sf ( clade A ) or the or133sf ( clade C ) . Cells labeled with a particular subfamily of OR probes were visually inspected for the presence or absence of guidance gene signal ( Fig 7A ) . Based on our single-cell transcriptome profiles , we would expect clade A enriched gene expression to overlap with or111sf expression , while clade C enriched gene expression to overlap with that of or133sf expression . As expected , nrp1a is co-expressed to a greater degree with or111sf than with or133sf expressing OSNs ( Fig 7B ) . Also as expected , nrp1b , robo2 , and pcdh11 were co-expressed to a greater degree in or133sf as compared to or111sf expressing OSNs . Thus , the pattern of association of guidance receptors with OR subfamilies as seen by FISH is consistent with the single-cell RNA-seq approach . It is important to note that the overlap is a qualitative rather than a quantitative measure and may not reflect true differences in expression levels . In order to test if the candidate genes identified by sc-RNA-seq have a functional role in axon guidance , we used OR reporter lines that express both an OR and the GAL4 activator directly under the control of the OR promoter . Tg ( BACor111-7:IRES:GAL4 ) expresses GAL4 in CZ-projecting OSNs that belong to clade A OSNs , and ( TgBACor130-1:IRES:GAL4 ) expresses GAL4 in DZ-projecting clade C OSNs . GAL4 expressing axons are visualized when these lines are crossed to a transgenic UAS:Citrine reporter line [36] . We first asked if nrp1a is required for the targeting of either of these axon populations . Our data shows that nrp1a transcripts are highly enriched in clade A cells which are known to project to the CZ , and further , that or111-7:IRES:GAL4 expressing OSNs have a similar expression profile as clade A OSNs . In contrast , nrp1a is not highly expressed in clade C sensory neurons . We therefore predict that nrp1a should be required for the correct targeting of or111-7 axons but dispensable for the targeting of or130-1 axons . The point mutant nrp1asa1485 has a nonsense mutation at aa206 which produces a premature stop in Nrp1a [50] . Either Tg ( nrp1a+/-; BACor111-7:IRES:GAL4 ) or Tg ( nrp1a+/-; BACor130-1:IRES:GAL4 ) fish were crossed with Tg ( nrp1a+/-; UAS:Citrine ) fish and their progeny genotyped and analyzed . Consistent with expectations , or111-7 expressing , citrine labeled neurons that should project exclusively to the CZ protoglomerulus , show significantly higher ectopic misprojections to the DZ in nrp1a -/- mutant as compared to their WT siblings ( Fig 8A–8C ) . In contrast , or130-1 expressing neurons navigate normally to the DZ in nrp1a -/- mutants ( Fig 8J–8L ) . We next tested the contribution that nrp1b makes to the targeting of OSNs expressing either or111-7 or or130-1 . nrp1b might be expected to contribute to DZ targeting as it is more highly expressed in clade C as compared to other OSNs . nrp1bfh278 carries a nonsense mutation at aa116 that leads to the generation of truncated protein [50] . WT and mutant larvae were generated as described above for analysis . As predicted , or130-1 expressing OSNs that should project exclusively to the DZ protoglomerulus have significantly more ectopic CZ misprojections in nrp1b-/- mutants as compared to their WT siblings ( Fig 8M–8O ) . In contrast , or111-7 neurons projected without additional errors to the CZ ( Fig 8D–8F ) . Thus , nrp1b is required for normal axonal guidance of clade C but not clade A OSNs . Lastly , the contribution of robo2 to DZ protoglomerular targeting was tested . robo2 is expressed more highly in clade C as compared to clade A OSNs . We employed the robo2ti272 mutation that carries a nonsense codon at aa635 and generates a truncated robo2 protein . This mutation has been shown to cause a dramatic misprojection phenotype in the retinal axon projection to the tectum [51] . WT and mutant larvae were generated as described above for analysis . Abnormal projections of BACor130-1:IRES:GAL4 expressing OSNs were observed at a higher rate in robo2 mutants as compared to WT larvae , particularly misprojections to the MG protoglomerulus ( Fig 8P–8R ) . In contrast , there is no increase in mistargeting of CZ projecting BACor111-7:IRES:GAL4 ( clade A ) expressing OSNs ( Fig 8G–8I ) . These data are consistent with clade C OSNs , but not clade A OSNs , requiring robo2 for their normal targeting to the DZ protoglomerulus . In each of the three mutants we examined , OSNs expressing a candidate guidance molecule at a higher level required that candidate for normal axonal targeting . Those OSNs expressing the same candidate at a lower level did not require it for targeting . These results strongly suggest that the candidate genes identified by single cell RNA-seq are likely to play a role in guidance and targeting of the clade-specific OSNs in which they are enriched .
We have identified subsets of candidate axon guidance-related genes whose differential expression is related to the clade from which an OSN chooses its OR . We have confirmed that three of these candidates: nrp1a , nrp1b , and robo2 , are functionally required for normal olfactory axon targeting in specific subsets of OSNs that express particular ORs . Not only axon guidance related genes , but also transcription factors , are differentially expressed in an OR clade-specific manner . These findings are consistent with a model in which distinct subsets of transcription factors coordinate general aspects of OR choice and axon targeting in OSNs . The dimension reduction technique , PLDA , has proven to be highly useful in this study for the identification of differentially expressed class-specific genes , guidance factors , and transcription factors . This approach has previously been successfully employed to characterize human brain cell types [52] . By focusing on a small number of genes that provide the most discriminative power for each gene set we analyzed , we have tried to avoid over-fitting problems that can arise with large multidimensional datasets . PLDA requires a priori knowledge of cell classification in order to maximize class separation using the available set of features . It would be useful in many studies where cells can be grouped based on molecular markers , developmental age or cellular state [53–55] . It is however not applicable to studies where there is no prior knowledge of sub-population characteristics and would be a poor choice to identify entirely novel cell sub-types . This method was suitable in our study since we were able to classify OSN transcriptomes into three categories based upon the predominant OR transcript expressed in each OSN . One caveat of using single cell RNA-seq datasets to identify differentially expressed genes is that the high frequency with which transcripts fail to be identified in any given cell can lead to the appearance of differentially expressed genes that arise solely from random fluctuations in transcript detection . Similarly , noise in the data could result in a failure to detect the actual differential expression of a gene . For these reasons , we have taken a conservative approach in identifying differentially expressed genes by choosing a limited number of genes with the highest coefficients of discrimination for each determinant . The weighted contribution of each gene to the discrimination between cell classes can be ranked and sorted to identify the most discriminating genes , although the cut-off for the most informative genes is somewhat arbitrary . These considerations make it essential to employ tests of identified genes for true patterns of differential expression and/or their expected biological function . One test we performed was to use a pair of transgenes to isolate a population of OSNs that extend axons exclusively to the CZ protoglomerulus . The transcriptomes of these cells matched with reasonable fidelity clade A-specific marker genes as well as the guidance-related gene set for clade A OSNs . As all of the OR clade A expressing OSNs we have tested project to the CZ protoglomerulus [36] , these findings provide a strong affirmation of the differential expression of these genes in CZ projecting OSNs . In a more direct approach , all four of the candidate genes chosen for FISH showed expression patterns that were consistent with the single-cell RNA-seq data . Finally , a test of biological function delivered the expected phenotypes for three candidate genes . The axons of OSNs in which each candidate was highly expressed projected aberrantly in mutants , while the axons of OSNs in which the candidate was expressed at a lower level projected normally . These results give us confidence that the guidance-related genes identified by PLDA in this study are good candidates for roles in OSN axon guidance . A single cell approach has enabled us to characterize subsets of OSNs in the absence of any known molecular markers other than the ORs they express . We confirmed our working hypothesis that OSNs choosing to express ORs from each of the three clades of OR genes can be divided into three separate categories based on their differential expression of specific transcription factors , guidance related molecules , or other miscellaneous genes . It is interesting that clade A and B OSNs can be distinguished in this way , as both appear to project to the CZ protoglomerulus [36] . Future studies will be directed towards determining if clade A and B OSNs target different domains within the CZ . The CZ is the largest protoglomerulus at 72 hpf and is reported to split in two to form the maG and the vmG/vaG protoglomeruli by 96 hpf [56] . It will be of interest to determine whether these two protoglomeruli are differentially targeted by clade A and clade B OSNs . Although our findings support the OR clade-based categorization we imposed on OSNs , it does not preclude additional subcategories within each clade-specific category , or even alternative categories that cut across the clade-related categories . We find that nrp1a is required in the zebrafish for the proper targeting of OSN axons to the CZ protoglomerulus , but not to the DZ protoglomerulus . In contrast , nrp1b is required for axons targeting the DZ , but not the CZ . As mentioned previously , semaphorins play important roles in organizing the olfactory projections in both the fly and mouse . In the fly , semaphorins organize the positioning of dendrites of projection neurons within the antennal lobe that serve as the targets for sensory axons , determine the ventromedial trajectory of OSN axons , and participate in glomerular targeting of sensory axons [10 , 15 , 57–59] . In the mouse , Sema3A helps to organize fasciculation of OSN axons in the olfactory nerve and contributes to the positioning of glomeruli in the bulb [60–63] . The clade and target specific requirement for nrp1a as compared to nrp1b observed for these two paralogs demonstrates that they have evolved to acquire independent functions in axon targeting in zebrafish . Whether this difference originates in their differential patterns of expression , or differences in their functional responses to ligands , is unknown . Previous work has shown that CZ projecting axons have an nrp1a-dependent repellent response that is mediated by sema3d expressed anteriorly in the bulb [64] , but many other potential semaphorin ligands are expressed in the bulb and some of them could have additional guidance effects . The ligand ( s ) that mediate Nrp1b’s contribution to DZ targeting are unknown . Further defining the specific expression requirements and ligand responsive properties of these two neuropilins will provide an interesting perspective on the evolution and specialization of these very important signaling molecules . We find that robo2 is required for the normal targeting of DZ but not CZ projecting axons . In the fly , slit/robo interactions are required for the normal dorsal-ventral positioning of glomeruli [65] . Similarly , Slit/Robo signaling is an important determinant of dorsal-ventral targeting of OSN axons in the mouse main olfactory bulb [11–12 , 66] . In the accessory olfactory bulb , Robo2 is essential for basal vomeronasal sensory axons to specifically target the posterior compartment of the accessory olfactory bulb [67] . It is unclear whether our results are consistent with a shift along the dorsal ventral axis , or whether they represent the spreading of axons into a new compartment . To address this issue in the future , it will be necessary to identify the location at which pathfinding errors first occur and the source of the guidance signal that Robo2 is responding to . Robo2 has previously been observed to play a key role in olfactory axon guidance and entry into the olfactory bulb of fish [68] . Loss of robo2 induced defasciculation of the ON , errant OSN axons failing to enter the bulb , impaired protoglomerular organization , and in the adult , mispositioned glomeruli . Miyasaka et al . suggested that some of these effects might be a consequence of the errors made by a group of early projecting pioneer axons in the mutant , and that later extending OSNs project abnormally because they follow the misprojecting pioneers . If there was any disruption of protoglomeruli in the robo2 mutant larvae we examined , it was not sufficient to interfere with protoglomerular identification . Only occasionally did we detect BAC labeled sensory axons failing to enter the bulb ( 2/75 preparations ) . As we observed the specific perturbation of axon projections to the DZ but not the CZ , it is unlikely that the misguidance phenotype we observed was caused by general defasciculation of the ON . It is possible that Slit expression surrounding the ON helps to organize the selective fasciculation of Robo2 expressing axons within the nerve , just as Sema3A expressed in the tissues surrounding the mouse ON help to organize the selective fasciculation of Nrp1 expressing axons [61] . Alternatively , slits expressed within the telencephalon or in immediately surrounding tissues [68] may affect the trajectories of Robo2 expressing axons . Our findings suggest that OSNs first project to a particular protoglomerulus depending upon the general category of OR they choose to express . Thus , OSNs expressing many different clade A ORs all project to the CZ protoglomerulus . This is in stark contrast to the adult olfactory bulb in which axons project to a much larger number of OR-specific glomeruli . Nevertheless , the early olfactory map is not formed at random . OSNs expressing the most closely related ORs specifically target the same protoglomerulus . We were able to identify clusters of transcription factors and guidance-related molecules that are differentially expressed between three OR clade-related subgroups of OSNs . We propose a model whereby the choice of clade from which an OR is expressed is coordinated with the expression of particular axon guidance receptors . We hypothesize that some of the differentially expressed transcription factors we have identified could help coordinate OR and guidance factor expression . This model makes the initial targets of OSN axons , the protoglomeruli , intermediate targets that are further subdivided and refined later in development as the axons of OSNs expressing a particular OR aggregate together and segregate from all others to form OR-specific glomeruli . In conclusion , we have exploited the zebrafish model system and the power of single cell RNA-seq to characterize three distinct OSN subpopulations at a very early stage of olfactory system development . The guidance-related candidates we have identified likely play a crucial role in targeting specific axonal populations to their first targets in the olfactory bulb . Further studies are needed to characterize the physiological roles of the remaining candidate guidance genes in the formation of the early olfactory map . A second avenue of future work will be to explore whether and how the transcription factors we identified coordinate the expression of ORs and guidance related factors . Finally , the expression-based approach we took to identify important transcription and guidance factors should be generally applicable to many other developing systems .
Animal husbandry and procedures adhered to the guidelines prescribed by National Institutes of Health Guide for the Care and Use of Laboratory Animals . All animal studies were fully approved by the University of Pennsylvania Institutional Animal Care and Use Committee ( IACUC ) protocol numbers 804944 and 804895 . Adult zebrafish were raised and maintained according to standard procedures [69] . Veterinary care was supervised by University Laboratory Animal Resources ( ULAR ) at the University of Pennsylvania . Larvae were staged based on hours post fertilization ( hpf ) and were raised at 28°C . Tg ( omp:lyn-RFP ) line was a gift from the Yoshihara laboratory at the RIKEN Brain Science Institute , Saitama , Japan [37] . The Tg ( or111-7:or111-7-IRES:GAL4 ) , Tg ( omp:GAL4 ) and Tg ( UAS:gap43-citrine ) lines were described by Lakhina et al [48] . Tg ( BACOR111-7:IRES:GAL4 ) and Tg ( BACOR130-1:IRES:GAL4 ) were described by Shao et al [36] . Adult heterozygote fish carrying the zOMP:RFP transgene were crossed to each other to obtain embryos that expressed zOMP:RFP . Timed matings allowed us to follow the developmental stages . Embryos were bleached at 24hpf and allowed to develop till 48hpf . Chorions were removed by treatment with pronase at 24hpf . Embryos were anesthetized using tricain and dissected in Ringer’s solution to isolate the anterior half of the head containing the olfactory pits . Approximately 200 embryos were dissected for each FACs experiment . The olfactory pits were dissociated in 2 . 5% trypsin at 28°C for 30 min . At the end of incubation , the trypsin was neutralized using soy-trypsin inhibitor and the cell suspension was spun down in a clinical centrifuge at 2000rpm for 5 min . The pellet was resuspended in Hank’s Balanced Salt Solution ( HBSS ) buffer ( Ca++/Mg++ free , without Phenol Red ) with 1% BSA and 25mM HEPES . Cells were sorted based on forward scatter , side scatter and RFP fluorescence using BD FACS Aria II SORP , BD FACS Aria II U , or BD-Influx sorters ( all BD Biosciences ) . Approximately 5500 OMP:RFP expressing cells were isolated for each of two independent experiments . For OR111-7:IRES:GAL4 bulk cell isolation , adult heterozyogtes carrying zOMP:RFP;UAS:citrine were crossed to adult heterozygotes for OR111-7:IRES:GAL4 . Embryos that expressed both OMP:RFP and OR111-7:IRES:GAL4;UAS:citrine were selected for sorting . 48hpf embryos were dissected as before except that cell suspensions were sorted based on GFP and RFP fluorescence . Approximately 1000 cells that expressed both were collected in each of 3 independent experiments . Approximately 5000 OMP:RFP-expressing cells were loaded onto a Fluidigm-C1 microfluidics chip ( Fluidigm ) ( size 5-10uM ) . Single cells were sorted into individual wells and visually inspected under a microscope for capture . 49 and 47 single cells , respectively were captured in two independent runs . Wells with no cells or more than one cell were marked to be discarded . cDNA synthesis and amplification was conducted using SMARTer amplification technology ( Clontech ) . Fluidigm spike-ins or ERCC spike-ins were included in the reaction mix . cDNA yield was assessed using Agilent BioAnalyzer ( Agilent Technologies , Santa Clara , CA ) . A minimum cDNA yield of 8ng per cell was required for library preparation . 20 cells from the first run and 38 cells from the second run passed this criterion . These were independently processed for pooled library preparation using Nextera XT DNA Library prep kit ( Illumina ) . The resulting libraries were deep-sequenced on HiSeq2500 ( Illumina ) using the paired end 100-base pair protocol to a depth of 50 million reads per cell at the Next-Generation Sequencing Core of the Perelman School of Medicine , University of Pennsylvania . Sequencing adapters , poly-A tails , and unknown nucleotides ( Ns ) were removed from the ends of reads using in-house trimming software . Additionally , low confidence nucleotides ( Phred score less than 20 ) were treated as unknown and replaced with Ns . Reads shorter than 30bp after trimming were discarded . Resulting reads were aligned to the Danio rerio reference genome , build GRCz10 , and to spike-in transcript sequences using STAR ( Spliced Transcripts Alignment to a Reference ) [70] aligner version 2 . 4 . 0 . VERSE [71] running in Hierarchical Assign mode was used to assign mapped reads to gene features . The average number of mapped reads per cell was 55 million . Two of the 58 cells had less than 3 million reads mapped to the zebrafish genome and were discarded . The remaining 56 single cell samples had an average of 7 , 483 genes per cell with an average of 2 , 565 reads per gene . GEO accession id GSE103692 . https://www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? acc=GSE103692 Zebrafish genes from GRCz10 version 80 were run on TargetP 1 . 1 server ( Center for Biological Sequence Analysis , Technical University of Denmark http://www . cbs . dtu . dk/services/TargetP/ ) [42] . Genes with reliability class ( RC ) score = 1 were selected as containing a signal peptide and included in a list of candidate guidance genes . AnimalTFDB2 . 0 ( Guo Lab , College of Life Science and Technology , HUST , China http://bioinfo . life . hust . edu . cn/AnimalTFDB/ ) [47] was used to download a list of zebrafish genes with known or putative transcription factor activity . The single-cell dataset of mature cells was filtered to include only those genes that were expressed in three or more cells at a minimum raw read count of 100 . We identified linear combinations of genes that maximized between-class separation in high-dimensional gene expression covariation matrices ( where the number of features exceeds the number of observations ) using the ‘penalizedLDA’ R package version 1 . 1 , which implements the Penalized LDA as originally conceived by Witten and Tibshirani [41] . A lasso penalization of 0 . 05 was applied to the feature coefficients in the discriminant vectors . Reduced discriminant functions were generated by using the top 20 genes ranked by the absolute value of their loading values in each discriminant axis and re-computing the Penalized LDA function using this sparse 40-gene feature set . Six-fold cross-validation was implemented via the ‘penalizedLDA’ package in R to test the accuracy of the ‘sparse’ 40-gene PLDA classifier , as well as to select the appropriate lambda tuning parameter value . Top genes from single-cell PLDA analysis the using entire transcriptome or TargetP subset were used to compute a PCA function on single-cell and OR111-7-IRES-Gal4 bulk RNA-seq dataset using gene expression values that were DESeq2 normalized [72] . Correlation analysis of transcription factors and guidance genes was performed by computing the Pearson’s correlation co-efficients between genes followed by hierarchical clustering using average linkage . All heatmaps were plotted using ggplots package in PIVOT [73] using normalized gene expression values . nrp1asa1485 , nrp1bfh278 and robo2ti272 mutants have been described previously [51 , 64] . These were crossed to Tg ( BACOR111-7-:RES:GAL4 ) , Tg ( BACOR130-1:IRES:GAL4 ) , or Tg ( UAS:citrine ) to generate adult heterozygotes that carried the appropriate transgenes . Heterozygous parents were mated to obtain fluorescent progeny that were collected at 72 hpf . Genomic DNA was extracted from tails and genotyped using PCR for nrp1b ( described in [64] ) and robo2 ( described in [74] ) or KASP assay for nrp1a ( described in [64] ) . Matched heads from embryos of the same genotype were pooled and were processed for immunohistochemistry . Immunohistochemistry was performed as previously described [48] . Larvae were fixed in 4% paraformaldehyde in PBS and dehydrated in methanol . To visualize Citrine positive axons , larvae were permeabilized in acetone for 20 min at −20°C and stained with goat anti-GFP ( 1:100; Rockland Immunochemicals , 600-101-215 ) and donkey anti-goat IgG Alexa Fluor 488 ( 1:500; Invitrogen ) . Propidium iodide staining was performed following secondary antibody treatment as described by Brend and Holley , 2009 [75] with the exception that larvae were not treated with RNase . Confocal microscopy was performed on an inverted Leica SP5 using a 40× oil-immersion lens . Stacks were acquired through the entire OB with optical sections taken 1 μm apart . Double-label in situ hybridization was performed using antisense digoxigenin ( DIG ) RNA probes to guidance genes and fluorescein-labeled probes to OR subfamily genes as previously described [36 , 64] , with the exception that RNase treatment was not performed after probe removal . Following probe hybridization and removal , embryos were incubated in anti-DIG-POD ( 1:500; Roche , 11207733910 ) and the DIG label was amplified using the cyanine 5-coupled tyramide kit to label axon guidance genes . OR transcripts were detected using anti-fluorescein-POD ( 1:500; Roche , 11426346910 ) and the fluorescein label was amplified using a fluorescein-coupled tyramide kit ( PerkinElmer , NEL741001KT ) . Propidium iodide labeling and imaging were performed following the second tyramide amplification [75] . The plasmids used to make probes targeting nrp1a and nrp1b were as described by Dell et al . , 2013 [76] and Taku et al . , 2016 [64] . For robo2 ( refseq accession number NM_131633 . 1 , nucleotides 3445–4382 ) and pcdh11 ( refseq accession number XM_005173190 . 3 nucleotides 220–839 ) sequences were amplified from 48hpf zebrafish cDNA and cloned into pcRII-TOPO Dual promoter TA cloning kit ( Invitrogen , K460001 ) for probe synthesis . Full-length probes were used in all hybridization experiments except for robo2 where the probe was hydrolyzed prior to incubation . The number of OBs containing axons terminating in either individual protoglomeruli or non-protoglomerular regions were counted . Axons were scored as projecting to a particular protoglomerulus only if they terminated in that protoglomerulus and not if they passed though it en route to another location . The percentage of OBs with axons in each protoglomerulus was computed and a two-tailed Fisher’s exact test was used to determine statistical significance . Error bars represent standard error of the mean .
|
Assembling the brain circuitry responsible for our sense of olfaction is a formidable task . The sensory neurons that detect odorants in the nasal epithelium , the olfactory sensory neurons , choose to express a single odorant receptor from a large gene repertoire . Sensory neurons that choose a particular odorant receptor are intermixed with other neurons that have chosen different receptors . Remarkably , all the sensory neurons that choose a particular odorant receptor project axons to the same target region in the brain , while those choosing to express other odorant receptors project axons to different target regions . Presumably , this specificity in axon targeting is related to the differential expression of guidance related genes in subsets of sensory neurons . Is there a systematic method for their identification ? Here we use single-cell RNAseq to characterize the transcriptional profiles of individual olfactory sensory neurons . We identify candidate guidance related genes whose expression is correlated to the simultaneous expression of particular odorant receptors . We show in loss of function studies that some of these guidance related genes are required for the guidance of specific subsets of olfactory sensory neurons . This general approach may open the door to the systematic identification of genes that guide axons to their specific targets during the assembly of developing neural circuits .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"fish",
"gene",
"regulation",
"regulatory",
"proteins",
"dna-binding",
"proteins",
"vertebrates",
"neuroscience",
"animals",
"dna",
"transcription",
"animal",
"models",
"osteichthyes",
"model",
"organisms",
"experimental",
"organism",
"systems",
"transcription",
"factors",
"genome",
"analysis",
"nerve",
"fibers",
"research",
"and",
"analysis",
"methods",
"developmental",
"neuroscience",
"genomics",
"animal",
"cells",
"axons",
"proteins",
"gene",
"expression",
"olfactory",
"receptor",
"neurons",
"biochemistry",
"zebrafish",
"cellular",
"neuroscience",
"eukaryota",
"cell",
"biology",
"axon",
"guidance",
"neurons",
"genetics",
"transcriptome",
"analysis",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"afferent",
"neurons",
"computational",
"biology",
"organisms"
] |
2018
|
Coordination of olfactory receptor choice with guidance receptor expression and function in olfactory sensory neurons
|
Antibiotic-resistant Salmonella enterica serovar Paratyphi A , the agent of paratyphoid A fever , poses an emerging public health dilemma in endemic areas of Asia and among travelers , as there is no licensed vaccine . Integral to our efforts to develop a S . Paratyphi A vaccine , we addressed the role of flagella as a potential protective antigen by comparing cell-associated flagella with exported flagellin subunits expressed by attenuated strains . S . Paratyphi A strain ATCC 9150 was first deleted for the chromosomal guaBA locus , creating CVD 1901 . Further chromosomal deletions in fliD ( CVD 1901D ) or flgK ( CVD 1901K ) were then engineered , resulting in the export of unpolymerized FliC , without impairing its overall expression . The virulence of the resulting isogenic strains was examined using a novel mouse LD50 model to accommodate the human-host restricted S . Paratyphi A . The immunogenicity of the attenuated strains was then tested using a mouse intranasal model , followed by intraperitoneal challenge with wildtype ATCC 9150 . Mucosal ( intranasal ) immunization of mice with strain CVD 1901 expressing cell-associated flagella conferred superior protection ( vaccine efficacy [VE] , 90% ) against a lethal intraperitoneal challenge , compared with the flagellin monomer-exporting mutants CVD 1901K ( 30% VE ) or CVD 1901D ( 47% VE ) . The superior protection induced by CVD 1901 with its cell-attached flagella was associated with an increased IgG2a∶IgG1 ratio of FliC-specific antibodies with enhanced opsonophagocytic capacity . Our results clearly suggest that enhanced anti-FliC antibody-mediated clearance of S . Paratyphi A by phagocytic cells , induced by vaccines expressing cell-associated rather than exported FliC , might be contributing to the vaccine-induced protection from S . Paratyphi A challenge in vivo . We speculate that an excess of IgG1 anti-FliC antibodies induced by the exported FliC may compete with the IgG2a subtype and block binding to specific phagocyte Fc receptors that are critical for clearing an S . Paratyphi A infection .
Four human host-restricted Salmonella enterica serovars cause clinically indistinguishable typhoid ( Salmonella Typhi ) and paratyphoid ( S . Paratyphi A , B and [uncommonly] C ) fever [1] . Multiply antibiotic resistant S . Paratyphi A have emerged in Asia , accompanied by increased incidences of paratyphoid fever in endemic populations [2] , [3] and in travelers [4] . Whereas vaccines exist to prevent typhoid fever , there is no licensed vaccine to prevent S . Paratyphi A disease . Vaccines for preventing typhoid fever include the purified Vi capsular polysaccharide administered parenterally and attenuated Vi-negative strain Ty21a given orally as a live vaccine [5] . Vi conjugated to recombinant exotoxin A of Pseudomonas aeruginosa conferred on Vietnamese children a high level of efficacy in a field trial [6] , [7] . Vi-based vaccines cannot protect against paratyphoid disease as S . Paratyphi A does not express Vi . Oral Ty21a confers moderate cross protection against S . Paratyphi B [8] but not S . Paratyphi A [9] . Despite the public health need [4] , there have been few reports on modern S . Paratyphi A vaccine development [10] , [11] . Attenuated Salmonella strains can be employed as mucosally-delivered vaccines or as “reagent strains” to achieve safe , high-yield production of purified antigens for manufacture of parenteral ( conjugate ) vaccines . A Salmonella surface antigen that has generated renewed interest in the role that it may play in protection is the flagellum . Flagella mediate intestinal epithelial and macrophage inflammation following infection and contribute to early host innate immune responses against Salmonella [12] . Flagellin ( FliC ) , the monomer of flagellar filaments that induces these effects , is being incorporated into fusion proteins linked to otherwise poorly immunogenic antigens and haptens , providing adjuvant activity to enhance immune responses to those moieties [13]–[15] . The flagellum is a complex motility organelle composed of >20 different proteins that form a basal body , hook , filament and an export system . The major extracellular part of the flagellum comprises ∼20 , 000 FliC monomers that are exported and assembled at the terminus of a growing filament . Between the hook and filament is a short junction formed by two hook-associated proteins , FlgK and FlgL [16] , [17] . S . Typhimurium mutants defective in FlgK or FlgL synthesize FliC monomers that do not polymerize and are released into the culture medium [17] . A capping structure of five FliD molecules at the end of the filament also promotes FliC polymerization [16] , [18] , [19] . Deletion of fliD in S . Typhimurium incapacitates the ability of transported FliC to polymerize [20] , [21] . Integral to our efforts to develop a S . Paratyphi A vaccine , we addressed the role of flagella as a potential protective antigen by comparing cell-associated flagella with exported flagellin subunits expressed by attenuated strains . Mutants were constructed with deletions in fliD or flgK , resulting in export of unpolymerized FliC , without impairing its overall expression . These strains allowed us to investigate whether expression of FliC as cell-attached flagellin filaments versus exported monomers , would influence the immune response or protection elicited by these live vaccines .
All animal experiments carried out in this work were approved by the University of Maryland Baltimore Office of Animal Welfare Assurance ( OAWA ) , under approved Animal Use Protocol 0409006 . S . Paratyphi A wild-type and mutant strains ( Table 1 ) were propagated on animal product-free LB Lennox medium ( Athena ES , Baltimore , MD ) . Lennox agar plates were prepared by addition of 1 . 5% agar ( Difco , BD , Franklin Lakes , NJ ) . Guanine ( 0 . 001% v/v ) was added for ΔguaBA mutant strains . Liquid cultures were incubated at 37°C , 250 rpm , at a ratio of 1∶10–1∶20 vol∶vol medium∶flask ( high-aeration conditions ) . For low-aeration growth conditions , the flasks were filled to 75% of their volume with the medium , and shaken at 80 rpm . Time course experiments in liquid culture were seeded with an overnight culture , inoculated to 0 . 01 OD600; samples were removed at regular intervals for determining culture turbidity at OD600 or plating . For each growth experiment , two flasks were cultured per strain , and each experiment was performed twice . Swimming and swarming assays were carried out on fresh Lennox plates containing 0 . 3% and 0 . 5–0 . 7% agar , respectively . Glucose was added to a final concentration of 0 . 5% for swarm plates . The swim plates were inoculated by stabbing the center with bacteria harvested from 1 . 5% Lennox plates . The swarm plates were inoculated by pipetting a 10-µl fresh culture , grown in liquid Lennox media to 0 . 5–0 . 6 OD600 , onto the surface of the center of the agar plate . All motility experiments were performed in triplicates and repeated at least twice . Swim and swarm ability are expressed as the radius of the mobility zone; for no swim , the radius of growth was >1 mm and for no swam , 10 mm . Deletion of fliD and flgK genes was performed by λ Red-mediated mutagenesis [22] essentially as described [23] . Primers ( listed in Table 2 ) were designed to replace most of the gene of interest with a kanamycin resistance cassette flanked by the Flippase Recombination Targets , FRTs . The kanamycin cassette was later deleted via λ Red recombinase , leaving an FRT scar sequence . Bacterial protein samples were normalized as follows . Cell pellets were washed in 0 . 125 M Tris-HCl , pH 6 . 8 , brought to 10 OD600 in the same buffer and diluted 1∶3 with Laemmli sample buffer ( Bio-Rad Laboratories , Hercules , CA ) . Supernatants were brought to the equivalent lowest OD600 culture per experiment , by addition of 0 . 125 M Tris-HCl , pH 6 . 8 , and diluted 1∶1 with Laemmli buffer . The protein samples were boiled for 10 min , and 10-µl aliquots were loaded onto 10% SDS-PAGE gels . For anti-FliC blots , monoclonal antibodies ( BioVeris , Gaithersburg , MD ) diluted 1∶1000 were used for 1 h incubation . Detection was performed with secondary peroxidase-labeled goat anti-mouse IgG ( KPL , Gaithersburg , MD ) , followed by application of the ECL PLUS Western blotting detection system ( GE Healthcare , Buckinghamshire , UK ) . Coomassie blue-stained gels and developed blots were scanned with a V700 Photo EPSON Scan ( digital ICE technologies ) using SilverFast SE imaging software ( LaserSoft Imaging , Sarasota , FL ) , and quantitated with QuantityOne software ( Bio-Rad ) . S . Paratyphi A FliC was prepared from strain CVD 1902 using the shearing . CVD 1902 was chosen for purification of flagella for two reasons . First , it was genetically engineered from the attenuated strain CVD 1901 to hyper-express flagellin by deletion of the clpX gene ( Table 1 ) , which together with clpP encodes the ClpXP ATP-dependent protease that degrades the master flagella positive regulator complex FlhD/FlhC , resulting in large amounts of flagella being over-produced [24] . Second , it is an attenuated strain and as such does not pose an occupational risk when cultured in large volumes for antigen purification . Bacterial cultures were grown overnight under low aeration conditions in 2-liter flasks containing Lennox broth supplemented with guanine . Cell pellets were washed and resuspended in PBS , and sheared for 3 min at high speed in a Waring laboratory blender . The sheared suspension was centrifuged twice at 7 , 000×g for 10 min , and the supernatant was collected and centrifuged at 100 , 000×g for 3 h to pellet the filaments . The pellet was suspended in saline at 4°C overnight , centrifuged at 7 , 000×g , and the clear supernatant containing flagellar filaments was transferred to a new tube . Protein concentration was determined with the BCA assay ( Pierce , Rockford , IL ) . Purity was assessed by SDS-PAGE and Coomassie blue staining . The amount of contaminating LPS was quantified using the resorcinol sulfuric acid assay [25] using a standard curve generated with purified S . Paratyphi A LPS . This FliC preparation was determined to be 98 . 3% pure . UF membranes with different Mr cut-offs were used for gradient separation of flagellin monomers from purified filaments . Supernatants collected from bacterial cultures grown under high-aeration conditions and containing flagellin monomers and/or sheared flagella , were first normalized to equal concentrations of FliC by passing supernatants through a 30-kDa cut-off Amicon membrane ( Millipore , Billerica , MA ) . The concentrated retentants were then passed through a 100-kDa cut-off Amicon membrane . The resulting filtrates were further passed through an additional 30-kDa membrane . Bacteria collected from swarm colonies were suspended in PBS to an OD600 of 1 . 0 and were incubated with 300 mesh Formvar coated copper grids ( Electron Microscopy Services , Hatfield , PA ) for 20 min . Grids were gently blotted and placed on 50 µl drops of 2% ammonium molybdate ( Sigma Aldrich , St . Louis , MO ) for 2 min . After air-drying , grids were observed with a JEOL electron microscope JEM-1200EX ( JEOL , Toyko , Japan ) . For all experiments , 6 week-old female BALB/c mice were purchased from Charles River Breeding Laboratory Inc . ( Wilmington , MA ) and maintained in a biohazard animal facility . Anesthesia ( isofluorane dispensed through a precision vaporizer ) was used for blood collection from the retro-orbital plexus . All studies were approved by the Institutional Animal Care and Use Committee ( IACUC ) of the University of Maryland Baltimore School of Medicine , and conducted in accordance with NIH guidelines [23] . For assessment of virulence , the hog gastric mucin assay was used [26] . Mice were injected by the intraperitoneal route ( i . p . ) with increasing 10-fold dilutions of bacteria; bacteria were harvested from overnight Lennox plates and suspended in PBS mixed with 10% ( wt/v ) Difco hog gastric mucin ( Becton-Dickinson , Sparks , MD ) to a final volume of 0 . 5 ml per mouse . Groups of six mice per dose per strain were tested . Mice were observed twice daily for 3 days for mortality or any signs of significant morbidity ( ruffled fur , weight loss of 20% or more , collapse , difficulty breathing or severe dehydration ) , and those showing the above signs were euthanized according to IACUC directives . 72 h post-challenge , surviving mice were euthanized using CO2 asphyxiation followed by cervical dislocation . LD50 values were calculated by logistic regression analysis . Fresh vegetative cultures of CVD 1901 , CVD 1901D or CVD 1901K were pelleted , washed with PBS , and brought to a final concentration of ∼1011 cfu/ml . 10-µl aliquots were applied intranasally on day 0 , 14 , and 28 to mice ( 5 µl/nostril , ∼109 cfu per mouse; 10–15 mice per group ) . A group immunized with PBS served as a negative control . Blood samples were collected prior to and after immunization and sera were stored at −70°C . Mice were challenged on day 56 with 3 . 3×105 cfu per mouse of wild-type ATCC 9150 S . Paratyphi A , freshly prepared as described above for the LD50 studies . Iron ( 8 µl of 5% ammonium iron ( III ) citrate to 1 ml of mucin ) was added to the bacterial suspension to increase virulence [27] . Following the challenge , mice were monitored every 6 h for 72 h for mortality or any signs of significant morbidity . Total IgG antibodies and IgG subclasses against S . Paratyphi A flagella were determined by ELISA as previously described [28] . Briefly , 96-well plates were coated with S . Paratyphi A flagella ( 5 µg/well ) or LPS ( 10 µg/ml ) . Samples were diluted in 10% dried milk in PBS containing 0 . 05% Tween 20 ( PBSTM ) and tested in duplicates . Specific antibodies were detected using HRP-labeled goat anti-mouse IgG , IgG1 and IgG2 ( KPL Inc . Gaithersburg , MD ) diluted in PBSTM followed by TMB Microwell Peroxidase Substrate solution ( KPL ) . Titers were calculated by interpolation in a standard curve as the inverse of the dilution that produces an OD value of 0 . 2 above the blank ( ELISA units/ml ) . Bactericidal activity was assessed by a complement-mediated lysis of S . Paratyphi A using sera from immunized mice . Fresh vegetative wild-type ATCC 9150 ( flagellated ) or 9150K ( non-flagellated ) bacterial suspensions ( 106 cfu/ml ) were mixed with 30% guinea pig complement ( Sigma-Aldrich , St . Louis , MO ) , added ( 1∶1 ) to heat-inactivated ( 56°C , 20 min ) serially-diluted mouse sera , and incubated for 1 h at 37°C . Following incubation , bacteria were counted by plating . End point titers were defined as the last dilution that induced a ≥50% reduction in the number of bacteria incubated with complement alone without addition of serum . Antibody-mediated bacterial uptake by macrophages was measured by seeding J774A . 1 cells into 24-well microdilution plates and growing in DMEM supplemented with 5% FCS at 37°C with 5% CO2 to a confluent layer ( 2×105 cells/well ) . Fresh vegetative wild-type ATCC 9150 bacteria were incubated with heat inactivated mouse serum ( 10% in PBS ) for 30 min on ice , then added to the cell monolayer at a ratio of 1∶1 . Following centrifugation ( 100×g , 10 min ) , the microdilution plate containing the monolayer and opsonized bacteria was incubated at 37°C with 5% CO2 for 30–45 min . External bacteria were removed by replacing the media with fresh media containing 100 µg/ml gentamicin , incubating for 30–45 min , followed by three PBS washes . The cells were lysed with 0 . 5% Triton x-100 and internalized bacteria were counted by plating . LD50 was estimated by logistic regression analysis . Continuous variables were compared among the groups using Kruskal-Wallis analysis of variance . Proportions in two groups were compared using the Fisher's exact test . Two-sided p-values<0 . 05 and one-sided p-values<0 . 025 were considered statistically significant .
Four S . Paratyphi A wild-type strains were compared for growth rate , bacterial cell yield and ability to express FliC , including American Type Culture Collection strain ATCC 9150 and three clinical isolates Q82b ( Mali ) , EAR6473 ( Chile ) and 15 . 067 ( Chile ) . No significant differences in growth rate ( Fig . 1A ) or overall protein electrophoretic profiles ( Fig . 1C ) were observed among the strains , but ATCC 9150 consistently reached the highest cell yield when grown in rich liquid medium ( Fig . 1B ) . FliC was a major component of the secretomes of all four strains ( Fig . 1C , middle panel ) . The high amounts of flagellar protein found in the supernatants are a consequence of shear forces acting upon the cells during growth in shake flasks , causing filament shearing from the cell surface . Among the four strains , the Chilean isolate 15 . 067 expressed lower levels of FliC ( Fig . 1C , lower panel ) . To compare flagellar protein expression further , “swim” and “swarm” motilities were tested , each providing evidence of flagella functionality [29] . Swimming is assayed by growing the bacteria on semisolid medium ( 0 . 2–0 . 4% agar ) where bacterial cells swim through water-filled channels in the agar , whereas swarming is observed following inoculation on the surface of solid medium ( 0 . 5–0 . 8% agar ) . Notably , swarming is associated with greater flagella expression than swimming [29] . In accordance with FliC expression ( Fig . 1C ) , strain 15 . 067 had reduced motility ( Fig . 1D ) . We next tested whether variations of flagellar expression affect virulence in mice . Since S . Paratyphi A is avirulent in mice when administered orally or intranasally ( i . n . ) , we adopted a mouse model used for S . Typhi ( another human host-restricted pathogen ) to determine LD50 , in which bacteria are suspended in hog gastric mucin prior to intraperitoneal ( i . p . ) injection of BALB/c mice . This model has been used to assess the attenuation of candidate oral S . Typhi vaccines pre-clinically [26] , [30] and reasonably predicted responses of humans given those strains in Phase 1 trials [31] , [32] . Accordingly , young mice were injected i . p . with 10-fold dilutions of bacteria in 10% ( w∶v ) hog gastric mucin . ATCC 9150 was the most virulent , with an LD50 value of 52 cfu/mouse , while strains Q82b , EAR6473 and 15 . 067 exhibited LD50 values of 846 , 199 and 692 cfu/mouse , respectively ( Fig . 1E ) . ATCC 9150 , with its excellent growth characteristics , copious flagella production and high virulence in mice , was therefore selected as the wild-type parent for construction of our vaccine candidates; the available genomic sequence of ATCC 9150 provided another rationale for using this strain [33] . To export flagellin as monomers , we targeted two chromosomal loci , FliD ( flagellar cap protein ) and FlgK ( a hook-filament junction protein ) , shown in S . Typhimurium to encode hook-associated proteins [17] , [20] , [34] . To assure the safety of our candidate strains , we first introduced a deletion in the chromosomal guaBA operon of ATCC 9150 , which encodes two essential enzymes , inosine monophosphate dehydrogenase ( GuaB ) and guanine monphosphate synthetase ( GuaA ) , involved in the de novo guanine nucleotide biosynthesis pathway . Resulting strain CVD 1901 was then further deleted for fliD ( yielding CVD 1901D ) or flgK ( CVD 1901K ) . The fliD or flgK deletions were also introduced into ATCC 9150 , leading to 9150D and 9150K . ATCC 9150 and the resulting five isogenic strains possessed indistinguishable growth rates when cultured under high aeration in appropriately supplemented rich liquid medium ( data not shown ) . Thus , neither the guaBA , fliD or flgK mutations impaired growth , although bacterial cell yields were somewhat lower compared to ATCC 9150 ( Fig . 2A top panel ) . As expected , analysis of secreted protein confirmed that higher FliC levels were found in supernatants of the ΔfliD and ΔflgK mutants compared to supernatants from their parental strains ( Fig . 2A lower panel ) . Differences in FliC expression between the fliD and flgK mutants and their parents were pronounced when bacteria were propagated as stationary broth cultures ( low-aeration ) where shear forces acting upon the cells are much lower . Under these conditions , no free FliC was observed in supernatants of either parental strain , yet FliC levels in supernatants of the fliD and flgK mutant cultures were as high as when grown under aerated conditions ( shown for CVD 1901 and derived mutants , Fig . 2B ) . Immunoblotting with anti-FliC under low-aeration growth showed that the parental strain retained most flagella on the bacterial cells . In contrast , the mutants exhibited almost undetectable levels of FliC on the cell surface; CVD 1901D showed some residual FliC , while none was detected on the surface of CVD 1901K ( Fig . 2C ) . Export of FliC from the mutants was elucidated by a three-step characterization ( Fig . 3A schema ) , using serial ultrafiltration ( UF ) membranes . First , conditioned media of CVD 1901 , CVD 1901D and CVD 1901K cultures were adjusted to equivalent FliC concentrations using 30-kDa cutoff UF ( Fig . 3A upper gel ) . Then , 100-kDa ( Fig . 3A middle gel ) followed by 30-kDa cutoff membranes allowed separation of monomeric from polymeric flagellin . Following the final 30-kDa passage , CVD 1901K supernatant contained the highest amount of flagellin , with less in the CVD 1901D supernatant and almost none in concentrate of the parental CVD 1901 strain ( Fig . 3A lower ) . Thus , flagellin molecules in supernatants from cultures of the fliD and flgK mutants are in the unassembled form . However , in accordance with the results shown in Fig . 2C , some flagellin expressed by the CVD 1901D mutant is cell-associated , sheared off the cell surface during growth , and retained within the 100-kDa filter . As a control , all culture supernatants were heat-treated to dissociate the polymeric flagella and indeed some FliC was observed in the filtrate of the parental CVD 1901 strain following the 100-kDa UF ( Fig . 3A middle and lower ) . Second , comparing flagella functionality revealed identical swim and swarm diameters ( % of plate ) for ATCC 9150 versus CVD 1901; 39% versus 43% for swim and 45% versus 69% for swarm , respectively ( Fig . 3B ) . Strains 9150D and CVD 1901D showed only swimming motility with a diameter of 11% and no swarming , while 9150K and CVD 1901K showed neither swim nor swarm motility . Decreasing the agar concentration in the swarming plates allowed swarming of the ΔfliD mutants , while the ΔflgK mutants remained non-motile ( Fig . 3B ) . Finally , bacterial samples from the 0 . 7% swarm plates were examined by electron microscopy ( EM ) . The negative stained bacterial cell images establish that the ΔflgK mutant is completely devoid of surface flagella , while the ΔfliD mutant carries one or two filaments ( Fig . 3C ) . Virulence of the mutants was compared to wild-type ATCC 9150 by inoculating 6 week-old mice i . p . with bacteria suspended in hog gastric mucin . An LD50 value of 8 . 8 bacteria per mouse was calculated for the wild-type ( Fig . 4 ) . In contrast , CVD 1901 showed an LD50 of 3 . 0×107 cfu/mouse . The LD50s for 9150D , 9150K , CVD 1901D and CVD 1901K were 17 , 49 , 3 . 4×106 and 2 . 0×107 cfu/mouse , respectively ( Fig . 4 ) . These results show clear attenuation only for strains harboring the guaBA deletion . Deletion of either fliD or flgK from ATCC 9150 attenuated the resulting strain by only half a log and did not work synergistically with the guaBA deletion . Whereas deletion of flgK or fliD in S . Paratyphi A did not alter bacterial virulence in this model , these mutations may nevertheless influence the protection conferred by vaccine strains also deleted in guaBA . Accordingly , we examined the ability of these mutants to protect against a S . Paratyphi A lethal challenge . Mice were immunized i . n . with ∼1×109 cfu of CVD 1901 , CVD 1901D or CVD 1901K on days 0 , 14 and 28 . Control mice received PBS . The i . n . route was chosen based on the robust immunity [35] , [36] and protection [37] induced by attenuated S . Typhi administered by this route . Three weeks after the last immunization , mice were challenged with a lethal dose of wild-type ATCC 9150 ( 3 . 3×105 cfu/mouse plus iron , see Materials and Methods ) . All control mice succumbed within 24 hours post-challenge ( Table 3 ) . The vaccine strains differed in their protective capacity , with CVD 1901 conferring significantly superior protection compared to CVD 1901D or 1901K . Serum IgG antibodies against FliC and LPS rose progressively after each immunization ( Fig . 5A ) , reaching similar levels for all strains . Unlike CVD 1901 ( which expresses many ) and CVD 1901D ( expresses a few ) surface-associated flagella , CVD 1901K is devoid of flagella at the time of administration . Hence , FliC antibodies induced by CVD 1901K represent responses to de novo FliC synthesized in vivo , rather than antigen present at the time of immunization . The slightly higher titers detected in the mice immunized with CVD 1901D may reflect the combined effect of surface-associated and secreted FliC . Overall , there was no significant correlation between anti-FliC or anti-LPS antibody titers and survival ( Fig . 5B ) , nor a correlation of survival with antibodies to S . Paratyphi A outer membrane protein fractions ( data not shown ) . Thus , no serum IgG responses against major Salmonella surface antigens correlated with protection in this model . Since the overall level of anti-FliC IgG among the immunized groups did not correlate with efficacy , we examined IgG antibody subtypes . While all three vaccine strains induced similar levels of anti-FliC IgG2a antibody , CVD 1901D and CVD 1901K induced strikingly high levels of anti-FliC IgG1 , which were 50- and 10-fold higher than those induced by CVD 1901 , respectively ( Fig . 6A , p = 0 . 012 ) . The IgG2a∶IgG1 geometric mean ratios were 1 . 2 , 0 . 036 , and 0 . 051 for CVD 1901 , CVD 1901D , and CVD 1901K respectively , implying that a functional IgG2a-biased response , rather than elevated ( and likely competing ) IgG1 antibodies , might correlate with enhanced protection ( Fig . 6B; p = 0 . 042 ) . Noting the differences in anti-FliC IgG subtype antibody responses induced by the different live vaccines , we next studied functional activity of the antibodies . Antibody switching to different IgG subclasses requires T-cell help ( TH ) during antigen priming; the presence of IgG1 reflects TH2 subset activity , whereas IgG2a indicates a TH1-type response . Since live vaccine carrying cell-associated FliC exhibited higher potency compared with flagellin-secreting strains , we further examined the potential contribution of the TH1-associated antibody response induced by these strains in protection against S . Paratyphi A challenge . The TH1 subset is responsible for many cell-mediated functions and favors the production of IgG2a antibodies with opsonophagocytic capacity that bind to high-affinity Fc receptors on macrophages [38] . These antibodies activate the complement system more readily than IgG1 antibodies [39] and efficiently mediate antibody-dependent cell-mediated cytotoxicity [40] . Complement-mediated antibody killing ( bactericidal ) of wild-type ATCC 9150 was assessed by incubating bacteria with serial dilutions of heat-inactivated sera from immunized mice to which guinea pig complement was added . Sera from naïve mice established the background activity . For the three sera sets , analogous bactericidal titers against ATCC 9150 were detected with no significant differences between the groups ( Fig . 7A , p = 0 . 28 ) . Similar results were obtained when the assay was repeated with the non-flagellated 9150K strain ( Fig . 7A , p = 0 . 22 ) , indicating that anti-FliC antibodies do not play a major role of S . Paratyphi A complement-mediated killing in this mouse model . We next examined opsonophagocytic activity using a macrophage culture assay that probes the ability of the sera to facilitate uptake of ATCC 9150 . Average numbers of intracellular bacteria of 3862 , 2383 and 2131 per 5×106 bacteria per well were recovered for sera from mice immunized with CVD 1901 , CVD 1901D , and CVD 1901K , respectively ( Fig . 7B ) , indicating a clear increased uptake for CVD 1901 sera ( p = 0 . 0002 ) . These data suggest enhanced anti-FliC antibody-mediated clearance of the organism by phagocytic cells induced by vaccines expressing cell-associated rather than exported FliC , which might be contributing to the vaccine-induced survival from S . Paratyphi A challenge in vivo .
Flagellar protein is highly immunogenic and immunomodulatory via stimulation of TLR5 , yet questions remain over its role in mediating protection against Salmonella [41]–[45] . Whereas purified Phase 1 flagella filaments or FliC subunits from S . Typhimurium [46] or S . Paratyphi A [10] inoculated parenterally protect mice against parenteral challenge with wild-type Salmonella of the homologous serovar , equipoise exists over whether flagellar protein contributes to protection when presented by live mucosal or parenteral inactivated whole cell vaccines . Flagellin expression is not needed for live oral S . Typhimurium vaccines to protect against wild-type challenge [47] , while human studies indicate an important role for cell-associated flagella in the protection conferred by parenteral inactivated whole cell typhoid vaccines . Inactivated whole cell vaccines ( most derived from wild-type strain Ty2 ) that provided superior protection also elicited higher anti-flagellar antibodies [48] , [49] . Importantly , no efficacy was observed in a large-scale controlled field trial when the inactivated vaccine was based on non-flagellated S . Typhi mutant TNM1 , derived from strain Ty2 [50] , suggesting that inactivated whole cell vaccines must express flagella in order to protect humans [50] . We employed attenuated S . Paratyphi A to investigate the protective capacity of the flagellar subunit protein FliC expressed by live mucosal vaccines . Since mucosally-administered live vaccines assure in vivo expression and presentation of flagellar antigens in a native form , we engineered S . Paratyphi A ATCC 9150 with specific deletions affecting flagella filament biosynthesis . SDS-PAGE and western blotting ( Fig . 2A–C ) , UF fractionation ( Fig . 3A ) , motility assays ( Fig . 3B ) and EM ( Fig . 3C ) established the phenotypes of these FliC-exporting mutants . While ATCC 9150 and CVD 1901 almost exclusively produce flagellar protein as polymer filaments , 9150D and CVD 1901D ( lacking flagellum cap protein ) express only one or two filaments . In contrast , 9150K and CVD 1901K ( lacking a flagellar hook-associated protein ) are completely devoid of flagella . Studies with S . Typhimurium explain why fliD mutants carry a few intact flagellum polymers . In ΔfliD S . Typhimurium mutants [17] , the tips of the hooks are intact and serve as effective heteronuclei for soluble FliC units to re-associate and form a functional flagellum . A filament already initiated has higher affinity for the newly added monomers , which elongate the single filament rather than form more short filaments . Our engineered strains mutated in fliD or flgK but unaltered in growth characteristics or virulence ( Fig . 4 ) provided a uniform background to explore the contribution of flagellar protein to immunity and protection . Mucosal immunization of mice with the live vaccines followed by subsequent i . p . challenge with virulent S . Paratyphi A showed prominent differences in the level of protection conferred in the face of a potent challenge that killed 100% of control mice ( Table 3 ) . CVD 1901 carrying many intact cell-attached flagella conferred 90% vaccine efficacy ( p = 0 . 0002 ) , while CVD 1901D with just a few cell-attached flagella provided 47% efficacy ( p = 0 . 026 ) ; CVD 1901K with no attached flagella elicited only 30% efficacy ( p = 0 . 15 ) . Mortality was significantly lower in CVD 1901 recipients ( 1/10 mice ) than in mice immunized with CVD 1901K ( 7/10 mice , p = 0 . 01 ) or CVD 1901D ( 8/15 mice , p = 0 . 04 ) . These data indicate an advantage for live mucosal Salmonella vaccines having cell-attached FliC filaments . The serum anti-FliC IgG titer did not correlate with protection . In fact , the less protective live vaccines that export flagellin subunits actually stimulated slightly higher anti-FliC IgG antibody titers ( Fig . 5A–B ) . However , when we dissected the anti-FliC response we found that while all three vaccine strains induced IgG2a anti-FliC , the flagellin-secreting strains ( CVD 1901D and 1901K ) induced significantly higher IgG1 anti-FliC titers ( Fig . 6 , p = 0 . 012 ) . IgG1 or IgG2a serum antibody responses in mice imply induction of TH2- or TH1-type subsets , respectively . Thus , both soluble and cell-attached polymeric FliC evoke TH1-directed switching to IgG2a but the IgG1 response is related to context . Whereas strong serum IgG1 anti-FliC responses were elicited by soluble exported FliC , this protein did not induce a strong IgG1-dependent TH2 response when presented as a bacterial cell-attached polymer , as observed for CVD 1901 . Others have reported that mucosal administration of purified S . Typhimurium flagellin elicits a strong TH2-type response [51]–[53] , while attached flagella on S . Typhimurium induce predominantly a TH1-dependent response [53] . Thus , the type of response against S . Typhimurium FliC did not seem to be determined by any intrinsic properties of FliC but rather appeared to be influenced by the form in which FliC was encountered , either as a soluble or cell-associated antigen [53] . Both monomeric and polymeric FliC induced TH2 responses provided these proteins were intact and not attached to cells , and only FliC on cells induced mainly a TH1 response [53] . Similar observations have been shown for Lactobacillus-associated and soluble FliC [54] , indicating a general mechanism for anti-FliC antibody switching . We document the identical behavior for S . Paratyphi A FliC protein . However , our study advances the field since , in contrast to earlier reports , we presented FliC in vivo as attached whole flagella or as FliC monomers exported by live bacteria; previously , flagellin monomers or polymers were administered as purified protein . Although many studies have described the immune responses to FliC , we report an association between a specific IgG subtype and protection . Vaccination with monomeric flagellin-exporting live vaccines induced stronger IgG1 anti-FliC responses but less protection against challenge with virulent S . Paratyphi A , suggesting that a pronounced TH2 response does not predict functional immunity . One must ponder why certain specific antibodies elicited following vaccination with whole organisms fail to protect . One possibility is that some antibodies in excess may compete or block specific phagocyte Fc receptors that endocytose or phagocytose antibody-coated microorganisms [55] . Overwhelming Fc receptors may interfere with clearance of the pathogen . An excess of IgG1 anti-FliC antibodies induced by the exported FliC may prevent binding of the IgG2a subtype anti-FliC that is critical for clearing S . Paratyphi A infection . Alternatively , certain antibodies that lack relevant biological activity may actually enhance rather than control infection; thus , the high IgG1 anti-FliC may enhance the uptake of bacteria into cells without triggering killing . Antibodies that engage Fc receptors and enhance infection have long been known for Chlamydia trachomatis [56] . The TH1 subset is responsible for cell-mediated functions such as activation of cytotoxic T cells and production of opsonization-promoting IgG antibodies that bind to high-affinity Fc receptors and interact with the complement system . TH1 cells produce IL-2 and IFN-γ that promote the differentiation of fully cytotoxic TC cells , which are suited to respond to intracellular pathogens . IFN-γ is a defining cytokine of the TH1 subset and activates macrophages to increase microbicidal activity [57] . IFN-γ secretion by TH1 cells also induces antibody class switching to IgG classes ( IgG2a in the mouse ) that support phagocytosis and complement fixation . Finally , in immunized mice most Salmonella-specific cells secreting IFN-γ are also FliC-specific [46] . Complement-dependent serum bactericidal antibodies [SBA] are one functional serological response in humans exposed to Salmonella pathogens or vaccines [58] . Our results show that SBA did not correlate with protection ( Fig . 7A ) . Replacing the flagellated WT S . Paratyphi A ATCC 9150 utilized in the SBA assay with non-flagellated strain 9150K , did not alter SBA titers , indicating that anti-FliC antibodies did not function in this assay or contributed only inconsequentially . SBA must bind antigenic determinants on or very near the bacterial surface so that complement may interact with lipid membrane components . Thus , antibodies directed against flagella may be ineffective as SBA . By contrast , we found antibodies mediating opsonophagocytic activity to be a useful functional correlate of protection against S . Paratyphi A ( Fig . 7B ) , as has also been proposed for invasive non-typhoidal Salmonella infections [59] . A difference was observed in the opsonophagocytic activity of sera collected from mice immunized with CVD 1901 versus sera from mice immunized with CVD 1901D or 1901K . Sera from CVD 1901 mice readily mediated uptake of virulent S . Paratyphi A by phagocytes . For every human vaccine for which a correlate of immunity exists [60] , the correlate is a serum antibody . Live Salmonella vaccines elicit both cell-mediated and antibody responses and our studies shed light on the role of immune responses to flagellar protein elicited by a rationally-attenuated S . Paratyphi A live vaccine . In addition , the unpolymerized FliC monomers produced by our engineered strains could be exploited as subunit vaccines or as a platform for engineered FliC-heterologous antigen fusions that might not otherwise be supported within the confines of a polymerized flagellum . An additional byproduct of our research is the demonstration of a small animal model that allows , following mucosal immunization , the efficacy of live S . Paratyphi A vaccines to be assessed , including discernment among various candidates .
|
Salmonella enterica serovar Paratyphi A is a pathogen that causes a systemic disease that is marked by serious complications and , if untreated , high mortality . The study of S . Paratyphi A pathogenesis and vaccine development has been extremely challenging since S . Paratyphi A is human host-restricted and no appropriate animal model exists . Since there is currently no licensed vaccine to prevent paratyphoid fever caused by this organism , our study represents a pioneering attempt to develop and refine a vaccine against S . Paratyphi A . We employed live attenuated strains which allow in vivo presentation of bacterial antigens via the natural route of infection , without the complications associated with antigen production and purification for subunit vaccines . For determining protective immunity against infection , we developed a mouse model that allowed evaluation of vaccine efficacy . We used our system to examine the protective capacity of a major Salmonella antigen , the flagellum . Due to its unique immunogenic properties , the flagellum is considered a major immune mediator , but its role in protection is controversial . We clearly show that cell-associated flagellar protein , presented by mucosally administered attenuated bacterial live vaccines , provides superior protection when compared to strains exporting FliC monomers , and we discuss possible mechanisms of immunity .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"humoral",
"immunity",
"medicine",
"genetic",
"mutation",
"immunity",
"to",
"infections",
"immunology",
"microbiology",
"bacterial",
"diseases",
"adaptive",
"immunity",
"infectious",
"disease",
"control",
"bacterial",
"pathogens",
"immunizations",
"infectious",
"diseases",
"biology",
"salmonella",
"immune",
"response",
"mutagenesis",
"immunity",
"genetics",
"genetics",
"and",
"genomics"
] |
2011
|
Cell-Associated Flagella Enhance the Protection Conferred by Mucosally-Administered Attenuated Salmonella Paratyphi A Vaccines
|
Siderophore biosynthesis by the highly lethal mould Aspergillus fumigatus is essential for virulence , but non-existent in humans , presenting a rare opportunity to strategize therapeutically against this pathogen . We have previously demonstrated that A . fumigatus excretes fusarinine C and triacetylfusarinine C to capture extracellular iron , and uses ferricrocin for hyphal iron storage . Here , we delineate pathways of intra- and extracellular siderophore biosynthesis and show that A . fumigatus synthesizes a developmentally regulated fourth siderophore , termed hydroxyferricrocin , employed for conidial iron storage . By inactivation of the nonribosomal peptide synthetase SidC , we demonstrate that the intracellular siderophores are required for germ tube formation , asexual sporulation , resistance to oxidative stress , catalase A activity , and virulence . Restoration of the conidial hydroxyferricrocin content partially rescues the virulence of the apathogenic siderophore null mutant ΔsidA , demonstrating an important role for the conidial siderophore during initiation of infection . Abrogation of extracellular siderophore biosynthesis following inactivation of the acyl transferase SidF or the nonribosomal peptide synthetase SidD leads to complete dependence upon reductive iron assimilation for growth under iron-limiting conditions , partial sensitivity to oxidative stress , and significantly reduced virulence , despite normal germ tube formation . Our findings reveal distinct cellular and disease-related roles for intra- and extracellular siderophores during mammalian Aspergillus infection .
Animals strategically withhold iron during infection to combat invading microbes [1 , 2] . Consequently , the ability to obtain iron from the host , both for essential metabolism and to cope with reactive oxygen species produced by phagocytic cells [3] , is a feature of most pathogens . The intimate coupling of iron uptake and storage with resistance to oxidative stress requires that all organisms strike a fine balance between the two; for example , catalases and peroxidases need heme as a cofactor [4] , but iron overload or incorrect storage can result in , or exacerbate , oxidative stress via Haber–Weiss/Fenton chemistry [5] . Aspergillus fumigatus is a saprophytic mould that has become the most common airborne fungal pathogen to cause disease in humans . Global ubiquity , and the infectious cycle of this pathogen , is perpetuated by prolific production of asexual spores termed conidia . Conidial germination in the human lung , following spore inhalation , represents the initiating event of pulmonary disease . Three important steps can be distinguished during spore germination: activation of the resting spore to appropriate environmental conditions , isotropic growth that involves water uptake and wall growth ( termed swelling ) , and polarized growth that results in the formation of a germ tube from which the new mycelium originates [6 , 7] . A . fumigatus causes a spectrum of diseases , depending upon the status of the host . Individuals with pre-existing structural lung disease , atopy , occupational exposure , or impaired immunity are susceptible to infection [8] . Invasive aspergillosis is now the most common cause of death due to fungal infection , occurring in up to one-quarter of transplant recipients or patients undergoing therapy for haematological malignancies , and 3% of AIDS patients . Typically , mortality associated with this disease reaches 50%–100% , due to difficulties with diagnosis and treatment [9] . Allergic bronchopulmonary aspergillosis is an A . fumigatus–induced respiratory disease usually found in atopic individuals [10] that can be life threatening , and frequently occurs in patients suffering from bronchial asthma , bronchiectasis , or cystic fibrosis [11] . A . fumigatus lacks specific uptake systems for host iron sources such as heme , ferritin , or transferrin . However , it employs two high-affinity iron uptake systems , reductive iron assimilation ( RIA ) and siderophore-assisted iron uptake , both of which are induced upon iron starvation [12] . RIA involves reduction of ferric to ferrous iron and subsequent uptake of ferrous iron by the FtrA/FetC complex , an activity that is blockable with the ferrous iron-specific chelator bathophenanthroline disulfonate ( BPS ) [12] . Siderophores are low molecular mass , ferric iron–specific chelators [13 , 14] . Like its mildly pathogenic relative Aspergillus nidulans , A . fumigatus produces three hydroxamate-type siderophores: extracellular fusarinine C ( FSC ) and triacetylfusarinine C ( TAFC ) , and intracellular ferricrocin ( FC ) [12 , 15] . FSC is an N2-acetyl-lacking precursor of TAFC , a cyclic tripeptide consisting of three N2-acetyl-N5-cis-anhydromevalonyl-N5-hydroxyornithine residues linked by ester bonds . FC is a cyclic hexapeptide with the structure Gly-Ser-Gly- ( N5-acetyl-N5-hydroxyornithine ) 3 [13] . Once secreted , FSC and TAFC mobilize extracellular iron for subsequent uptake [16] , and in A . nidulans , FC is involved in intracellular conidial and hyphal iron storage [17] . The postulated biosynthetic pathway according to Plattner and Diekmann [18] is shown in Figure 1 . The initial biosynthetic step , shared by pathways of both intra- and extracellular siderophore biosynthesis , is catalyzed by the ornithine-N5-monooxygenase SidA , the mutational inactivation of which abolishes siderophore biosynthesis , completely attenuating virulence in neutropenic mice and demonstrating the importance of the siderophore system in establishing infection [12] . Apart from A . nidulans SidC , the nonribosomal peptide synthetase essential for FC synthesis , other components of Aspergillus siderophore biosynthetic pathways have remained uncharacterized at the molecular level . In this study we identified the A . fumigatus FC synthetase SidC , and the TAFC biosynthetic enzymes SidD , SidF , and SidG . Moreover , we describe a novel A . fumigatus siderophore , hydroxyferricrocin ( HFC ) , employed for conidial iron storage . Phenotypic characterization of respective mutant strains reveals distinct roles for extracellular and intracellular siderophores in iron metabolism , resistance to oxidative stress , and virulence .
Most ascomycetes and basidiomycetes produce siderophores , but well-known exceptions are Saccharomyces spp . , Candida spp . , and Cryptococcus spp . [19] . As a first step in our analyses , we screened the recently released A . fumigatus genomic sequence [20] for genes putatively involved in siderophore biosynthesis . We expected such genes to be conserved among siderophore-producing organisms , including fungi and bacteria , and to be upregulated by iron starvation . Northern analysis identified four candidate genes in A . fumigatus , termed sidC ( Afu1g17200 ) , sidD ( Afu3g03420 ) , sidF ( Afu3g03400 ) , and sidG ( Afu3g03650 ) ( Figure 2 ) . SidC and SidD are nonribosomal peptide synthetases involved in biosynthesis of FC and TAFC , respectively . SidC displays high identity ( 55% ) to A . nidulans SidC [21] , and is conserved in most siderophore-producing fungi . SidD is conserved among siderophore-producing ascomycetes and is subject to iron-mediated transcriptional repression [22] . With respect to its two-module structure and amino acid sequence , SidD displays high similarity to NPS6 , which was recently shown to be involved in biosynthesis of TAFC- and coprogen-type siderophores in the plant pathogens Cochliobolus heterostrophus , Cochliobolus miyabeanus , Alternaria brassicicola , and Fusarium graminearum [23] . SidF homologs can be found among hydroxamate-producing fungi and numerous bacterial species; for example , the Escherichia coli homolog IucB , which displays 38% identity , is an N5-hydroxylysine:acetyl coenzyme A–N5-transacetylase that is essential for synthesis of the siderophore aerobactin [24] . SidG homologs , which belong to the N-acetyltransferase family ( GNAT , pfam00583 . 11 ) , are known to exist thus far in only two other fungal species , A . nidulans and Giberella zeae , but are present in several bacteria , such as Caulobacter crescentus , Desulfovibrio vulgaris , Streptomyces coelicolor , and Pseudomonas aeruginosa ( http://www . ncbi . nlm . nih . gov/BLAST/ ) . To analyze the respective functions of SidC , SidD , SidF , and SidG , we generated gene deletion mutants in A . fumigatus ATCC46645 ( wt ) by replacement with the hygromycin B resistance ( hph ) marker termed ΔsidC , ΔsidD , ΔsidF , and ΔsidG , respectively . To assess the impact of gene deletion on siderophore biosynthesis in each mutant , background siderophore production was quantified using high-performance liquid chromatography ( HPLC ) analysis of both culture supernatants and cell extracts following growth for 24 h under siderophore-derepresssing ( see Figure 2 ) iron-depleted conditions ( Figure 3 ) and compared to that of the wt and of the ornithine-N5-monooxygenase mutant , ΔsidA , which lacks both intra- and extracellular siderophores . The wt accumulated 7 mg FC per gram dry weight of mycelium , and excreted 42 mg TAFC per gram dry weight . The supernatant also contained the direct TAFC precursor FSC , approximating 12% of the mycelial TAFC content , and FC was detectable in trace amounts ( Figure 3A ) . SidC deletion abolished FC synthesis as measured from hyphal extracts but had no influence on TAFC and FSC production . Conversely , deletion of sidD or sidF prevented synthesis of TAFC and FSC without affecting FC production . SidG deletion eliminated TAFC production but increased FSC production about 10-fold , that is , the FSC reached in amount about the TAFC content of the wt . FC accumulation was not affected in ΔsidG . Taken together , the siderophore production pattern of the deletion mutants , the features of the gene products , and the predicted siderophore biosynthetic pathway [18] strongly suggest that ( i ) sidC encodes the nonribosomal peptide synthetase involved in ferricrocin biosynthesis , ( ii ) sidD encodes the nonribosomal peptide synthetase responsible for TAFC biosynthesis , ( iii ) sidF encodes N5-hydroxyornithine:cis anhydromevalonyl coenzyme A–N5-transacylase , and ( iv ) sidG encodes FSC–acetyl coenzyme A–N2-transacetylase ( Figure 1 ) . For each deletion mutant , three independent complemented strains were subjected to siderophore and phenotypic analysis . Complementation ( see Materials and Methods ) reversed all mutant phenotypes ( Figure S1 ) , definitively linking the phenotypes to inactivation of the respective gene . A . nidulans , Aspergillus ochraceus , and Neurospora crassa employ FC as both a hyphal and a conidial iron storage compound [21 , 25] . In stark contrast , HPLC analysis of A . fumigatus wt conidial extract revealed a lack of FC . Instead , we were able to detect the presence of a structurally distinct siderophore indicated by a different retention time ( Figure 3A ) . Its synthesis requires SidA and SidC but not SidD , SidF , or SidG , as indicated by conidial siderophore analysis of the respective deletion mutants ( Figure 3B ) . When the spores were generated on medium supplemented with FC , but not TAFC , ΔsidA and ΔsidC conidia contained this novel siderophore , suggesting that it is derived from FC . TAFC addition did not lead to HFC formation ( Figure 3B ) , negating the possibility that FC-bound iron induces HFC formation . Consistently , high-resolution mass spectrometry of this compound gave two molecular masses m/z ( M+H ) + 787 . 2432 matching C28H45N9O14Fe ( calculated molecular mass 787 . 2435 ) , and ( M+Na ) + 809 . 2253 matching C28H44N9O14FeNa ( calculated molecular mass 809 . 2255 ) ( unpublished data ) , suggesting that the conidial siderophore of A . fumigatus is derived from ferricrocin by hydroxylation; therefore , we termed it hydroxyferricrocin ( HFC ) . Importantly , these data also show that conidia of ΔsidA and ΔsidC strains , which lack all intracellular siderophores , can be loaded with HFC by supplementation with FC during sporulation . Notably , the total iron content of HFC-deficient conidia of both ΔsidA and ΔsidC strains was reduced about 67% compared to that of wt ( Table 1 ) . Conidial loading of either mutant with HFC by supplementation of the sporulation medium with the HFC precursor FC largely reconstituted the wt iron content ( Table 1 ) , demonstrating that HFC represents a major A . fumigatus conidial iron storage compound . All aspects of deletion-associated defects described above are completely rescued by gene-mediated complementation ( Figure S1 ) , directly linking observed phenotypic traits to deletion of the relevant siderophore biosynthetic pathway genes . As HFC was detected neither in the supernatant nor in hyphae of wt liquid cultures ( Figure 3 ) , we analyzed its potential developmental regulation by assessing siderophore content , conidia production , and expression of brlA , which encodes an early conidiation-specific transcription factor [26 , 27] , over a time course spanning conidiation . HFC synthesis started concomitantly with brlA expression ( Figure 4 ) and paralleled conidia production , indicating that its synthesis is developmentally regulated . We have previously shown that SidA deficiency causes absence of asexual sporulation during iron-depleted conditions partially curable by increased iron availability and completely restored by FC supplementation [12] . ΔsidC displayed a similar conidiation phenotype as ΔsidA ( Figure 5 ) , but ΔsidC produced more conidia than ΔsidA at the same iron concentration . Hence , FC or HFC appears to be specifically important for optimal sporulation . ΔsidD and ΔsidF showed decreased conidiation only during iron-depleted conditions , suggesting that this defect is caused by iron deficiency due to lack of siderophore-mediated iron uptake . ΔsidG displayed a wt conidiation rate . In view of sporulation defects in siderophore mutant backgrounds , generation of spores from all strains for which phenotypic and virulence testing is described was performed on iron-supplemented medium unless otherwise stated . The importance of FC for efficient germ tube formation from A . nidulans conidia , where it maintains the conidial siderophore , has been previously demonstrated [17] . In order to analyze the role of A . fumigatus siderophore production during growth initiation , we assessed the time course of conidial swelling and germ tube formation by A . fumigatus wt , ΔsidA , ΔsidC , ΔsidD , ΔsidF , and ΔsidG strains under iron-depleted conditions , iron-depleted conditions in the presence of BPS ( blocking RIA ) , and under iron-replete conditions having either iron sulphate or TAFC as iron source ( Tables 2 and 3 ) [7 , 28] . wt conidia were largely unaffected by the different conditions , as were those of ΔsidD , ΔsidF , and ΔsidG strains ( unpublished data ) , indicating that , in the presence of sufficient intracellular siderophore , extracellular iron mobilization is not a requirement for efficient growth initiation under any of the conditions tested . In contrast , conidial swelling and germ tube generation by ΔsidA conidia were not observable until 4 . 7 and 6 . 3 h , respectively , after the wt in iron-depleted conditions , demonstrating that siderophore-mediated iron storage or utilization of intracellular iron is required for efficient growth initiation under iron limitation ( Tables 2 and 3 ) . ΔsidA swelling and germ tubes were completely absent under iron depletion in the presence of BPS , demonstrating that RIA can promote initiation of growth in the absence of siderophores , albeit at a considerably slower rate . An increase of extracellular iron availability partly overcame the ΔsidA defects . TAFC was marginally more effective for this than ferrous sulphate , indicating that a lack of HFC can be compensated by uptake of extracellular iron and that siderophore-mediated iron uptake can more readily support efficient germ tube formation than RIA in the absence of HFC . Consistently , therefore , ΔsidC conidia displayed a delay in both swelling and germ tube production during iron-depleted conditions , but to a lesser extent than that of ΔsidA conidia , emphasizing that the ability to synthesize extracellular siderophores supports efficient germ tube development in the absence of HFC . As noted above , ΔsidA and ΔsidC conidia can be loaded with the HFC precursor FC , and such conidia approximated wt initiation of growth ( Table 3 ) , establishing that the described growth defects are indeed due to lack of HFC . Notably , FC supplementation ( as the sole means of acquiring iron ) permits a conidial HFC iron content sufficient to allow germ tube formation and elongation up to a length of 500 μm in vitro , as HFC-containing ΔsidA conidia stop growing at this stage in the presence of BPS ( unpublished data ) . To analyze the role of extracellular and intracellular siderophores in resistance to iron starvation and to oxidative stress , radial growth rates of ΔsidA , ΔsidC , ΔsidD , ΔsidF , and ΔsidG were compared to that of wt under various stress-inducing growth conditions ( Figure 6 ) . In all assays performed ( see below ) , ΔsidG behaved comparably to the wt , demonstrating that the TAFC precursor FSC , at least at the elevated levels observed in this mutant , can fully compensate an absence of TAFC . As shown previously [12] , a complete lack of siderophores ( ΔsidA ) reduces the growth rate significantly during iron-depleted conditions . In comparison , absence of either the intracellular ( ΔsidC ) or the extracellular ( ΔsidD , ΔsidF ) siderophores under iron limitation caused a mild growth reduction , suggesting some redundancy in function for intra- and extracellular siderophores under these conditions . Interestingly , ΔsidF was less affected than ΔsidD . During iron-depleted conditions , inhibition of RIA by BPS completely impaired the growth of ΔsidA , ΔsidD , and ΔsidF and reduced the growth rate of ΔsidC . Compared to ΔsidF , ΔsidA and ΔsidD required a higher extracellular iron concentration to compensate the defect . Taken together , these data suggest that , in the absence of siderophore-mediated iron mobilization , RIA is an absolute requirement for surviving iron limitation , being able to compensate partially for a lack of intracellular or extracellular siderophores in vitro . Furthermore , intracellular siderophores are also required for promoting growth in the absence of RIA , although to a lesser extent compared to the extracellular siderophores because ΔsidC is less affected by BPS in iron-limiting conditions . Detoxification of hydrogen ( peroxide ) depends on iron because catalases and peroxidases require heme as cofactor . Consistently , iron-depleted A . fumigatus is more sensitive to hydrogen peroxide than iron-replete cells ( unpublished data ) . On the other hand , inappropriate iron storage can catalyze formation of reactive oxygen species . Deficiency of total ( ΔsidA ) and intracellular ( ΔsidC ) siderophores caused hypersensitivity to hydrogen peroxide during iron-depleted growth ( Figure 6B ) . This defect was cured by an increase of extracellular iron availability , suggesting that the major role of intracellular siderophore is efficient iron utilization rather than iron detoxification . Deficiency in extracellular siderophores ( ΔsidD and ΔsidF ) rendered cells partially sensitive to hydrogen peroxide , and again , this defect was limited to iron-depleted conditions ( Figure 6B ) . To investigate the oxidative stress sensitivity of strains deficient in the extra- and/or intracellular siderophore in more detail , we analyzed conidial and hyphal killing by hydrogen peroxide in each mutant background . HFC-lacking conidia of both ΔsidA and ΔsidC showed increased sensitivity to hydrogen peroxide , and this defect was cured by HFC loading ( Figure 6C ) . These data indicate that deficiency in conidial siderophore iron storage causes increased susceptibility to killing by hydrogen peroxide . In a hyphal killing assay , ΔsidA was significantly more sensitive and ΔsidC slightly more sensitive than wt ( Figure 6D ) , demonstrating that FC also plays a role in hyphal oxidative stress resistance . In contrast to ΔsidA and ΔsidC , conidia of ΔsidD , ΔsidF , and ΔsidG were as resistant as wt to hydrogen peroxide ( unpublished data ) , suggesting that the increased sensitivity of ΔsidD and ΔsidF , found in the plate assays ( Figure 6B ) , is due to sensitivity of hyphae only—probably due to iron deficiency caused by lack of extracellular siderophores . To test the hypothesis that catalase deficiency underlies the observed oxidative stress sensitivity of ΔsidA and ΔsidC , we performed catalase staining on hyphal and conidial protein extracts . A . fumigatus produces three active catalases , Cat1 and Cat2 in hyphae , and CatA in conidia [29] . Mycelium of strains lacking both Cat1 and Cat2 exhibit only slightly increased sensitivity to hydrogen peroxide , whereas CatA deficiency results in significant increased sensitivity to hydrogen peroxide of conidia [29] . Catalase activity zymograms demonstrated that activity of hyphal Cat1 and Cat2 is decreased during iron-depleted compared to iron-replete conditions , but did not display any difference between wt , ΔsidA , ΔsidC , ΔsidD , ΔsidF , and ΔsidG strains ( unpublished data ) . These data suggest that the increased oxidative stress sensitivity of the siderophore mutant strains during iron-depleted conditions is not due to decreased hyphal catalase activity . CatA activity was about the same in wt , ΔsidD , ΔsidF , and ΔsidG strains but significantly decreased in ΔsidA and ΔsidC conidia ( Figure 7 ) , which agrees with the increased hydrogen peroxide sensitivity of conidia from these strains . Supplementation of the sporulation medium with FC reconstituted CatA activity of ΔsidA and ΔsidC conidia , demonstrating that the catalase defect is caused by lack of conidial siderophore . Taken together , all mutant impairments were most severe under iron-depleted conditions and at least partially reverted by an increase of extracellular iron availability , indicating that the observed defects are related to gene deletion–induced iron deficiency . Alternatively , or additionally , the defects might be related to accumulation of toxic intermediates , produced only during iron-depleted conditions . In this respect , the differing behaviour of ΔsidD and ΔsidF might indicate that abolition of TAFC synthesis at different steps of the biosynthetic pathway has varying consequences dependent upon different pathway intermediates interfering with metabolism . To assess the relative contributions of intra- and extracellular siderophores to virulence of A . fumigatus , we compared the survival of neutropenic mice following infection with 5 × 105 A . fumigatus ΔsidA , ΔsidC , ΔsidD , ΔsidF , or ΔsidG conidiospores to that of mice ( n = 13 ) infected with an equivalent dose of the corresponding complemented strain ( Figure 8 ) . Following intranasal inoculation with a saline conidial suspension , mice were monitored for signs of respiratory distress and weighed daily . A cumulative weight loss of 20% body weight relative to that measured on the day of infection was taken as a stand-alone endpoint of experimentation . The median survival time of mice ( n = 13 ) infected with the parental isolate ATCC46645 was 6 d , and 100% mortality was recorded for this group since no mouse infected with this wt isolate survived beyond the 11th day post-infection ( Figure S2 ) . Histopathological analysis of wt-infected lung tissue sections revealed numerous germinated spores and branching primary hyphae appearing as discreet pulmonary lesions in the bronchioles and alveoli at 24 h post-infection ( Figure 9 ) and having an even distribution throughout the sections examined . Weight loss was steady from day +1 of the infection ( unpublished data ) , and recruitment of inflammatory cells to foci of wt A . fumigatus infection was evident at 24 h , and substantial at 72 h , post-infection . This pathology stands in direct contrast to that of ΔsidA-infected mice , where virulence is completely attenuated [12] and neither germinated spores nor hyphal elements are observable at this infectious dose at similar time points of infection ( Figure 9 ) . In comparison to survival following infection with a minimum of three independently gene-complemented strains ( n = 15 , Figure S3 ) , survival of ΔsidC-infected mice ( n = 23 ) was significantly increased ( p = 0 . 0017 , by log rank survival analysis ) , leading to 41% mortality and median survival time of 8 d among mice succumbing to infection ( Figure 8 ) . ΔsidC infection was characterized by a marked reduction in germinated spores at 24 h post-infection and a concomitant absence of inflammation relative to wt infection ( Figure 9 ) . Many phagocytosed conidia were evident at this time point ( unpublished data ) , as well as extremely scarce incidences of germination and tissue invasion ( Figure 9 ) . At 72 h post-infection , hyphal moieties were detectable , but very scarce compared to the wt , and discreet inflammatory lesions were associated with such foci as evidenced by hematoxylin and eosin staining ( Figure 9 ) . Comparative infection with ΔsidF ( n = 14 ) and corresponding complemented strains ( n = 14 , Figure S3 ) resulted in similarly attenuated virulence ( p = 0 . 0006 , by log rank survival analysis ) resulting in 36% mortality and median survival of 6 d among mice succumbing to infection ( Figure 8 ) . Germinated spores were detectable by histopathological examination at the 24-h time point of infection; however , they were difficult to locate due to their sparse distribution within the sections examined . Mild inflammation was evident at this time point , as well as many instances of phagocytosed conidia ( unpublished data ) . At 72 h post-infection , germinated spores and mycelial growths were evident with accompanying inflammatory cell recruitment ( Figure 9 ) . Once more , the frequency of such lesions is markedly reduced in comparison with the wt infection . ΔsidD infection ( n = 13 ) , however , was not fatal in neutropenic mice . With a single exception , mice survived the infectious challenge ( Figure 8 ) and no significant weight loss was recorded despite renewed immunosuppression throughout the course of experimentation ( unpublished data ) . ΔsidD is therefore severely ( if not completely ) attenuated for virulence in neutropenic mice ( p < 0 . 0001 compared to n = 13 complemented strains ) . Viable conidia were not recovered from the lung of the mortally afflicted subject , though a degree of bacterial colonization of the lung was evident upon plating lung homogenates , and a severe drop in body weight commencing on day +4 was observed . We therefore reserve judgement on the cause of this particular fatality . Histopathology was comparable to that of ΔsidA infection , with a similar lack of germinating spores at any time point examined and no evidence of inflammatory response to the inoculum at the time points tested ( Figure 8 ) . Conversely , infection with ΔsidG was as virulent as the wt and demonstrated a similar pathology of infection ( Figures 8 and 9 ) . We previously demonstrated an absolute requirement for A . fumigatus siderophore biosynthesis for pathogenicity in neutropenic mammalian infections [12] . The analyses described here identify a clear requirement for both intra- and extracellular siderophores for full pathogenicity since abrogation of either category leads to attenuation of virulence . ΔsidA attenuation therefore likely results from the compound phenotype associated with total siderophore abrogation , including delayed germination , increased conidial and hyphal sensitivity to oxidative stress , and sub-optimal growth under iron-limiting conditions . To determine the contribution of germinative capacity and oxidative stress resistance to overall virulence phenotype , we exploited the capacity to load conidia artificially with HFC ( through FC supplementation of the growth medium ) and examined ΔsidA and ΔsidC virulence in the presence , and absence , of HFC . We reasoned that if the lack of siderophore-assisted germination and oxidative stress resistance was the sole basis of ΔsidA attenuation , then HFC-mediated rescue of ΔsidA germination in vivo would also completely rescue virulence . FC supplementation of ΔsidA growth for 5 d prior to infection partially restored pathogenicity , resulting in a 50% recovery of virulence ( Figure 10 ) . However , no increase in ΔsidC virulence was observed under similar experimental conditions ( unpublished data ) . Complementation ( see Materials and Methods ) reversed all mutant virulence phenotypes ( Figures 8 and S3 ) , definitively linking attenuated phenotypes to inactivation of the respective gene .
The microbial quest for iron in mammalian hosts is crucial for successful pathogenesis as , in this environment , iron is tightly bound by carrier proteins such as transferrin , leaving free iron concentrations insufficient for sustenance of microbial growth . Most aerobic bacteria and fungi have genes encoding iron transport systems that become induced under iron limitation [30 , 31] , among which siderophore-mediated iron transport provides a means of uptake , even for organisms which cannot themselves synthesize such molecules . Various bacteria produce extracellular siderophores , and in many cases their involvement in virulence has been demonstrated [30] . However , bacteria do not produce intracellular siderophores , but use ferritin and bacterioferritin for this purpose [32] . We have previously demonstrated [12] that in the aggressive , but poorly characterized , fungal pathogen A . fumigatus , a single genetic locus , sidA , directs biosynthesis of the extracellular siderophores FSC and TAFC ( for mobilization of extracellular iron ) , and a hyphal siderophore FC ( for hyphal iron storage ) . Coupled with a second high-affinity iron uptake mechanism , RIA , these low molecular mass ferric iron–specific chelators ensure a steady supply , and appropriate storage , of cellular iron . Abrogation of A . fumigatus siderophore biosynthesis by sidA deletion prevents initiation of mammalian infection , which cannot be supported by RIA alone . Consistently , inactivation of RIA by deletion of the high-affinity iron permease–encoding ftrA is inconsequential for virulence , identifying siderophore biosynthesis by this organism as paramount to successful pathogenesis . The absence of such biosynthetic pathways in mammals lends much promise to siderophore biosynthesis as the basis for therapy . To this end we have genetically delineated pathways of siderophore biosynthesis in A . fumigatus . Previously , the ornithine monooxygenase–encoding sidA was the only known gene , and its deletion , resulting in absence of all siderophore types , caused a compound phenotype comprising hypersensitivity to hydrogen peroxide , increased iron demand for germination , and reduced growth rate , in particular during iron-depleted conditions in vitro ( this study and [12] ) . Here , we describe the identification and comparative mutational analysis of four novel A . fumigatus iron-regulated genes , sidC , sidF , sidD , and sidG , whose expression is repressed by iron , allowing elucidation of the A . fumigatus siderophore biosynthetic pathways downstream of sidA ( Figure 1 ) and facilitating analysis of the relative contributions of extra- and intracellular siderophores in germ tube formation , sporulation , tolerance to iron depletion , oxidative stress resistance , and virulence . Initial comparisons of siderophore production in mutant and wt backgrounds were highly informative with respect to ordering genetic loci within the expected biosynthetic pathways . The nonribosomal peptide synthetase SidC is required for biosynthesis of FC and HFC , whereas the acetyl transferase SidG , the acyltranferase SidF , and the nonribosomal peptide synthetase SidD are essential for biosynthesis of TAFC , suggesting the biosynthetic pathway shown in Figure 1 . Analyses of respective siderophore biosynthetic mutants in comparison to the wt and ΔsidA isolates revealed a range of phenotypes attributable to siderophore deficiency . Moreover , these analyses identified a novel conidial siderophore in A . fumigatus , derived from FC by hydroxylation , which we termed hydroxyferricrocin . We found that similar to FC in A . nidulans [17] , HFC is required for conidial iron storage in A . fumigatus ( Table 1 ) and that its lack causes increased sensitivity to oxidative stress ( Figure 6C ) , likely due , at least in part , to CatA deficiency ( Figure 7 ) , and delayed swelling and germ tube formation of conidia during iron depletion ( Tables 2 and 3 ) . Remarkably , therefore , the iron homeostatic machinery of A . fumigatus differs from that in the closely related , but negligibly virulent , model ascomycete A . nidulans in at least two aspects . A . fumigatus uses HFC as the conidial iron storage compound instead of FC and is able to assimilate iron reductively ( RIA ) , which makes it more versatile with respect to iron acquisition [31] . The partial rescue of ΔsidA attenuation following reconstitution of the conidial HFC , which is possible by supplementation of the sporulation medium with the HFC precursor FC , in the absence of de novo synthesis of both extracellular and intracellular siderophores demonstrates the importance of the conidial siderophore during the initial phase of infection . Prevention of HFC biosynthesis , as evidenced by measurement of conidial iron content in ΔsidA and ΔsidC backgrounds ( Table 1 ) , reduced conidial iron content by 67% compared to the wt isolate , and negated a role for extracellular siderophores in conidial iron storage since no difference in conidial iron content was discernable between ΔsidA , which lacks extracellular siderophores , and ΔsidC , which has a full complement of extracellular siderophores . In fact , such comparative analyses of ΔsidA and ΔsidC provide a useful basis for determining the contribution of extracellular siderophores to several physiological processes . For example , while ΔsidA germination is significantly delayed or completely absent during iron deficiency whether in the presence or absence of RIA , ΔsidC displays a far milder phenotype under both conditions , highlighting the importance of extracellular iron mobilization as a compensator for deficiency in intracellular iron storage and identifying an interdependency between different types of siderophore for certain cellular processes , such as germination . Other phenotypic manifestations of siderophore-mediated iron storage deficiency include growth retardation under iron limitation ( regardless of RIA activity ) and extreme sensitivity to oxidative stress , as measured by radial growth on solid agar , which was equally potent in both ΔsidA and ΔsidC backgrounds . The fact that sidA and sidC are equally required for full conidial resistance to oxidative stress but ΔsidA is avirulent and ΔsidC only partially attenuated demonstrates that this common feature cannot be the sole virulence determinant of these mutants , and indicates a crucial role also of extracellular siderophores , which are still produced by ΔsidC . Closer examination of the oxidative stress sensitivities revealed an important distinction between the two strains where , despite identical conidial sensitivities to hydrogen peroxide ( Figure 6C ) , hyphal sensitivity to hydrogen peroxide was significantly greater for ΔsidA ( Figure 6D ) , implicating extracellular siderophore production in hyphal tolerance to oxidative stress . Both mutants suffer catalase A deficiency ( Figure 7 ) , which correlates well with conidial sensitivity to hydrogen peroxide ( Figure 6C ) . Distinguishable severity of phenotype in these two mutant backgrounds extends also to pathogenicity , where the partial attenuation of virulence observed following ΔsidC infection ( Figure 8 ) presumably reflects the delayed germination , growth retardation , and oxidative stress phenotypes observed for this mutant in vitro . All observed defects were more pronounced under iron depletion , suggesting that the intracellular siderophore is required for optimal iron storage and possibly iron distribution rather than iron detoxification . Consequently , the increased sensitivities to oxidative stress in ΔsidA and ΔsidC backgrounds might be due to hampered detoxification that requires iron as cofactor; for example , the heme-containing catalases and peroxidases [4] . In light of the fact that catalase A deficiency alone has no role in virulence [29] , resistance to oxidative stress in vivo , regardless of origin , must be supported by other as yet unidentified functions . An alternative explanation for sidA- and sidC-supported pathogenicity might be that disruption of iron homeostasis due to lack of the intracellular siderophore increases intrinsic oxidative stress . As found previously for A . nidulans [21] , deficiency in the intracellular siderophore caused reduced production of spores , indicating a crucial role in conidiogenesis ( Figure 5 ) . Considering similarly the effects of extracellular siderophore abrogation , certain physiological consequences are clearly unique to elimination of these compounds in A . fumigatus . In contrast to the iron deficiency–mediated growth retardation observed in the ΔsidC background , ΔsidD and ΔsidF mutants are completely incapable of growth in the absence of RIA when challenged with iron shortage . This highlights a dramatic shortfall in terms of compensatory mechanisms for extracellular siderophore deficiency , which is likely to extend to infection scenarios where the severe lack of available free iron would render RIA suboptimally effective . Partial sensitivity to oxidative stress is measurable by radial growth retardation for both ΔsidD and ΔsidF ( Figure 6B ) and both strains demonstrate attenuated virulence , although this phenotype is far more pronounced ( or possibly absolute ) for ΔsidD . By our analyses , the only basis for this difference lies in the amount of free iron supplementation required for each strain to rescue the effect of RIA inactivation , which is much higher for ΔsidD than ΔsidF . One might therefore hypothesize that RIA is able to support ΔsidF virulence to a greater extent than ΔsidD virulence . Alternatively , or additionally , the defects might be related to accumulation of intermediates of the blocked pathway , which is induced only during iron-depleted conditions , and interfere with metabolism . Notably , reversed-phase HPLC analysis combined with mass spectrometry indicated that the supernatant of iron-depleted ΔsidD , but not ΔsidF , contains elevated amounts of two compounds compared with wt . One of these could be identified as N5-cis-anhydromevalonyl-N5-hydroxyornithine , the direct precursor of FSC ( unpublished data ) . In this respect , the increased sensitivity to iron depletion and oxidative stress of ΔsidD , compared to ΔsidF , suggests that blockage of TAFC synthesis at different steps of the biosynthetic pathway has different consequences . In accord with the in vitro phenotypes , deletion of sidD had a greater impact on virulence as compared to sidF . All ΔsidD and ΔsidF defects were compensated by an increase in extracellular iron availability , suggesting that these impairments are related to gene deletion–induced iron deficiency ( Figure 6B ) . The attenuated virulence of ΔsidD and ΔsidF ( Figure 8 ) clearly indicates induction of this pathway during infection . Notably , Cramer et al . [33] showed that sidD is the most highly expressed A . fumigatus NRPS-encoding gene following incubation with macrophages . Furthermore , we found significant induction of sidC , sidD , sidF , and sidG at the level of gene expression at an early stage of infection in neutropenic mice in genome-wide expression analyses ( A . McDonagh , personal communication ) . Iron deficiency is unlikely to result from immunosuppressive regimen alone since many other microbes suffer iron stress–induced attenuation in immunocompetent murine models of infection , with the pulmonary pathogen Mycobacterium tuberculosis providing a good example [34] . During iron depletion , A . fumigatus usually excretes high amounts of TAFC and low amounts of the ultimate TAFC precursor FSC . Strains producing FSC in amounts comparable to those of TAFC produced by wt , but lacking TAFC due to deficiency in SidG , behave like wt under all conditions investigated , including virulence , demonstrating that FSC can satisfactorily replace TAFC as a siderophore in vitro and in vivo . This study demonstrates the distinct roles of intra- and extracellular siderophores in iron homeostasis of A . fumigatus and reveals that the complete complement of intra- and extracellular siderophores is required for full virulence of this species . The predominant role for intracellular siderophores appears to lie with promoting germination and resisting oxidative stress , both of which require extracellular siderophores , whose absence or malfunction can be supported by RIA . The predominant role of extracellular siderophores , however , is to facilitate hyphal growth under iron limitation , particularly in circumstances where RIA is ineffective . Recently , RIA was found to be dispensable for virulence of the plant pathogen Fusarium graminearum [35] , whereas deficiency in the A . fumigatus SidD ortholog NPS6 caused loss of extracellular siderophores and reduction of virulence of F . graminearum , Cochliobolus heterostrophus , C . miyabeanus , and Alternaria brassicicola [23] , demonstrating that siderophores are a common virulence determinant of at least some animal and plant pathogenic fungal species . Moreover , the intracellular siderophore has also been implicated in the virulence of Mangaporthe grisae in rice [36] . Nevertheless , the role of individual iron homeostasis-maintaining mechanisms in virulence largely depends on the pathogen–host system because the siderophores produced by the phytopathogenic basidiomycetes Ustilago maydis and Microbotryum violaceum do not contribute to their virulence [37 , 38] , and because there are siderophore-lacking animal pathogenic ascomycetes , for example , C . albicans , and basidiomycetes , for example , C . neoformans . In C . albicans and U . maydis , RIA was found to be crucial for virulence [39 , 40] . Our analysis depicts complementary , but differential roles for distinct A . fumigatus siderophores , which appear to be employed for different purposes in vitro , and during infection , across a developmental spectrum . Under most circumstances , the combined abolishment of intra- and extracellular siderophore biosynthesis is required for extreme debilitation . An exception to this rule is posed by sidD inactivation , which matches ΔsidA in terms of virulence attenuation ( Figure 8 ) . Sharing absolute growth inhibition in the absence of RIA under iron depletion , the functions lacked by these mutants would seem to be equally attractive from a therapeutic standpoint .
Fungal strains ( Table 4 ) were cultured at 37 °C in +Fe-Aspergillus minimal medium ( AMM , iron-replete conditions ) according to Pontecorvo et al . [41] containing 1% ( wt/vol ) glucose as carbon source , 20 mM glutamine as nitrogen source , and 10 μM FeSO4 . For iron-depleted conditions , iron was omitted . For growth assays , the respective strains were point inoculated in a concentration of 104 colony forming units ( cfu ) /ml on appropriately supplemented AMM plates and incubated for 48 h at 37 °C . In view of sporulation defects in siderophore mutant backgrounds , generation of spores from all strains for which phenotypic and virulence testing is described was performed on medium supplemented with 1 . 5 mM FeSO4 , unless otherwise stated . For supplementation with FC or TAFC , the siderophores were used in the holo form ( iron containing ) . Oligonucleotides are listed in Table S1 . Total RNA was isolated using TRI reagent ( Sigma-Aldrich , http://www . sigmaaldrich . com/ ) and northern analysis was performed according to Sambrook et al . [42] . Hybridization probes were generated by PCR using primer oAf538PS1-f and oAf538PS1-r for sidD , oAf538AT1-f and oAf538AT1-r for sidF , oAfAT1me and oAfAT2me for sidG , oAf267PS1-f and oAf267PS1-r for sidC , and oTubfum1 and oTUBFUMr for β-tubulin encoding tubA . For extraction of genomic DNA , mycelia were homogenized , and DNA isolated according to Sambrook et al . [42] . E . coli DH5α strain was used as the host for general DNA propagations . For inactivation of sidC , sidD , sidF , and sidG , respective fragments including flanking regions were amplified from genomic DNA by PCR and subcloned into plasmid vectors . Genes were replaced by the hygromycin resistance marker gene hph [43] using standard techniques . Details are provided in Text S1 . For transformation , deletion constructs were released from bacterial vectors and used for transformation of wt protoplasts . For the generation of ΔsidF and ΔsidG mutant strains , the bipartite marker technique was used [44] . Accurate gene deletion was confirmed by Southern hybridization . Mutant strains were complemented by co-transformation of respective plasmids ( sidF and sidG ) or cosmids ( sidC and sidD ) and plasmid pSK275 carrying the pyrithiamine resistance gene as a selection marker [45] . Direct selection of reconstitution alleles having single homologous insertions proved impossible despite several attempts , so we chose instead to complement each mutant strain by ectopic integration and test three independent complemented strains for each genetic defect generated . Single-copy ectopic integration in complemented mutant strains was verified by PCR and Southern hybridization . Characterization and quantification of siderophores was performed by reversed-phase HPLC chromatography according to Konetschny-Rapp et al . [46] , as described by Oberegger et al . [15] . To analyze the conidial siderophore HFC , freeze-dried conidia were solubilized in a Mixer Mill 300 ( Retsch , http://www . retsch . com/ ) and resuspended in 50 mM potassium phosphate buffer . After purification via an Amberlite XAD-16 resin ( CWG , http://www . cwg . hu/ ) column , aliquots were analyzed by reversed-phase HPLC . The sensitivity of conidia to killing by hydrogen peroxide was assayed as described by Han et al . [47] . Briefly , conidial suspensions approximating 105 cfu/ml were incubated for 30 min at 20 °C with varying concentrations of hydrogen peroxide . To determine the number of surviving conidia , the spore suspensions were diluted 50-fold and plated on +Fe-AMM . Following incubation for 24 h at 37 °C , colonies were counted and normalized to that without hydrogen peroxide treatment . The sensitivity of hyphae to hydrogen peroxide was estimated by using a modification of the protocol of Kawasaki et al . [48] . Approximately 150 conidia were plated on iron-replete minimal medium and grown at 37 °C for 19 h—at this time point small colonies were countable . Subsequently , the plates were overlaid with 4 . 5 ml of the same medium as top-agar but containing the indicated concentration of hydrogen peroxide . After further incubation for 24 h at 37 °C , colonies able to resume growth were counted as survivors and normalized to the number before hydrogen peroxide treatment . Murine infections were performed under UK Home Office Project Licence PPL/70/5361 in dedicated facilities at Imperial College London . Outbred male mice ( strain CD1 , 18–22 g; Harlan Ortech , http://www . harlan . com/ ) were housed in individually vented cages . Mice were immunosuppressed as previously described [12] . A . fumigatus spores for inoculations were grown on Aspergillus complete medium , containing 5 mM ammonium ( + ) -tartrate , 200 mM NaH2PO4 , and 1 . 5 mM FeSO4 for 5 d prior to infection . Conidia were freshly harvested using sterile saline ( Baxter Healthcare , http://www . baxterhealthcare . co . uk/ ) and filtered through Miracloth ( Calbiochem , http://www . emdbiosciences . com/g . asp ? f=CBC/home . html ) . Conidial suspensions were spun for 5 min at 3 , 000g , washed twice with sterile saline , counted using a hemocytometer , and resuspended at a concentration of 6 . 25 × 106–1 . 25 × 107 cfu/ml . Viable counts from administered inocula were determined following serial dilution by plating on Aspergillus complete medium containing 5 mM ammonium ( + ) -tartrate , 200 mM NaH2PO4 , and 1 . 5 mM FeSO4 and growth at 37 °C . Mice were anesthetized by halothane inhalation and infected by intranasal instillation of 2 . 5 × 105–5 × 105 conidia in 40 μl of saline . Mice were weighed every 24 h from day 3 and visual inspections made twice daily . In the majority of cases the end point for survival experimentation was a 20% reduction in body weight measured from the day of infection , at which point mice were sacrificed . Lungs for histological sectioning were removed immediately after sacrifice and fixed in 4% v/v formaldehyde ( Sigma ) . Lungs were embedded in paraffin prior to sectioning and stained with hemotoxylin and eosin or light green and Grocott's Methenamine Silver .
The A . fumigatus GenBank/NCBI ( http://www . ncbi . nlm . nih . gov/ ) nucleotide and amino acid sequences , respectively , for the genetic loci described in this publication are sidC ( XM_747995 and XP_753088 . 1 ) , sidD ( XM_743569 and XP_748662 . 1 ) , sidF ( XM_743567 and XP_748660 . 1 ) , and sidG ( XM_743592 and XP_748685 . 1 ) .
|
Patients with suppressed immune systems due to cancer treatments , HIV/AIDS , organ transplantation , or genetic disorders are at high risk of infection with the ubiquitously present fungal pathogen Aspergillus fumigatus . Treatments for this disease , collectively termed invasive aspergillosis , are often not successful , and prospects for survival can be slim . A . fumigatus produces small molecules , termed siderophores , for acquisition and storage of iron , an element essential for growth . We found that these siderophores are crucial for virulence of A . fumigatus because their removal ( by gene deletion ) prevents or lessens disease in a mouse model of invasive aspergillosis . Siderophores are not produced by humans so they present good prospects for new therapies , as drugs that specifically target siderophore production , rather than activities shared by humans and fungi , are less likely to affect patients adversely .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"biochemistry",
"infectious",
"diseases",
"yeast",
"and",
"fungi",
"eukaryotes",
"microbiology",
"mus",
"(mouse)",
"genetics",
"and",
"genomics"
] |
2007
|
Distinct Roles for Intra- and Extracellular Siderophores during Aspergillus fumigatus Infection
|
Human malaria parasite species were originally acquired from other primate hosts and subsequently became endemic , then spread throughout large parts of the world . A major zoonosis is now occurring with Plasmodium knowlesi from macaques in Southeast Asia , with a recent acceleration in numbers of reported cases particularly in Malaysia . To investigate the parasite population genetics , we developed sensitive and species-specific microsatellite genotyping protocols and applied these to analysis of samples from 10 sites covering a range of >1 , 600 km within which most cases have occurred . Genotypic analyses of 599 P . knowlesi infections ( 552 in humans and 47 in wild macaques ) at 10 highly polymorphic loci provide radical new insights on the emergence . Parasites from sympatric long-tailed macaques ( Macaca fascicularis ) and pig-tailed macaques ( M . nemestrina ) were very highly differentiated ( FST = 0 . 22 , and K-means clustering confirmed two host-associated subpopulations ) . Approximately two thirds of human P . knowlesi infections were of the long-tailed macaque type ( Cluster 1 ) , and one third were of the pig-tailed-macaque type ( Cluster 2 ) , with relative proportions varying across the different sites . Among the samples from humans , there was significant indication of genetic isolation by geographical distance overall and within Cluster 1 alone . Across the different sites , the level of multi-locus linkage disequilibrium correlated with the degree of local admixture of the two different clusters . The widespread occurrence of both types of P . knowlesi in humans enhances the potential for parasite adaptation in this zoonotic system .
The epidemiological emergence of infections can be traced by genotypic analyses , with a high level of resolution when pathogens have a high mutation rate , as illustrated by recently emerged viruses that now have a massive impact on global public health [1 , 2] . Such analysis is more challenging for eukaryote pathogens with low mutation rate , although it is now clear that the major human malaria parasites Plasmodium falciparum and P . vivax have been endemic for many thousands of years after having been acquired as zoonotic infections from African apes [3 , 4] . In contrast , natural human infections by P . knowlesi were almost unknown [5] until a large focus of cases in Malaysian Borneo was described a decade ago [6] . Infections have since been reported from throughout southeast Asia , within the geographical range of the long-tailed and pig-tailed macaque reservoir hosts ( Macaca fascicularis and M . nemestrina ) and mosquito vectors ( of the Anopheles leucosphyrus group ) [7] . The most highly affected country is Malaysia , where there have been thousands of reported cases and P . knowlesi is now the leading cause of malaria in most areas [8 , 9] . It is vital to determine the causes of this apparent emergence , as P . knowlesi can cause severe clinical malaria with a potentially fatal outcome [10–12] . Increasing rates of case detection may reflect better diagnosis , increased transmission by mosquitoes from reservoir host macaques to humans , or parasite adaptation to humans . Molecular tools to discriminate P . knowlesi from other malaria parasite species were not widely applied until the zoonosis became known , but analysis of DNA in archived blood samples from Malaysia and Thailand shows that it was already widespread twenty years ago [13 , 14] . Sequences of parasite mitochondrial genomes and a few nuclear gene loci indicate ongoing zoonotic infection , as human P . knowlesi genotypes share most alleles identified in parasites sampled from wild macaques [15–17] . To understand this zoonosis , and to identify whether human-to-human mosquito transmission is occurring , analyses of parasite population genetic structure in humans and macaques should be performed by extensive population sampling and characterisation of multiple putatively neutral loci . This study presents a P . knowlesi microsatellite genotyping toolkit and its application to the analysis of a large sample of isolates from human cases at ten different sites , as well as from both species of wild macaque reservoir hosts . Results reveal a profound host-associated sympatric subdivision within this parasite species , as well as geographical differentiation indicating genetic isolation by distance . The existence of two divergent parasite subpopulations , and their admixture in human infections provides unparalleled opportunity for parasite hybridisation and adaptation . Observations of some clinical infections with parasite types that appear intermediate between the two subpopulations may reflect this process , and are a possible result of human-to-human mosquito transmission .
Hemi-nested PCR assays were developed for amplification of 19 tri-nucleotide simple sequence repeat loci from throughout the genome of P . knowlesi and tested for species-specificity using control DNA from all 10 known parasite species of humans , long-tailed or pig-tailed macaques , as well as human and macaque DNA to identify those suitable for genotyping samples from all hosts ( S1 Table ) . Assays for 11 loci were entirely species-specific for P . knowlesi , and 10 of these gave a clear single electrophoretic peak for each allele without any stutter bands ( S2 Table ) . These were used to genotype P . knowlesi infections in a total of 599 humans and wild macaques with a high rate of success , 556 ( 92 . 8% ) scoring clearly for all 10 loci ( S3 Table and S1 Dataset ) . Numbers of alleles at each locus ranged from 7 ( for locus NC03_2 ) to 21 ( for locus CD05_06 ) ( S1 Dataset ) . We first compared parasites from different host species sampled from Kapit where high numbers of clinical cases are seen ( Fig 1 ) , analysing the infections with complete 10-locus genotype data . Almost all P . knowlesi infections in macaques contained multiple genotypes , with no significant difference between long-tailed macaques ( 88% of 34 were mixed , a mean of 2 . 71 genotypes per infection ) and pig-tailed macaques ( 100% of 10 were mixed , a mean of 2 . 70 genotypes per infection; P = 0 . 65 for comparison between macaque host species ) , whereas only a minority of human P . knowlesi infections had multiple genotypes ( 35% of 167 , a mean of 1 . 40 genotypes per infection; P < 10–15 for comparison between humans and macaques ) ( Fig 2A ) . To allow equally weighted sampling per host , the predominant allele at each locus within each infection was counted for subsequent analysis ( S1 Dataset ) . Pairwise comparisons of each of the complete 10-locus profiles revealed that all infections in Kapit were genotypically distinct , except for one identical pair and one identical triplet of human infections ( Fig 2B , S4 Table ) . There was a much higher average proportion of shared alleles among pig-tailed macaque infections than among those in long-tailed macaques or humans ( medians of 5 , 3 and 2 identical alleles out of 10 loci respectively ) . Analysis of allele frequencies revealed that P . knowlesi parasites from pig-tailed macaques are very highly divergent from those in long-tailed macaques ( FST = 0 . 217 , P < 0 . 001 ) , whereas those in humans have an intermediate level of relatedness ( FST = 0 . 067 versus long-tailed macaques , FST = 0 . 104 versus pig-tailed macaques; P < 0 . 001 for both ) . A Bayesian model-based STRUCTURE analysis of multi-locus genotype data from all hosts sampled in Kapit clearly indicated the existence of two sub-population clusters of P . knowlesi ( K = 2; ΔK = 936 . 75 based on Evanno’s estimation of K-population ) ( Fig 3A , S1 Fig and S1 Dataset ) . An individual infection genotype was assigned to be predominantly of a particular cluster if the STRUCTURE analysis score exceeded 0 . 5 for that cluster . All except one of the long-tailed macaque infections were assigned to the Cluster 1 subpopulation , whereas all pig-tailed macaque infections were assigned to the Cluster 2 subpopulation , while 71% of human infections were assigned to Cluster 1 and 29% to Cluster 2 ( Fig 3A , S5 Table ) . A small minority of those which were primarily assigned to either cluster appeared to have a degree of mixed assignment , with scores nearer 0 . 5 than either zero or 1 . 0 for the alternative clusters ( S1 Dataset ) , which is analysed in a separate section below . An independent scan by principal component analysis ( PCA ) showed an almost complete separation between parasites from long-tailed macaques and pig-tailed macaques along the first principal component , while parasites from humans covered the whole distribution and overlapped with all of the samples from both of the macaque hosts ( Fig 3B ) . We analysed a further 367 human P . knowlesi infections from nine other geographical sites ( Fig 1 ) . Most human infections had single P . knowlesi genotypes ( Fig 4A and S1 Dataset ) , and there were no differences in the proportions of mixed genotype infections across all sites ( Comparison across 10 sites including Kapit: Pearson’s X2 , P = 0 . 096; 32% of infections having > 1 genotype overall ) . There were no differences in allelic diversity among the different sites ( HE estimates between 0 . 67 and 0 . 75 , P > 0 . 1 for all pairwise Wilcoxon Signed Rank tests across all 10 loci , S3 Table ) . Pairwise comparisons among genotypes from different infections showed a similar level of diversity at each site , with a median of 2 or 3 identical alleles out of 10 loci in each site ( Fig 4B , S4 Table ) . Every infection had a different multi-locus genotype , and there were virtually none that shared alleles at more than 7 loci , except for nine pairs of identical haplotypes ( three pairs in Betong , three in Miri , and one in each of Sarikei , Tenom and Kelantan ) ( Fig 4B ) . Each identical haplotype pair was shared by infections from different individuals sampled at the same site within the same year , except for two of the identical haplotype pairs in Miri , shared by individual infections sampled one and two years apart ( S4 Table ) . There were two subpopulation clusters ( K = 2 , ΔK = 174 . 94 , S1 Fig ) throughout all of these sites , as had been seen in Kapit , but the relative frequency of the clusters varied geographically ( P < 0 . 0001 , Fig 4C ) . The Cluster 1 subpopulation was more frequent overall , but Cluster 2 was also common at each of the sites in Sarawak , particularly in Miri and Kanowit where it was more frequent than Cluster 1 ( S5 Table and S1 Dataset ) . Over all human infections , there was a similarly high level of divergence in allele frequencies between the two subpopulation clusters as was seen between parasites from the two different macaque host species ( FST = 0 . 194 , P < 0 . 001 ) . As expected , the degree of cluster admixture at each sampling site ( p1*p2 , where p1 and p2 are the local frequencies of Cluster 1 and Cluster 2 respectively ) correlated positively with the ( ISA ) index of multi-locus linkage disequilibrium ( Spearman’s Rho = 0 . 678 , P = 0 . 015 , Fig 5 and S5 Table ) . Analysis of geographical divergence on the basis of FST indices derived from population allele frequencies ( S6 Table ) identified a pattern strongly consistent with isolation by distance ( Mantel test of matrix correlation P < 0 . 0001 , Fig 6A ) . The greatest level of divergence was seen between peninsular Malaysia and Borneo as expected , although isolation by distance was also apparent within Borneo ( Mantel test P = 0 . 0016 ) . The overall pattern consistent with isolation by distance remained when only infections with Cluster 1 genotypes were analysed ( P = 0 . 0016 , Fig 6A ) . There was a similar trend for the smaller number of samples with Cluster 2 genotypes , although this was not significant ( P = 0 . 0922 , S2 Fig ) , indicating that the majority of the geographical differentiation is independent of the Cluster subpopulation structure . A principal component analysis of all individual infection genotypes showed that most of the overall diversity is among those defined as Cluster 1 by the STRUCTURE analysis ( Cluster 2 infections covered only part of the first principal component distribution ) , and infections from peninsular Malaysia are restricted to part of the second principal component distribution ( Fig 6B ) . Combination of the macaque samples together with all of the human samples across the 10 geographical locations confirmed the definition of the two P . knowlesi subpopulation clusters , which correspond to those shown above ( S3 Fig ) . Allele frequency distributions showed that some loci were particularly differentiated between the subpopulation clusters , with FST > 0 . 3 for loci NC03_2 and CD13_61 ( S4 Fig ) . The robustness of the two assigned clusters was confirmed even with the exclusion of these most highly differentiated loci in the STRUCTURE analysis ( S5 Fig ) . Most individual infection genotypes had a clear majority of putative ancestry assignment to either Cluster 1 or Cluster 2 , but a small minority of infections had a more intermediate profile ( Figs 3A and 4C ) . Quantitative analysis of the proportional Cluster 1 and Cluster 2 ancestry assignments for each infection genotype based on the STRUCTURE analysis yielded an index of the degree of intermediate cluster assignment for each infection . This has a maximum possible value of 0 . 5 , although most infections had values closer to zero . The intermediate cluster assignment indices showed no difference between single and mixed genotype human infections ( Mann-Whitney test P = 0 . 20 , Fig 7A ) , whereas both of these independently had higher indices than the macaque infections ( P < 0 . 001 for both comparisons ) . When analysis was focused on Kapit alone , the distribution of intermediate cluster assignment indices were not significantly different between human and macaque infections ( P = 0 . 25 , Fig 7B ) . However , there were geographical differences , with human infections from Kelantan having a significantly higher distribution of values compared to five of the sites in Borneo ( Mann-Whitney test P < 0 . 05 for each comparison after Bonferroni correction , Fig 7B ) . Across the different sites , there was no significant correlation between the local population admixture of both clusters ( p1*p2 , S5 Table ) and the mean or variance of intermediate cluster assignment indices ( P = 0 . 33 and P = 0 . 59 respectively ) . Infections which had intermediate cluster assignment ( index values > 0 . 25 ) were not particularly closely related , having a similar degree of allele sharing as seen in the general local populations ( S6 Fig , compared with Figs 2B and 4B ) .
We show that human P . knowlesi is an admixture of two divergent parasite populations associated with different forest-dwelling macaque reservoir hosts . In human infections , the long-tailed macaque-associated P . knowlesi type ( Cluster 1 ) is most common overall and at most of the geographical sites , while the pig-tailed macaque-associated type ( Cluster 2 ) is also common at sites in Sarawak . The estimate of divergence between these two sympatric parasite subpopulations ( FST index of ~ 0 . 22 averaged over 10 microsatellite loci ) may be conservative , due to high allelic diversity of the microsatellite loci which restricts the potential upper range of fixation indices [18 , 19] . The differentiation varied among the loci , with two of the microsatellite loci being particularly highly differentiated between the clusters ( FST ~ 0 . 35 ) , so the robustness of the two assigned clusters was confirmed by repeat analyses which excluded these . Previous analysis of P . knowlesi mitochondrial DNA sequences from a relatively small number of human and long-tailed macaque infections in Kapit did not indicate two divergent lineages [15] , although analysis of samples from Sabah suggests that sequences from pig-tailed macaque infections are differentiated from sequences from long-tailed macaque infections [20] . The results confirm that humans have mostly single genotype P . knowlesi infections whereas macaques have polyclonal infections , supporting the expectation that there is a higher rate of transmission among macaques [15 , 21] . The estimated number of genotypes per infection here is a minimum number based on the alleles detected , and it is possible that some infections may have contained additional parasite clones that were not detected , due to having low density in the blood or having similar alleles to the ones detected . The number of P . knowlesi genotypes detected per infection in humans is lower than was previously seen in microsatellite analyses of the endemic human malaria parasites P . falciparum and P . vivax in some of the same areas in Malaysia [22 , 23] , whereas the number of P . knowlesi genotypes per infection in macaques is much higher . Levels of multi-locus linkage disequilibrium in P . knowlesi here are lower than reported in P . vivax or P . falciparum in these areas [22 , 23] , indicating that recombination in P . knowlesi probably commonly occurs in mosquitoes containing a macaque blood meal with multiple parasite genotypes . It is unknown how the two sympatric P . knowlesi subpopulations are genetically isolated . The observation of a single long-tailed macaque with a P . knowlesi Cluster 2 type infection ( otherwise only seen in pig-tailed macaques and humans ) suggests there is not an absolute barrier in terms of primate host susceptibility , although there are differences in ecology . Additional sampling of both long-tailed and pig-tailed macaques will be important to confirm the host associations of different parasites [20] . Both macaque species are widespread , but long-tailed macaques prefer secondary forest near human settlements where they have access to farms for food , whereas pig-tailed macaques spend more time in ground foraging in primary forests , generally having less frequent contact with humans [24] . There may be differential susceptibility of mosquito species to the respective parasite types , as suggested for subpopulations of another malaria parasite elsewhere [25] , or different mosquitoes may feed on the respective macaque host species . Genetic differentiation in P . knowlesi was also strongly correlated with geographical distance , overall and for the Cluster 1 parasites . The observation of highest FST values between populations from Malaysian Borneo and Peninsular Malaysia was expected , as the South China Sea has separated macaques in these areas since the last glacial period [26] , but a test for isolation by distance remained significant when analysing only sites within Borneo . A small minority of human infections had intermediate cluster assignment indices , which could potentially result from occasional crossbreeding between the two genotypic clusters , although this cannot be concluded from these data alone . Hybridisation between species or sub-species can offer opportunities for adaptation , and has been associated with emergence of novel host-specificity or pathogenicity in other parasitic protozoa [27] and fungi [28] . Switching of host species has occurred repeatedly in malaria parasites of birds [29] and small mammals [30] , as well as apes and humans [3 , 4] , but the occurrence of parasite hybridisation and introgression has not been investigated . The potential occurrence of inter-cluster hybridisation in even a minority of human P . knowlesi infections , combined with the possibility of human-mosquito-human transmission , may increase the potential for P . knowlesi adaptation to the human host or to mosquito species that are more abundant than the currently known forest-associated vectors . Genome-wide analysis of P . knowlesi populations would enable further evaluation of the genetic structure of this zoonotic parasite species , and allow scanning for loci under selection within each of the two subpopulations . Human clinical isolates containing single species infections would be relatively straightforward to analyse , as P . knowlesi sequences would be unmixed with those of other human malaria species . In contrast , as natural macaque infections usually contain a mixture of different malaria parasite species [15] , to obtain unambiguous genome sequences it may be necessary to sequence from individual parasites isolated from these hosts [31] . Although experimental studies on P . knowlesi are usually conducted in vivo in non-human primates [32–34] , new approaches to adapt the parasites to in vitro growth using human erythrocytes have been successful [35 , 36] . Analysis of phenotypic differences between the different host-associated types may be investigated using both in vivo and in vitro experimental systems , while continued epidemiological and clinical surveillance for increasing incidence or disease severity is of the highest priority .
Human blood samples were taken after written informed consent had been obtained from patients . This study was approved by the Medical Research and Ethics Committee of the Malaysian Ministry of Health ( Reference number: NMRR-12-1086-13607 ) , which operates in accordance to the International Conference of Harmonization Good Clinical Practice Guidelines . Animal sampling was carried out as previously described [15] in strict accordance with the recommendations by the Sarawak Forestry Department for the capture , use and release of wild macaques . A veterinarian took a venous blood sample from each macaque following anesthesia by intramuscular injection of tiletamine and zolazepam , and all efforts were made to minimize suffering by collecting blood at the trap sites and releasing the animals immediately after the blood samples had been obtained . The Sarawak Forestry Department approved the study protocol for capture , collection of blood samples and release of wild macaques ( Permit Numbers: NPW . 907 . 4 . 2–32 , NPW . 907 . 4 . 2–97 , NPW . 907 . 4 . 2–98 , 57/2006 and 70/2007 ) . A permit to access and collect macaque blood samples for the purpose of research was also obtained from the Sarawak Biodiversity Centre ( Permit Number: SBC-RP-0081-BS ) . A total of 599 DNA samples from different P . knowlesi infections of humans and macaques were analysed from collections performed at 10 different geographical sites ( Fig 1 ) . 552 samples were from human P . knowlesi malaria patients from all of the sites , eight in Malaysian Borneo ( Sarawak and Sabah states ) and two in Peninsular Malaysia ( Kelantan and Pahang states ) . For samples from Sarawak , DNA was extracted at the University Malaysia Sarawak ( UNIMAS ) in Kuching from previously reported blood samples collected between 2000 and 2011 [6 , 10 , 37 , 38] as well as new samples collected in 2012 and 2013 , allowing analysis of five sites: Kapit ( n = 185 ) , Betong ( n = 78 ) , Kanowit ( n = 34 ) , Sarikei ( n = 27 ) and Miri ( n = 50 ) . Samples from Sabah were collected in 2013 , and DNA was extracted by the Sabah Public Health Reference Laboratory , allowing analysis of three sites: Kudat ( n = 30 ) , Ranau ( n = 42 ) and Tenom ( n = 26 ) . For Peninsular Malaysia , blood samples collected from Kelantan ( n = 30 ) and Pahang ( n = 50 ) underwent DNA extraction at the Institute for Medical Research in Kuala Lumpur . We analysed DNA from blood samples of a total of 47 wild macaques ( long-tailed macaque , Macaca fascicularis n = 37; pig-tailed macaque , M . nemestrina n = 10 ) previously collected within 30 km radius of Kapit town in Sarawak [15] . The map locations and dates of the individual macaque sampling is shown in S7 Fig . The presence of P . knowlesi DNA was confirmed in all samples at UNIMAS by nested PCR assays [6 , 15] . A combination of three microsatellite mining tools ( iMEX [39] , mreps [40] , and MSATCOMMANDER [41] ) were used to identify simple sequence repeat loci from the P . knowlesi reference genome [42] . Loci with perfect tri-nucleotide simple repeat sequences were carefully selected using customised perl-script commands based on narrow criteria to maximise their likely utility for genotyping: i ) a minimum of 7 repeat copies in each microsatellite in the reference sequence , ii ) located at non-telomeric chromosomal regions as defined by regions syntenic with the P . vivax reference genome [43] , iii ) absence of any homopolymeric tracts adjacent to the microsatellite sequence that could give rise to additional size polymorphism . As a result , 19 trinucleotide repeat loci widely spaced in the genome ( S1 Table ) were shortlisted and PCR primers were designed using PrimerSelect software ( DNASTAR , USA ) for hemi-nested PCR assays . The specificity of PCR was tested using DNA controls of all human Plasmodium species , common malaria parasites of the Southeast Asian macaques ( P . knowlesi , P . coatneyi , P . inui , P . cynomolgi and P . fieldi ) , as well as human , long-tailed and pig-tailed macaque DNA . Loci for which primers showed complete specificity of amplification from P . knowlesi were tested further for genotyping performance . Genotyping of each microsatellite locus was performed using a hemi-nested protocol with a fluorescent dye-labelled inner primer during the second round PCR amplification ( primers listed in S1 Table ) . Both first and second round PCR amplifications were conducted in individual tubes or wells for each locus , in 11 μl reaction volume containing 0 . 2 mM each dNTP ( Bioline , UK ) , 2 mM MgSO4 , 1X ThermoPol II reaction buffer ( NEB , UK ) , 0 . 275 U Taq DNA polymerase ( NEB , UK ) , 0 . 1 μM of each forward and reverse primer , and 1 μl sample DNA template . The PCR cycling conditions were as follows: initial denaturation at 94°C for 2 min , followed by 28 cycles of 94°C for 30 sec , annealing at 56°C for 30 sec and elongation at 68°C for 30 sec , with a final elongation step at 68°C for 1 min . Final PCR products were pooled into three groups of loci with different product size and dye profiles together with Genescan 500 LIZ molecular size standards ( Applied Biosystems , UK ) and run on a Genetic Analyzer 3730 capillary electrophoretic system ( Applied Biosystems , UK ) . GENEMAPPER version 4 . 0 software ( Applied Biosystems , UK ) was used for scoring of allele electrophoretic size , and quantification of peak heights . Infections containing multiple haploid parasite genotypes were apparent as multiple electrophoretic peaks for a locus corresponding to different alleles . The apparent genotypic multiplicity of infection ( MOI ) was determined by the locus with the most alleles detected in the infection , considering peaks with height of at least 25% relative to the predominant allele within each isolate . The predominant allele per locus within each infection was counted for subsequent population genetic analyses . Allelic diversity at each locus was measured as the virtual heterozygosity ( HE ) using FSTAT software version 2 . 9 . 3 . 2 ( http://www2 . unil . ch/popgen/softwares/fstat . htm ) , and allele frequency distributions were also inspected using GenAlEx version 6 [44] within the Microsoft Excel platform . Genetic differentiation between each population was measured by pairwise fixation indices ( FST ) using FSTAT , with Bonferroni correction on a nominal significance level of 0 . 05 applied for multiple comparisons across the population pairs . To test for correlation between genetic differentiation and geographical distance , a Mantel test for isolation by distance was performed with Rousset’s linearised FST/ ( 1-FST ) plotted against the natural log of geographic distance using Genepop version 4 . 2 [45] . The relatedness of haplotypes between individual isolates was assessed by measuring the pairwise proportion of shared alleles , excluding samples with missing data at any locus . A matrix of pairwise similarity among isolates was calculated based on the identical or mismatched alleles from a complete set of loci and the distribution of shared alleles between sample pairs for each population was visualised using a customised perl-script command . To test for non-random allele assortment , multi-locus linkage disequilibrium ( LD ) was assessed by the standardised index of association ( IAS ) using LIAN version 3 . 6 [46] , with significance of the IAS values tested by Monte-Carlo simulation with 10 , 000 data permutations to generate the null distribution under linkage equilibrium . To explore evidence of population substructure in the entire population , a Bayesian analysis was performed using the STRUCTURE version 2 . 3 . 4 software [47] using samples with no missing data at any locus . Individuals in the population pool were clustered to the most likely population ( K ) by measuring the probability of ancestry using the multi-locus genotype data . The program parameters were set to admixture model with correlated allele frequency , with 50 , 000 burn-in period and 100 , 000 Markov chain ( MCMC ) iterations . To run the simulation , K value was predefined from 1–10 and the run was performed in 20 replicates for each K . The most probable K value was then calculated according to Evanno’s method [48] using the webpage interface STRUCTURE Harvester [49] . The assignment of a sample to a subpopulation cluster was based on the inferred cluster scores by STRUCTURE analysis , where samples with inferred cluster scores within a range in relation to the K-value were assigned together as one subpopulation cluster . The intermediate cluster assignment indices were calculated based on the proportion of shared cluster ancestries per individual isolate inferred by the cluster scores from the STRUCTURE analysis . We also independently performed a principal component analysis ( PCA ) using the GenAlEx package for the same purpose . Samples with missing data at any locus were excluded , and the genetic distance matrix was generated based on the allelic mismatches between pairs of isolates . A two-dimensional PCA plot was generated considering the first two highest eigenvalues , and genetic clusters were determined based on the eigenvector coordinates along the axes of variation .
|
Extraordinary phases of pathogen evolution may occur during an emerging zoonosis , potentially involving adaptation to human hosts , with changes in patterns of virulence and transmission . In a large population genetic survey , we show that the malaria parasite Plasmodium knowlesi in humans is an admixture of two highly divergent parasite populations , each associated with different forest-dwelling macaque reservoir host species . Most of the transmission and sexual reproduction occurs separately in each of the two parasite populations . In addition to the reservoir host-associated parasite population structure , there was also significant genetic differentiation that correlated with geographical distance . Although both P . knowlesi types co-exist in the same areas , the divergence between them is similar to or greater than that seen between sub-species in other sexually reproducing eukaryotes . This may offer particular opportunities for evolution of virulence and host-specificity , not seen with other malaria parasites , so studies of ongoing adaptation and interventions to reduce transmission are urgent priorities .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Admixture in Humans of Two Divergent Plasmodium knowlesi Populations Associated with Different Macaque Host Species
|
Visna/Maedi , or ovine progressive pneumonia ( OPP ) as it is known in the United States , is an incurable slow-acting disease of sheep caused by persistent lentivirus infection . This disease affects multiple tissues , including those of the respiratory and central nervous systems . Our aim was to identify ovine genetic risk factors for lentivirus infection . Sixty-nine matched pairs of infected cases and uninfected controls were identified among 736 naturally exposed sheep older than five years of age . These pairs were used in a genome-wide association study with 50 , 614 markers . A single SNP was identified in the ovine transmembrane protein ( TMEM154 ) that exceeded genome-wide significance ( unadjusted p-value 3×10−9 ) . Sanger sequencing of the ovine TMEM154 coding region identified six missense and two frameshift deletion mutations in the predicted signal peptide and extracellular domain . Two TMEM154 haplotypes encoding glutamate ( E ) at position 35 were associated with infection while a third haplotype with lysine ( K ) at position 35 was not . Haplotypes encoding full-length E35 isoforms were analyzed together as genetic risk factors in a multi-breed , matched case-control design , with 61 pairs of 4-year-old ewes . The odds of infection for ewes with one copy of a full-length TMEM154 E35 allele were 28 times greater than the odds for those without ( p-value<0 . 0001 , 95% CI 5–1 , 100 ) . In a combined analysis of nine cohorts with 2 , 705 sheep from Nebraska , Idaho , and Iowa , the relative risk of infection was 2 . 85 times greater for sheep with a full-length TMEM154 E35 allele ( p-value<0 . 0001 , 95% CI 2 . 36–3 . 43 ) . Although rare , some sheep were homozygous for TMEM154 deletion mutations and remained uninfected despite a lifetime of significant exposure . Together , these findings indicate that TMEM154 may play a central role in ovine lentivirus infection and removing sheep with the most susceptible genotypes may help eradicate OPP and protect flocks from reinfection .
Visna/Maedi virus ( VMV ) and caprine arthritis encephalitis virus ( CAEV ) are small ruminant lentiviruses ( SRLV ) of the retroviridae family [1] that infect sheep and goats in major sheep producing countries worldwide . The exceptions are Iceland where VMV was eradicated after a 30-year effort [2] , and Australia and New Zealand where VMV has not been reported in sheep but CAEV has been reported in goats [3] , [4] . Once infected , seroconversion typically occurs within weeks to months and the infection is incurable . Sheep do not usually display signs of clinical disease in the first two years of infection . The first signs of disease are often loss of body condition and indurative mastitis ( i . e . , thin ewe syndrome and hard udder ) . When disease develops , severe clinical signs may include difficulty breathing , chronic wasting , loss of motor control , and arthritis . Ovine progressive pneumonia virus ( OPPV ) is a closely related North American counterpart to VMV and typically produces an interstitial pneumonia . Seroprevalence studies of U . S . sheep have shown that 36% of sheep operations have infected animals and 24% of all animals tested were seropositive [5] . The impact of subclinical OPPV infection is significant and includes detrimental effects on sheep production from breeding through weaning [6] , [7] , [8] . Considering that losses are cumulative during an animal's lifetime , the negative effects on ewe production and the sheep industry are substantial . Natural transmission of ovine lentiviruses is primarily among adults , occurs most frequently after their first year [9] , [10] , [11] , [12] , [13] , and is by the respiratory route [14] , [15] , [16] . In addition , some infections occur in lambs by ingestion of infected colostrum and milk [8] , [17] , [18] , [19] , [20] , [21] . Ovine lentiviruses are macrophage-tropic but not T-lymphocyte-tropic and thus do not cause an immunodeficiency in sheep [22] , [23] , [24] , [25] , [26] . Persistence of ovine lentivirus in infected sheep is attributed to latent proviral DNA sequences integrated into the genome of a small fraction of monocytes circulating in the blood . Proviral DNA transcription and gene expression is suppressed until infected monocytes mature into macrophages as they migrate into the interstitial spaces of affected organs [27] , [28] . Once in the target organs , infected macrophages initiate viral replication , which induces an inflammatory cascade that ultimately attracts more infected monocytes and other leukocytes . These lesions increase progressively , terminating in disease and eventual death . Although there is no cure , the impact of disease can be reduced by lowering the prevalence . Voluntary SRLV control programs have been established in several European countries [29] , [30] , [31] , [32] , [33] , [34] . OPP can be eradicated by testing and removing infected animals or by isolating lambs from seropositive dams at birth . The lambs are then raised on uninfected colostrum and milk , and maintained separately from seropositive sheep thereafter . Either of these methods may be used alone , or in combination , to break the cycle of transmission . However , an OPP-free flock is still susceptible to infection if exposed to other infected sheep or goats [35] . Thus , efforts to eradicate OPP and maintain infection-free status would be facilitated if replacement breeding stock were genetically resistant to lentivirus infections . Examples of genetic resistance to lentivirus infection have been documented in human populations . Nearly all individuals who lack the lentivirus co-receptor CCR5 do not acquire human immunodeficiency virus ( HIV ) infection after significant exposure [36] , [37] , [38] . Moreover , an infected person receiving transplanted stem cells lacking CCR5 may be cured of HIV [39] . In the cases of VMV and OPPV , reports have suggested that host resistance to lentiviral infection may also occur in sheep [40] , [41] , [42] , [43] . Significant breed effects on seroprevalence have also been observed in comingled flocks of purebred sheep , further indicating possible host genetic restriction [10] , [12] . For example , in U . S . sheep the OPPV seroprevalence in purebred Finnsheep , Texel , and Suffolk was 77 , 65 , and 15% , respectively [12] . In Basque dairy-sheep , seroconversion was strongly associated with lifetime maternal VMV-serological status and was interpreted as evidence of genetic susceptibility [44] . The present article reports findings from a genome-wide association study ( GWAS ) that used naturally-exposed ewes , together with the International Sheep Genome Consortium SNP50 marker set , to test for genetic association with lentivirus infection . Ovine DNA sequence variation in a transmembrane protein gene ( TMEM154 ) was associated with lentivirus infection . The ancestral TMEM154 allele encodes a 191 amino acid polypeptide with glutamate ( E ) at position 35 and is associated with infection susceptibility . A mutant TMEM154 allele encodes lysine ( K ) at position 35 allele and is associated with reduced susceptibility . Two deletion mutations were also observed in TMEM154 , however there were not enough individuals with these deletions to test their effect . Together , these results suggest that TMEM154 may play a central role in ovine lentivirus biology .
The presence of OPPV infection was tested with a competitive enzyme-linked immunosorbent assay ( cELISA ) in 3 , 545 breeding-age sheep from purebred and crossbred research flocks in South Central Nebraska , USA . This cELISA has high sensitivity ( 98 . 6% ) and specificity ( 96 . 9% ) in sheep naturally infected with OPPV [45] . Analysis by age class showed OPPV infection was lowest in 1-year-olds ( 8% ) , increased with age , and peaked at age 5 ( 43% , Figure 1A ) . From age 5 to 8 years , the number and proportion of OPPV-infected sheep declined in each year , indicating that the older infected sheep were leaving the flock at a faster rate than their uninfected flock mates . These results indicated that , by age 4 , most sheep received sufficient OPPV exposure for infection to occur and that uninfected ewes appeared to have greater longevity in these flocks . Although age is a risk factor for infection , seroprevalence varied widely within age class , depending on breed composition ( Figure 1B ) . To examine the possibility that genetic risk factors may influence susceptibility to OPPV , matched case-control pairs consisting of infected and uninfected ewes were selected ( Table 1 ) . The strict matching criteria were intended primarily to reduce the variation in breed composition and OPPV exposure within each pair . The matching procedure identified 130 case-control pairs of 4- to 9-year-old ewes ( Table 1 ) . These pairs were used in a two-stage design with the goal of reducing falsely positive marker associations and minimizing the number of costly genome-wide scans . For the genome-wide association phase of the study , 69 pairs of 5- to 9-year-old ewes ( white bars in Figure 1C ) were evaluated first , while 61 matched pairs of 4-year-old ewes were held in reserve for verification of GWAS results . Single nucleotide polymorphisms ( SNPs ) in the Ovine SNP50 BeadChip array ( n = 54 , 241 ) were scored in 69 matched case-control pairs and tested for association with OPPV . The experimental design was estimated to have a detectable relative risk of genetic association that ranged from two to six in dominant and co-dominant models of inheritance , depending on marker allele frequency , and the extent of linkage disequilibrium ( LD ) between a marker and a disease allele ( Materials and Methods ) . Of the 54 , 241 SNPs tested , 50 , 614 had quality scores in the acceptable range as determined by clustering and genotype calling algorithms . A single SNP on chromosome 17 had an unadjusted p-value of 3 . 19×10−9 ( OAR17_5388531; Figure 2A ) . This was highly significant compared to the significance threshold of 1×10−6 ( i . e . , a significance level of 0 . 05 divided by 50 , 614 ) . Moreover , the Quantile-Quantile ( Q-Q ) plot showed no evidence of an inflated test statistic caused by population structure . The c/t SNP OAR17_5388531 was in intron 5 of an ovine gene homologous to the human TMEM154 gene on chromosome 4 . The “c” allele of SNP OAR17_5388531 was on the sense strand of TMEM154 , had a frequency of 0 . 257 , and was associated with infected sheep . Another SNP ( s46403 ) had the second lowest unadjusted p-value ( 2 . 22×10−6 ) and was on chromosome 13 in a gene similar to human angiopoietin 4 ( ANGPT4 ) . A third SNP ( OAR17_5405721 ) had the third lowest unadjusted p-value ( 6 . 71×10−6 ) and was located in the 3′UTR of ovine TMEM154 . The highly significant association of one SNP in ovine TMEM154 , together with the third best SNP association being located in the same gene , suggested that a genetic risk factor associated with OPPV infection existed in this genomic interval . Subsequent efforts were directed towards characterizing the genomic region of ovine TMEM154 , discovering additional polymorphisms , and testing them for association with infection . The complete sequence of the TMEM154 region was not available for sheep and thus was determined by identifying and sequencing four overlapping bacterial artificial chromosomes ( BACs ) spanning approximately 400 kb . A contiguous 78 kb region was assembled de novo and appeared to contain the complete TMEM154 gene region ( Figure 2B , GenBank Accession HM355886 ) . Other contigs from these BACs contained exons similar to human ARFIP1 and FBXW7 . The ovine genes appeared to be in the same orientation and approximate positions as those in reported for ARFIP1 , TMEM154 , and FBXW7 on human chromosome 4 and cattle chromosome 17 . Sanger sequencing of targeted genomic DNA fragments amplified by polymerase chain reaction ( PCR ) in the ARFIP/TMEM154/FBXW7 region revealed 128 additional SNPs in the 69 pairs of matched 5- to 9-year-olds . However , SNPs associated with OPPV infection were observed only within the TMEM154 gene ( Figure 2B ) . The results indicated that TMEM154 , and not flanking genes , was the likely source of the association . Although sequence variation in any number of gene elements can alter biological function , those that may directly affect the polypeptide sequence were evaluated first . The ovine TMEM154 genomic assembly contained seven TMEM154 exons encoding a 191 amino acid precursor protein ( Figure 3A ) . The precursor protein contained a putative signal peptide at the N-terminus with a cleavage site predicted between positions 30 and 31 and resulted in a mature protein of 161 amino acids . The predicted mature ovine TMEM154 protein was 92 . 5 , 67 . 3 , and 53 . 8% identical with those of cattle , humans , and mice , respectively . The extracellular domain and signal peptide accounted for most of the amino acid sequence differences in these comparisons ( 83 , 31 , and 25% identity , respectively ) . The ovine intron/exon junctions were established by comparing the genomic sequences with those from a 1 , 012 bp reverse transcription ( RT ) -PCR fragment amplified from cDNA of contemporary animals . The ovine TMEM154 mRNA and exon structure was similar to those reported for cattle , human , and mice ( data not shown ) . Although RNA samples were not available for the 69 pairs of case-control sheep , the transcript sequence was determined for 11 case-control pairs of contemporary sheep and seven other available sheep . In all 29 sheep tested , the expected full-length transcripts were observed and their sequences corresponded to those from genomic DNA . Thus , alternatively spliced TMEM154 transcripts did not explain the association observed with the SNP OAR17_5388531 . To evaluate whether amino acid sequence variation was encoded by ovine TMEM154 , the exons were amplified from genomic DNA and sequenced for a panel of 234 animals that included all 69 matched case-control pairs and 96 rams representing common U . S . sheep breeds . In these 234 animals , five missense SNPs ( T25I , D33N , E35K , T44M , N70I ) and two frameshift deletion polymorphisms ( R4AΔ , E82YΔ ) were observed in the predicted signal peptide and the extracellular domain ( exons 1 and 2 , Figure 3A ) . Conversely , nonsynonymous SNPs and frameshift polymorphisms were not observed in exons 3 through 7 in any of these sheep . TMEM154 exons 1 and 2 were then considered as potential “hotspots” for coding polymorphisms , and these exons were sequenced for more than 5000 sheep from research populations , revealing one additional missense SNP ( L14H ) . Combinations of the eight “coding” polymorphisms were observed on haplotypes encoding eight distinct precursor protein isoforms . Four haplotypes were predicted to encode full-length polypeptides with glutamate ( E ) at position 35 ( Figure 3B , designated 2 , 3 , 9 , and 11 ) . The E35 allele was in strong LD with the “c” allele of OAR17_5388531 associated with infected cases ( r2 = 0 . 98 ) . Two haplotypes ( designated 1 and 10 ) encoded full-length polypeptides with lysine ( K ) at position 35 . The remaining haplotypes ( 4 and 6 ) had frameshift deletions predicted to cause premature termination of translation and loss of the putative membrane spanning and cytoplasmic domains of TMEM154 . Comparing polymorphic ovine TMEM154 amino acid residues with those in related mammalian species indicated that haplotype 3 was the most likely ancestral isoform in sheep . Thus , the ancestral ovine precursor protein isoform is inferred to be a 191 amino acid polypeptide with a negatively charged E35 residue . Haplotype comparisons between mammalian species also showed the E35 residue is highly conserved in mammals and the positively charged K35 residue of TMEM154 was not observed in other species analyzed ( Figure 3B ) . A median-joining network of haplotypes encoding polypeptide isoforms of ovine TMEM154 showed that the two truncated isoforms were located on the distal branches of the tree ( Figure 3C , haplotypes 4 and 6 ) . Because the more recent haplotypes appeared to have evolved towards dysfunction and OPPV resistance , relationships presented in Figure 3C provide a framework for evaluating the potential role of TMEM154-encoded polypeptide isoforms in ovine lentivirus infection . We hypothesized that the more ancient full-length TMEM154 haplotypes encoding E35 were genetic risk factors for OPPV infection because these alleles were in strong LD with the “c” allele of OAR17_5388531 . Thus , sheep with one copy of haplotype 2 or 3 were compared to those without . Because the experimental design included paired samples , the McNemar's test for two correlated proportions was used for the analysis [46] . Also , the 61 matched pairs of 4-year-old ewes were deployed at this stage for comparison with results obtained from the 69 matched pairs of 5- to 9-year-old ewes . The dichotomous variable for this test was defined as having zero or one copy of the TMEM154 genetic risk factor ( i . e . , haplotype 2 or 3 ) . In one discordant type , the infected case had one copy of a TMEM154 risk factor , but the uninfected control did not ( Table 2 and Table S1 ) . This type of discordant pair ( n = 28 ) was consistent with the hypothesis . In the other discordant type , the infected case had zero copies of a TMEM154 risk factor , and the uninfected control had one copy . This type of discordant pair ( n = 1 ) was inconsistent with the hypothesis . In the 61 matched pairs of 4-year-old ewes , the odds ratio of these discordant pairs indicated that ewes with one copy of a full-length TMEM154 E35 allele were 28 times more likely to be infected than those without ( p-value<0 . 0001 , 95% CI 5–1100 ) . Compared to either SNP for OAR17_5388531 or E35K , the haplotype model for TMEM154 yielded more significant results ( Table 2 ) . In this study , the predictive value of having one versus two copies of TMEM154 risk factor alleles could not be determined because there were too few pairs of this type . Although an additive model was not excluded , one copy of the risk factor was significantly correlated with OPPV infection . If this were a diagnostic test for the 260 sheep in these 130 pairs , the predictive value ( PV ) for a positive test would be 85% with a sensitivity and specificity of 69% and 88% , respectively . These results were consistent with those of the GWAS and confirmed that TMEM154 haplotypes encoding polypeptide variants were associated with OPPV infection . Estimating the relative risk ( RR ) and PV of TMEM154 haplotype alleles 2 and 3 provides an indication of the potential impact of full-length E35 alleles in other affected flocks . A total of 2705 sheep , 3-years and older , were evaluated in nine cohort studies to test TMEM154 haplotype alleles 2 and 3 as risk factors for OPPV infection . These cohorts consisted of sheep from various breeds , ages , and production environments and were sampled over a span of seven years from research flocks in Nebraska and Idaho , and a private flock in Iowa . Animals were matched for gender and production environment in all cohorts; for three cohorts , animals were also matched for age at sampling . For each cohort , two-way contingency tables were used to analyze the relationship between the presence of TMEM154 haplotypes 2 or 3 and being infected with OPPV ( Table S2 and Table S3 ) . For all cohorts , the combined RR of OPPV infection for animals with TMEM154 haplotypes 2 or 3 was 2 . 85-times greater than those without ( 95% CI 2 . 36–3 . 43 , p-value<0 . 0001 ) . These cohort studies also confirmed that TMEM154 haplotype alleles 2 and 3 are significant risk factors for OPPV infection in various breeds , and may be associated with lentivirus infection in multiple geographic locations and environments . The frequency of TMEM154 haplotype risk factors within breeds provides an indication of their potential susceptibility of OPPV in production environments similar to those described here . The combined frequencies of risk factor alleles 2 and 3 was highest in Texel ( 0 . 74 ) and lowest in Rambouillet ( 0 . 035 , Table 3 ) and generally consistent with seroprevalence trends in the research flocks ( Figure 1B ) . The most common truncated isoform of TMEM154 was encoded on haplotype 4 , which was detected in Katahdin ( 0 . 15 ) , Suffolk ( 0 . 13 ) , Composites ( 0 . 033 ) , Rambouillet ( 0 . 005 ) , and Polypay ( 0 . 003 ) . Overall , ovine TMEM154 haplotypes encoding polypeptide isoforms 1 , 2 , 3 , and 4 accounted for more than 99% of the haplotypes observed .
Ovine gene regions including CCR5 and DRB1 were previously associated with OPP provirus levels in candidate gene studies [41] , [42] , [43] . However , the OvineSNP50 BeadChip marker density and distribution were not sufficient for evaluating whether CCR5 and DRB1 gene regions were associated with OPPV infection in the present study . The nearest SNPs flanking CCR5 ( OAR19_55954161 and s65253 ) were approximately 20 kb from the coding region and had unadjusted p-values of 0 . 62 and 0 . 75 , respectively . The degree of LD between the BeadChip SNP markers and those previously identified for CCR5 are unknown . The nearest SNPs flanking DRB1 ( OAR20_ 26932949 and OAR20_27259292 ) were greater than 100 kb from the coding region and had unadjusted p-values of 0 . 77 and 0 . 21 , respectively . For either gene , there were no markers within 1 Mb that had unadjusted p-values less than 0 . 01 . We acknowledge that a GWAS with 50 k SNPs and 69 case-control pairs limits the detectable genetic risk factors to those with large effects . Thus , other genetic risk factors for OPP may exist and were missed for lack of power in our study . In addition , it is not known whether TMEM154 genetic risk factors are associated with specific OPPV strains such as those found in Nebraska , Idaho , and Iowa , USA . Strain differences , together with adverse production conditions like high animal density , indoor housing with poor ventilation , and moist climates , may enhance transmission and overcome any or all host genetic resistance . Thus , it is not known whether genetic variants of TMEM154 will be useful predictors of OPPV infection under other management and/or environmental conditions . The success of the present GWAS study was dependent on seven key features of the research design , without which the allelic association may have escaped detection . The first important feature was the availability of a serological test with good sensitivity and specificity for correctly classifying infection status . Serological results from duplicate blinded samples tested in two laboratories indicated that less than 2% of the animals were inconsistently classified in the initial round of testing . Second , the matched case-control design with older sheep was a key feature for reducing variation in the management conditions , environment , breed composition , and pathogen exposure . The use of older sheep increased the chances that sufficient natural exposure had occurred so that a high proportion of susceptible individuals could become infected . Third , it was important to have sufficient numbers of older sheep to assemble enough matched pairs to detect an association . Of the older sheep tested , only 22% met the matching criteria . Fourth , the relatively diverse breed composition of the research flocks increased the likelihood that an association observed in the 69 matched pairs was not limited to one breed . Fifth , public availability of the OvineSNP50 BeadChip was essential to progress beyond a functional candidate gene approach for discovering gene-phenotype associations . TMEM154 had not previously been identified as a candidate gene in lentivirus biology for any species and thus would not have been considered . Sixth , the SNP marker spacing on the 50 k chip in the region of TMEM154 was fortuitous because a GWAS may have missed the TMEM154 association if the SNP density was lower or the distribution of SNPs happened to be less serendipitous . A higher density SNP chip would increase the chances of a marker SNP being in LD with a polymorphism that influences the trait of interest . A higher density SNP chip may also rule out the association of neighboring genes , and thereby narrow the region of focus . Seventh , TMEM154 in the study populations had three common haplotypes encoding polypeptide isoforms , two of which formed a risk factor group with a large effect . This was previously unknown and was determined by the evolutionary history of TMEM154 in these sheep . Nevertheless , this report demonstrates that a GWAS approach with 50 k SNPs and 69 matched case-control pairs was successful in sheep . Although TMEM154 haplotype risk factors 2 and 3 were strongly associated with OPPV infection , some animals without these haplotypes were also infected . For example , 36 of 139 sheep with a 1 , 1 diplotype were seropositive in the pairs of matched case-control sheep . This is consistent with the concept that host genetic resistance is conditional . Many factors may contribute to a virus overcoming host genetic resistance including: a high viral dose during an exposure event , a long duration of repeated viral exposures , viral genetic adaptation to host defenses , and multiple routes by which infection may occur . In the latter case , other host-encoded genes may play significant roles . Thus , comparing the relative level of resistance conferred by various TMEM154 haplotypes , together with the identification of additional host genetic risk factors , will be important for developing flocks that are genetically resistant to lentivirus infections .
Prior to their implementation , all animal procedures were reviewed and approved by the care and use committees at the United States Department of Agriculture ( USDA ) , Agricultural Research Service ( ARS ) Meat Animal Research Center ( USMARC ) in Nebraska , the USDA , ARS , Sheep Experiment Station ( USSES ) in Idaho , and Washington State University in cooperation with the USDA , ARS , Animal Disease Research Unit ( ADRU ) . The USMARC ( Nebraska ) sheep population was sampled in 2003 ( n = 3545 ) and used to select 69 matched case-control pairs of 5- to 9-year-old ewes for the GWAS . The same population was also used to select 61 matched case-control pairs of 4-year-old ewes for analyzing TMEM154 haplotypes as risk factors for OPP infection . Animals not used in matched case-controls were used in unmatched cohort studies for validation as shown in Table S2 . Animals were not members of more than one group . The USMARC sheep population is a relatively diverse flock with more than ten breeds representing genetic diversity for traits such as fertility , prolificacy , maternal ability , growth rate , carcass leanness , wool quality , mature weight and longevity [47] . The USMARC population was sampled again in 2010 and used to select a cohort of 280 ewes , 4- to 5-year-old , and raised in similar conditions as those sampled in 2003 . The purpose was to determine if the association of TMEM154 haplotypes with OPP infection was reproducible in animals sampled seven years later . The USSES ( Idaho ) sheep population was sampled in 2004 and 2008 and used to select cohorts of 309 and 365 mature ewes , respectively . The purpose was to determine if an association of TMEM154 haplotypes with OPP infection was evident in another research flock that was geographically and historically distinct from the Nebraska flock . The USSES sheep population contains Columbia , Rambouillet , and Polypay breeds . The private Polypay sheep flock ( Iowa ) was sampled in 2009 and used to select a cohort of 218 mature ewes . The purpose was to determine if an association of TMEM154 haplotypes with OPP infection was evident in a commercial flock distinct from those in Nebraska and Idaho . This commercial flock was chosen based on its availability . Whole blood samples for serum fractionation and DNA extraction were drawn from the jugular vein into S-Monovette serum Z and EDTA KE 9 ml syringes , respectively ( Sarstedt , Newton , NC , USA ) . Laboratory diagnosis for OPP was performed at the Washington Animal Disease Diagnostic Laboratory ( Pullman , WA , USA ) with a Caprine Arthritis Encephaltitis Virus ( CAEV ) competitive-inhibition ELISA ( cELISA ) . This CAEV cELISA is applicable for the detection of OPPV antibodies in sheep [45] , [48] . Briefly , this assay uses a proprietary monoclonal antibody derived from the fusion of goat splenocytes and mouse myeloma cells ( VMRD , Inc . , Pullman , WA , USA ) . This antibody is conjugated to horeseradish peroxide and is used to compete with serum antibodies for the CAEV antigen bound to the microtiter plate . Additional testing for OPP was performed at USMARC and ADRU with CAEV cELISA kits , according to manufacturer's instruction ( VMRD , Inc . , Pullman , WA , USA ) . Ovine BACs predicted to contain TMEM154 were identified from those mapped to the ovine draft genome sequence http://www . livestockgenomics . csiro . au/sheep/oar2 . 0 . php ) . BACs were isolated from an arrayed 10–12× sheep BAC library ( CHORI-243 , [52] ) , cultured , and the BAC DNA was purified . The BACs were derived from the Texel ram used for the ovine genome sequencing project ( USMARC animal no . 200118011 ) . Pooled samples of the four BACs ( CH243-492L14 , 270 kb; CH243-229A18 , 147 kb; CH243-363J1 , 329 kb; and CH243-426G18 , 279 kb ) were sequenced by synthesis with conditions optimized for 600 bp read lengths , according to manufacturer's instructions ( Roche Applied Sciences , Branford , CT , USA ) . DNA sequences were assembled de novo with Newbler software provided by the manufacturer and the contigs were evaluated and viewed with Consed [53] . Contig assembly made use of information and sequence available for cattle and sheep at the National Center for Biotechnology Information ( NCBI ) and International Sheep Genomics Consortium ( ISGC ) , respectively . A 78 kb region of genomic DNA sequence containing the complete predicted TMEM154 gene was assembled with 70 k reads and 25 Mb of sequence . Four large contigs were manually joined with information derived from ovine mRNA sequences , and the annotated 78 kb sequence was deposited in GenBank ( accession number HM355886 ) . Ovine TMEM154 exons were genotyped by Sanger sequencing of PCR fragments amplified from genomic DNA ( Table S4 ) . DNA extraction and genetic analyses were performed in a manner similarly to that previously described [47] . Briefly , a 1 , 000 bp PCR product containing each exon was sequenced in the 138 matched case-control sheep and 96 rams from a diverse panel of common U . S . sheep breeds ( MARC Sheep Diversity Panel version 2 . 4 ) [47] . After scoring polymorphisms from these 234 sheep in all exons , a second round of nested PCR fragments were designed so that: 1 ) a 700 bp amplicon was fully nested within each previous 1 , 000 bp amplicon , and 2 ) the amplification primers for the 700 bp products did not bind to polymorphic sites discovered from sequencing the 1 , 000 bp on the genome ( Table S4 ) . The combined Sanger sequences from each animal were scored and recorded manually . More than 60 thousand tracefiles and 6 . 9 million genotypes from the present report are publicly available via the internet ( http://cgemm . louisville . edu/USDA/index . html ) . For mRNA transcript analysis , ovine blood ( 3 mL ) was collected ( Tempus Blood RNA tubes , Life Technologies Corporation , Carlsbad , CA , USA ) and stored at −20°C prior to RNA extraction . Whole blood RNA was purified by centrifugation and filtration according to the manufacture's protocol ( Tempus Spin RNA isolation kits , Life Technologies Corporation ) . RNA quantity and quality were determined spectophotometrically ( ND-1000 , NanoDrop Technologies , Inc . , Wilmington , DE , USA; and Agilent 2100 Bioanalyzer ( Agilent Technologies , Inc . , Santa Clara , CA , USA ) . The complete TMEM154 mRNA coding region was amplified by PCR from cDNA ( SuperScript III One-Step RT-PCR System , Platinum Taq High Fidelity , Invitrogen Corporation , Carlsbad , CA , USA ) . The 25 µL reactions contained 1× of the manufacturer's reagent cocktail , 0 . 2 µM each of the sense and antisense primers ( Table S4 ) , 0 . 5 µL SuperScript III RT/Platinum Taq High Fidelity Enzyme Mix , and 30–50 ng of total RNA . Reaction conditions were the following: 1 cycle of cDNA synthesis at 55 °C for 30 minutes followed by pre-denaturation at 94 °C for 2 minutes; 40 cycles of PCR amplification at 94 °C for 15 seconds , 58 °C for 30 seconds , 68 °C for 1 minute; and 1 cycle of final extension at 68 °C for 5 minutes . As a control for DNA contamination and any putative TMEM154 pseudogenes , duplicate sample reactions to those described above were subjected to PCR without preceding cDNA synthesis . Successful amplification of 1 , 012 bp fragments was monitored by gel electrophoresis . Amplicons were not observed in RT-PCR reactions lacking cDNA synthesis . Following an Exonuclease I digestion [54] , TMEM154 RT-PCR amplicons were sequenced with dye-terminator chemistry and separated by capillary electrophoresis ( ABI 3730 , PE Applied Biosystems , Foster City , CA , USA ) . The oligonucleotide primers for PCR and sequencing are listed in Table S4 . Sequences were analyzed for polymorphisms and scored manually with Phred and Phrap [55] , [56] , Polyphred ( version 6 . 10 ) [57] and Consed software [53] . An artiodactyl species panel of DNAs similar to that described previously [58] was sequenced to provide an estimate of the likely ancestral state of the polymorphic ovine TMEM154 codons . This panel is composed primarily of species from the Pecoran clade , whose common ancestor dates to about 30 million years ago [59] . Oligonucleotide primers derived from ovine TMEM154 genomic sequences were used in PCR assays to amplify exons 1 and 2 and PCR products for both exons were produced for the following species: Wyoming bighorn sheep ( Ovis canadensis , n = 7 ) , American plains bison ( Bison bison , n = 7 ) , Alaskan caribou ( Rangifer tarandus , n = 7 ) Wyoming elk ( Cervus canadensis nelsoni , n = 7 ) , Texas exotic red deer ( Cervus elaphus , n = 2 ) , Texas exotic fallow deer ( Cervus dama , n = 1 ) , gaur ( Bos gaurus , n = 2 ) , domestic goat ( Capra hircus , n = 4 ) , Arkansas exotic water buffalo ( Bubalus bubalis , n = l ) , Wyoming mule deer ( Odocoileus hemionus , n = 7 ) , Wyoming white-tailed deer ( Odocoileus virginianus , n = 5 ) , Wyoming mountain goat ( Oreamnos americanus n = 8 ) , and Alaskan and Wyoming moose ( Alces alces , n = 8 ) , for a total of 66 non-ovine artiodactyl individuals . To ensure that amplified DNA sequences were not derived from spurious ovine DNA , only those sequences with distinctive species-associated nucleotide differences were included in the analysis . Proteins encoded by Pecoran species were more than 95% identical to that encoded by ovine TMEM154 haplotype 3 .
|
Ovine lentivirus targets the host immune system and causes persistent retroviral infections affecting millions of sheep worldwide . In primates , lentivirus resistance is attributed to mutant virus coreceptors that are not expressed . In sheep , some animals are resistant to lentivirus infection despite repeated exposure; however , the mechanism of resistance is unknown . We designed a genome-wide association study to test whether sheep might have genetic variation that protects against lentivirus infection . Our results showed that variation in an ovine gene ( TMEM154 ) was associated with infection . Sheep with the ancestral type of this gene were nearly three times more likely to become infected than those with mutant forms . We also discovered two mutant forms predicted to abolish the protein's function . Although the biological function of TMEM154 is unknown , our results indicate that it plays an important role in lentivirus infection in sheep . Producing sheep with the least susceptible form of TMEM154 may help eradicate the ovine disease caused by lentivirus .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"public",
"health",
"and",
"epidemiology",
"veterinary",
"diseases",
"veterinary",
"epidemiology",
"epidemiology",
"genetics",
"molecular",
"genetics",
"biology",
"genetics",
"and",
"genomics",
"veterinary",
"science"
] |
2012
|
Reduced Lentivirus Susceptibility in Sheep with TMEM154 Mutations
|
Vibrio vulnificus causes highly lethal bacterial infections in which the Multifunctional Autoprocessing Repeats-in-Toxins ( MARTX ) toxin product of the rtxA1 gene is a key virulence factor . MARTX toxins are secreted proteins up to 5208 amino acids in size . Conserved MARTX N- and C-terminal repeat regions work in concert to form pores in eukaryotic cell membranes , through which the toxin’s central region of modular effector domains is translocated . Upon inositol hexakisphosphate-induced activation of the of the MARTX cysteine protease domain ( CPD ) in the eukaryotic cytosol , effector domains are released from the holotoxin by autoproteolytic activity . We previously reported that the native MARTX toxin effector domain repertoire is dispensable for epithelial cellular necrosis in vitro , but essential for cell rounding and apoptosis prior to necrotic cell death . Here we use an intragastric mouse model to demonstrate that the effector domain region is required for bacterial virulence during intragastric infection . The MARTX effector domain region is essential for bacterial dissemination from the intestine , but dissemination occurs in the absence of overt intestinal tissue pathology . We employ an in vitro model of V . vulnificus interaction with polarized colonic epithelial cells to show that the MARTX effector domain region induces rapid intestinal barrier dysfunction and increased paracellular permeability prior to onset of cell lysis . Together , these results negate the inherent assumption that observations of necrosis in vitro directly predict bacterial virulence , and indicate a paradigm shift in our conceptual understanding of MARTX toxin function during intestinal infection . Results implicate the MARTX effector domain region in mediating early bacterial dissemination from the intestine to distal organs–a key step in V . vulnificus foodborne pathogenesis–even before onset of overt intestinal pathology .
The United States Department of Agriculture ( USDA ) reports that 1 in 6 Americans fall ill with a foodborne infection every year , resulting in more than 3000 fatalities and an estimated $15 billion in economic burden due to combined medical costs , productivity loss , and death . Of the cited foodborne illnesses , the most economically costly on a per-case basis are those caused by the Gram-negative , marine bacterial pathogen Vibrio vulnificus [1] . Since V . vulnificus can be concentrated out of the water by filter-feeders , such as oysters , susceptible individuals often contract gut infections from consumption of contaminated seafood [2 , 3] . Within the intestine , the microbe is adept at transversing the intestinal barrier to spread systemically , at which point gastrointestinal symptoms are supplanted by bacterial sepsis [4] These intestinal infections can be severe resulting in hospitalization rates of more than 90% and fatality rates that exceed 50% [1 , 5–10] . Although V . vulnificus cases remain rare , annual V . vulnificus case counts have increased over the past 15 years in the United States , and the infection is prevalent in endemic countries , including Japan , Taiwan , and South Korea [5 , 11 , 12] . Moreover , there is a growing risk of pathogen exposure in historically non-endemic areas as highlighted by case reports from countries including Sweden , Germany , France , and Denmark [13 , 14] . Thus , the rise in V . vulnificus infection incidence , while in part attributable to improved reporting , correlates with increasing sea surface temperature and incidence of this disease is expected to increase with climate change [5 , 15] . Pathogenic bacteria are known for secreting virulence factors that promote infection of host organisms . In the case of V . vulnificus , the rtxA1 gene encodes a secreted multifunctional autoprocessing repeats-in-toxins ( MARTX ) toxin that is the dominant virulence factor for intestinal infection [16] . The 5206 a . a . MARTX toxin produced by V . vulnificus strain CMCP6 contains long regions of highly conserved tandem amino acid repeats at the N- and C-termini . These regions are required for toxin secretion , and formation of the MARTX toxin pore , which has been estimated at 1 . 63nm [17 , 18] . Between the repeat regions are situated a cysteine protease domain ( CPD ) and a region of modularly organized effector domains [19] . The pore is thought to serve as the route for translocation of the central toxin region containing the CPD and effector domains across the eukaryotic plasma membrane from the extracellular space to the cytosol of target cells . Stimulated by the eukaryotic-specific molecule inositol hexakisphosphate , the CPD is activated to cleave after leucine residues located between effector domains , thereby releasing bacterial effector proteins into the cytosol of the targeted eukaryotic cell [20] . The V . vulnificus MARTX toxin is associated with numerous cytopathic and cytotoxic functions in vitro . Specifically , the toxin plays a role in lysis of numerous eukyarotic cell types [21 , 22] , cytoskeletal dysfunction as illustrated by epithelial cell rounding [18] , inflammasome activation [23] , inhibition of phagocytosis [24–26] , and induction of apoptosis [27] . Moreover , the discrete effector domains present in the inner region of MARTX toxins are being biochemically characterized to discern their enzymatic functions . Effector domains from various MARTX toxins are now known to inhibit Rho GTPases [28 , 29] , cleave Ras and Rap1 [30 , 31] , inhibit autophagy [32] , induce apoptosis [33] , and crosslink actin [34–36] Yet it has been challenging to discern functional relationships between discrete portions of the MARTX toxin and V . vulnificus virulence . We have previously demonstrated that the MARTX toxin effector domain region is dispensable for toxin secretion from the bacterium and toxin translocation to target cells , indicating that secretion and translocation functions are conferred by the conserved repeat regions [18] . In fact , the toxin repeat regions and the CPD comprise a sufficiently robust platform to deliver a heterologous beta-lactamase ( Bla ) in place of the native MARTX toxin effector repertoire [18 , 28] . Notably , the MARTX toxin pore is also sufficient to induce necrotic cytotoxicity of HeLa cells in vitro , even in the absence of effector domains [17 , 18 , 22] . In light of the intestinal tissue necrosis observed during V . vulnificus intragastric ( i . g . ) infection , it might follow that the MARTX toxin repeat region pore , sufficient for cellular necrosis in vitro , would be sufficient for MARTX-associated intestinal damage , bacterial dissemination , and death as linked to colonization of the liver and spleen [9 , 16 , 37–39] . Yet , the effector domain region is retained across V . vulnificus isolates and is required for in vitro cell rounding and apoptosis induced by the bacterium [18 , 33] . Therefore , we hypothesized that the MARTX toxin effector domain region is important for i . g . pathogenesis of V . vulnificus . In this study , we characterize the functional roles of MARTX toxin regions in V . vulnificus pathogenesis and virulence during i . g . infection . We find that sum of the MARTX toxin effector domains induce rapid loss of transepithelial resistance and increased paracellular permeability in vitro prior to the induction of intestinal epithelial cell lysis . Moreover , MARTX effector domains are required for bacterial dissemination from the intestines to the liver and spleen very early during i . g . infection . Surprisingly , overt intestinal epithelial necrosis is not a prerequisite to bacterial dissemination . Together these data indicate that the focus of the MARTX toxin as a virulence factor should de-emphasize its lytic function and emphasize toxic mechanisms of delivered effector domains that confer V . vulnificus virulence potential .
The MARTX cytotoxin product of rtxA1 is established as a potent virulence factor of V . vulnficus infection [16 , 25] . The VvhA cytolysin/hemolysin product of the vvhA gene is also cytolytic and functions additively to virulence , yet the MARTX toxin exerts the dominant effect [16] . To isolate MARTX-associated phenotypes from those of VvhA , we utilized a variant of V . vulnificus CMCP6 with an internal out-of-frame deletion in vvhA [18] . In this strain , the wild-type version of the MARTX toxin ( RtxA1 , Fig 1A ) is produced . The rtxA1 locus has then been manipulated in this ΔvvhA background . The ΔvvhA rtxA1::bla strain was previously generated by replacing the gene sequence encoding for the MARTX toxin effector domains with an in-frame sequence encoding TEM1 Bla to produce the modified RtxA1::Bla toxin ( Fig 1A ) . This effector-free strain expresses and secretes the modified Bla MARTX toxin , as previously characterized [18] . An rtxA1 null strain was also previously generated via internal deletion , resulting in the ΔvvhA ΔrtxA1 strain [18] . To verify that CMCP6 genetic manipulation did not compromise rtxA1 gene expression , quantitative real-time PCR ( qRT-PCR ) was employed . Gene expression from mutant strains was compared to that of the CMCP6rif parent strain following growth in Luria Burtani ( LB ) broth . At the mRNA level , there was no detectable difference of gene expression of rtxA1 between the ΔvvhA and ΔvvhA rtxA1::bla strains ( Fig 1B ) . Therefore , rtxA1 mRNA expression is not affected by the presence or absence of the gene sequence encoding effector domains . To examine the role of MARTX toxin effector domains in V . vulnificus virulence , mice were inoculated i . g . with either the parental ΔvvhA strain or the effector-free ΔvvhA rtxA1::bla strain ( Fig 1C , low dose ) . The ΔvvhA strain caused fatality in 100% of mice by 24 hours post infection ( hr p . i . ) . In contrast , the rtxA1::bla was lethal to just 7% of infected mice ( Fig 1C ) . Thus , the rtxA1::bla strain is significantly attenuated , demonstrating that the MARTX toxin effector domains are instrumental to toxin-associated virulence . To test for an association between in vitro cytolytic activity and virulence , mice were infected at a 100-fold higher dose with either ΔvvhA rtxA1::bla–which causes rtxA1-mediated lysis in vitro–or ΔvvhA ΔrtxA1 –which does not induce lysis in vitro . No significant differences in survival outcomes were observed between the two groups ( Fig 1D ) . Notably , some vvhA- and rtxA1-independent virulence was observed at these high-doses . This could be due to bacterial components , such as LPS , and/or minor virulence factors , such as the VvpE metalloprotease , all of which may influence pathogenesis at very high dose in the absence of dominant virulence factors [40 , 41] . Nonetheless , the ΔvvhA rtxA1::bla strain does not confer additional virulence over the null mutant [18] . Although the effector-free MARTX toxin expressed by ΔvvhA rtxA1::bla is sufficient to cause necrotic cytotoxicity in vitro , the kinetics of this process are delayed in the absence of MARTX toxin effector domains [18] . Therefore , slower cell lysis kinetics could contribute to the observed virulence attenuation of ΔvvhA rtxA1::bla compared to ΔvvhA . If this hypothesis were true , it should then follow that a strain with intermediate MARTX-dependent lysis kinetics would exhibit intermediate virulence . To address this question , a recently described rtxA1 variant with a catalytically-inactive MCF-like effector domain integrated into the rtxA1::bla gene to generate the modified toxin RtxA1::MCF*Bla ( Fig 1A ) [33] was used . In this strain , the presence of the effector stimulates the rate of in vitro cell lysis to a level intermediate to that of ΔvvhA rtxA1::bla and ΔvvhA [33] . However , the ΔvvhA mcf*::bla strain was completely attenuated during i . g . infection and showed no increased virulence above that of the ΔvvhA rtxA1::bla strain ( Fig 1C ) . Despite exhibiting intermediate rtxA1-dependent lysis kinetics in vitro , this strain does not exhibit intermediate virulence phenotype . Therefore , lysis kinetics do not correlate to strain virulence potential . Collectively , these studies indicate that the ability of a given V . vulnificus strain to induce lysis in vitro does not correlate with bacterial virulence for i . g . inoculated mice . Rather , MARTX effector domains delivered by the holotoxin , while not required for cell lysis , are essential for MARTX toxin-associated virulence . It is known that detection of bacteria in the liver and spleen correlates with lethality in mouse models of V . vulnificus infection [38 , 39] . Clinically , outcomes of V . vulnificus human infection are considerably worse once the infection has become septic [6] . Therefore , bacterial dissemination from the initial site of the infection constitutes a key step in V . vulnificus pathogenesis . Moreover , the rtxA1 gene has been linked to the dissemination process [16 , 25] . To examine the role of MARTX toxin effector domains in V . vulnificus intestinal colonization and dissemination , we inoculated mice i . g . with ΔvvhA , ΔvvhA rtxA1::bla , or ΔvvhA ΔrtxA1 at the same dose used to determine relative virulence of these strains . Knowing that initial lethality was observed approximately 8 hr p . i . , we selected 6 hr p . i . to euthanize mice and collect organs such that all mice could be examined prior to death . Isolated organs were homogenized and the resulting homogenate plated to determine the bacterial load per organ . No differences in bacterial recovery from the intestine were detected across mice infected with ΔvvhA , ΔvvhA rtxA1::bla and ΔvvhA ΔrtxA1 ( Fig 2A ) . Interestingly , this indicates that the bacterial load of the intestine at 6 hr p . i . is independent not only of the MARTX toxin effector domains , but also of the toxin in its entirety . Therefore , neither the MARTX toxin effector domain region nor the repeat regions influence intestinal bacterial load 6 hr p . i . Moreover , early intestinal colonization does not account for virulence differences among the inoculated strains . Bacterial load in the liver and spleen of the same mice were also quantified at 6 hr p . i . ( Fig 2B and 2C ) . While the ΔvvhA strain is detected in these organs at 2–5 x 104 CFU/organ , neither the ΔvvhA rtxA1::bla nor the ΔvvhA ΔrtxA1 strain disseminated to this level . In fact , no colonies grew from the majority of liver and spleen homogenates from ΔvvhA rtxA1::bla or ΔvvhA ΔrtxA1-infected mice even when plated at a detection limit of 10 CFU/organ . The inability to detect V . vulnificus in distal organs at a meaningful level in the absence of the MARTX toxin effector domain region reinforces the integral role of the MARTX toxin in V . vulnificus pathogenesis . In addition , this result indicates that lytic action conferred by MARTX toxin repeat regions is not sufficient to facilitate bacterial transit to distal organs . Rather , an intact MARTX toxin effector domain region is required for bacteria to reach , and survive in , distal organs including the liver and spleen . Bacterial dissemination from the intestine to distal organs necessitates bacterial transit across the protective intestinal epithelial barrier from intestinal lumen to bloodstream , and resistance to immune defense mechanisms , particularly phagocytosis , encountered at each of these locations . While the MARTX toxin is known to facilitate bacterial resistance to phagocytosis [25 , 26 , 28] , equivalent bacterial loads were observed in the intestines of mice at 6 hr p . i . Therefore , at this time point , at least in the intestine , V . vulnificus are not differentially susceptible to immune clearance dependent upon rtxA1 , though this occurs at later time points [16] . In previous studies , significant intestinal epithelial tissue damage has been observed in both mice and humans following i . g . V . vulnificus infection [9 , 16 , 37] . This damage has subsequently been attributed to the additive function of secreted exotoxins VvhA and MARTX [16] . To test the relationship between ΔvvhA rtxA1::bla lysis in vitro , intestinal epithelial damage during infection , and bacterial dissemination , mice were inoculated i . g . with the ΔvvhA , ΔvvhA rtxA1::bla , or ΔvvhA ΔrtxA1 strains . At 6 hr p . i . , the entire intestine was collected and analyzed with hemotoxalin and eosin ( H/E ) staining . Surprisingly , no significant pathology was observed in any intestinal tissues at 6 hr p . i . The epithelia remain intact in 4/4 ΔΔvvhA-infected mice , 3/4 ΔvvhA rtxA1::bla-infected mice , and 3/3 ΔvvhA ΔΔrtxA1-infected mice ( Fig 3 ) . In 1/4 ΔvvhA rtxA1::bla-infected mice , the intestine showed observable bacterial staining in the lumen and the small intestinal epithelium showed some damage at villous tips ( Fig 3 ) . This outlier sample demonstrates that rare events resulting in rapid bacterial expansion can lead to epithelial damage and supports a model in which bacterial outgrowth in the intestine at later infection time points ( 8–12 hr p . i . ) indeed causes pathological changes , as previously shown [16] . However , this event was not representative . Therefore , we conclude that V . vulnificus does not induce extensive intestinal epithelial necrosis by 6 hr p . i . and the overt damage previously observed during later infection or in neutropenic mice is not solely responsible for dissemination [16 , 37] . In fact , dissemination is initiated by the MARTX toxin effector domain region prior to the onset of fulminant intestinal tissue damage in vivo . The absence of fulminant intestinal tissue pathology during i . g . V . vulnificus infection suggested that the bacterium might instead be inducing localized apoptosis of intestinal epithelial cells rather than lysis . Indeed , it has been reported that V . vulnificus can induce mitochondrial-mediated apoptosis in an rtxA1-dependent manner [27] . Apoptotic cells can be observed by an experienced pathologist in H&E stains tissue sections at high magnification . However , no major differences among tissue samples were observed in the examined H&E sections . Nonetheless , this result was confirmed using an apoptosis-specific stain . The same embedded intestinal tissue used for H/E staining were also stained for cleaved caspase-3 . Sporadic apoptosis is observed in the intestinal epithelial layer as indicated by positive staining for cleaved caspase-3 ( S3 Fig ) . However , no differences were observed dependent upon rtxA1 . Therefore , differences in apoptosis do not account for differences in bacterial dissemination among V . vulnificus strains and do not provide a mechanism by which epithelial breach is occurring . Previous studies have linked V . vulnificus dissemination from the intestine to the liver and spleen with lethal infection outcomes [38 , 39] . To assess organ damage , the spleen and liver were isolated from mice infected with ΔvvhA , ΔvvhA rtxA1::bla , or ΔvvhA ΔrtxA1 at 6 hr p . i . and assessed by histological examination . Despite the presence of >104 CFU ΔvvhA in the spleen and liver of infected mice ( Fig 2 ) , these organs retain their normal morphology at 6 hr p . i . and tissues do not exhibit overt pathology ( Fig 4 ) . Therefore , disseminated V . vulnificus ΔvvhA do not induce direct spleen and liver organ damage at 6 hr p . i , despite the rapid onset of animal mortality beginning at 8 hr p . i . Absent overt pathology in the form of either tissue necrosis or apoptosis , we reasoned that the changes induced by V . vulnificus MARTX to facilitate bacterial dissemination must be subtler and not dependent on gross effects observable by pathology . Previous work that predominantly focused V . vulnificus research on cell lysis extensively uses lactate dehydrogenase ( LDH ) release from nonconfluent , adherent epithelial cell monolayers as an in vitro system for MARTX-dependent cytotoxicity [16–18 , 22 , 24] . For a more relevant three-dimensional culture model to monitor intestinal epithelial barrier breach events in vitro , the interaction between V . vulnificus and polarized confluent T84 colonic epithelial cells was studied ( Fig 5A ) . T84 cells were cultured as monolayers in transwells to transepithelial resistance ( TER ) of ≥1000 Ω*cm2 . Log phase bacteria were added to the apical surface to mimic luminal i . g . bacterial exposure . Based upon preliminary dosing experiments ( S1 Fig ) and bacterial ratios relevant to in vivo infection ( S1 Text ) , subsequent experiments were conducted at multiplicity of infection ( MOI ) 2 . 5 or 0 . 25 . Initial experiments examined the individual contributions of vvhA and rtxA1 to characterize interactions between the epithelial monolayer and bacterial exotoxins in vitro ( Fig 5B ) . When applied to the apical surface of T84 monolayers , CMCP6rif rapidly induced intestinal barrier dysfunction as demonstrated by a 50% decrease from initial TER after 60 minutes and more than 80% drop over 210 minutes . T84 monolayers exposed to V . vulnificus ΔvvhA exhibited a drop in TER identical to monolayers exposed to CMCP6rif . In contrast , T84 monolayers exposed to ΔrtxA1 retained initial TER to nearly 90 minutes , exhibiting a significant delay and attenuation of TER disruption relative to CMCP6rif and ΔvvhA-exposed monolayers . When monolayers were exposed to ΔvvhA ΔrtxA1 , TER was maintained over the course of the experiment to the endpoint at 210 minutes and , over multiple experiments , often resulted in a rise in TER . Together these results indicate that both the VvhA hemolysin and MARTX toxin are sufficient for TER drop over a 3-hour time scale , but only the MARTX toxin accounts for the rapid loss of TER initiated shortly after addition of bacteria . Further , the contribution of VvhA that occurs after 120 min is not essential , additive or synergistic ( Fig 5B ) . Having established the importance of the MARTX holotoxin in TER loss , the role of MARTX toxin regions was examined ( Fig 5C ) . T84 monolayers were co-incubated with V . vulnificus ΔvvhA , ΔvvhA rtxA1::bla , or ΔvvhA ΔrtxA1 . Monolayers exposed to ΔvvhA dropped to 50% resistance in 60–75 minutes of co-incubation , as previously observed . However , the integrity of the monolayers exposed to ΔvvhA rtxA1::bla was maintained to approximately 100 minutes . Therefore , compared to the ΔvvhA strain , the ΔvvhA rtxA1::bla strain is significantly delayed in its ability to disrupt T84 monolayer integrity . T84 monolayers exposed to ΔvvhA rtxA1::bla for more than 100 minutes gradually exhibited a drop in TER , though TER loss induced by the ΔvvhA rtxA1::bla strain was attenuated throughout the duration of the experiment compared to that induced by ΔvvhA . Monolayers exposed to ΔvvhA rtxA1::bla strain were also distinct from the ΔvvhA ΔrtxA1-exposed monolayers , as the double mutant did not experience any disruption of TER . Thus , MARTX effector domains are necessary for rapid-onset monolayer disruption , but pore formation by the MARTX repeat regions is sufficient for T84 disruption past 100 minutes of co-incubation . To test whether different toxin translocation and lysis kinetics previously observed in HeLa cells influenced bacterial interactions with T84 monolayers , the ΔvvhA mcf*::bla strain was again employed ( Fig 5D ) [33] . With the goal of observing even subtle differences between the strains , the applied dose was reduced 10-fold to MOI 0 . 25 . However , even at this lower dose there was no detectable difference between the ΔvvhA mcf*::bla and ΔvvhA rtxA1::bla-exposed T84 monolayers . Both strains remained drastically attenuated in the ability to disrupt T84 TER compared to ΔvvhA . The ΔvvhA strain induced 50% TER loss by approximately 100 minutes–a delay relative to the 10-fold higher dose , but an increase relative to either of the strains lacking active MARTX effector domain regions . At the same 100-minute time point , monolayers exposed to ΔvvhA mcf*::bla or ΔvvhA rtxA1::bla retained approximately 90% initial resistance . Surprisingly , this loss of TER by 60 mins was not due to overt actin depolymerization ( Fig 6 ) . Indeed , despite dropping to 45% initial resistance , the ΔvvhA-exposed monolayer exhibits actin morphology akin to the PBS control monolayers . Specifically , all samples retained characteristic honeycomb-like actin morphology in the x-y plane and columnar cellular structure in the monolayer z-plane . Therefore , the rapid initial loss of TER upon addition of V . vulnificus to T84 monolayers is not an artifact of differences in lysis kinetics of the wild-type RtxA1 toxin compared to RtxA1::Bla and is not due to extensive loss of actin structure . A common mechanism by which the integrity of a monolayer can be compromised is disruption of cell-cell junctions that results in increased paracellular permeability . To examine mechanisms of MARTX-induced barrier dysfunction , paracellular permeability to small molecules was examined with use of a fluorescently tagged , 3-kD dextran . The molecule cannot pass through cells and is likewise typically excluded from passing between cells by tight junctions . However , when intercellular junctions are disrupted , the dextran molecule gains passage between cells . Following application of dextran to the apical transwell chamber , PBS or bacteria were added to monolayers and basal media was sampled for dextran transit over time . Appreciably greater amounts of dextran were sampled from the basal media of ΔvvhA-exposed monolayers compared to monolayers exposed to ΔvvhA rtxA1::bla , or ΔvvhA ΔrtxA1 ( Fig 5E ) . Neither PBS mock-exposed nor ΔvvhA ΔrtxA1-exposed monolayers allowed more than 0 . 09 μg of dextran to cross from the apical to the basal compartment of the T84 transwells over the tested 180 minute timecourse . Similarly , a maximum of 0 . 15 μg of dextran was detected in ΔvvhA rtxA1::bla-exposed monolayers . However , the ΔvvhA-exposed monolayers allowed passage of 0 . 45 +/- 0 . 01 μg of fluorescent dextran ( Fig 5E ) . In ΔvvhA-exposed transwells , an upward trend in basal dextran levels began between 45 and 60 minutes . By 90 minutes , significant differences in dextran transit between the ΔvvhA monolayers and all other monolayers were evident ( Fig 5E ) . In contrast , only basal levels were observed in ΔvvhA rtxA1::bla and ΔvvhA ΔrtxA1-exposed monolayers prior to 135 minutes . From 135–180 minutes , increased levels of dextran were sampled from the basal media of monolayers exposed to ΔvvhA rtxA1::bla compared to ΔvvhA ΔrtxA1 or PBS-exposed monolayers , though these amounts were still considerably less than the basal dextran sampled from ΔvvhA-exposed monolayers . These in vitro monolayer experiments revealed that both TER disruption and dextran flux exhibit biphasic characteristics . Specifically , the ΔvvhA strain caused rapid decay of TER , while ΔrtxA1 ( Fig 5B ) and ΔvvhA rtxA1::bla ( Fig 5C ) caused only late onset TER disruption . The ΔvvhA strain rapidly induced paracellular permeability , but the rate of dextran flux between 135 and 180 minutes ( 900 ng/hr/cm2 ) was significantly greater than the flux rate from 60–135 minutes ( 350 ng/hr/cm2 ) ( Fig 5F ) . In ΔvvhA rtxA1::bla-exposed monolayers , dextran flux rates were significantly greater in the 135–180 minute time frame ( 300 ng/hr/cm2 ) compared to 60–135 minutes ( 100 ng/hr/cm2 ) . Overall , it was concluded that the MARTX toxin , when interacting with a polarized columnar monolayer , exerts two MARTX-dependent mechanisms of barrier disruption: one shortly after addition of bacteria and mediated by the effector domains and one later after addition of bacteria linked to the repeat regions . Since the repeat regions are known to be sufficient for lysis of unpolarized cells , the contributions of V . vulnificus MARTX toxin regions to T84 cell lysis was explored as a mechanism for late vs early onset loss of TER . Monolayers were exposed to ΔvvhA , ΔvvhA rtxA1::bla , or ΔvvhA ΔrtxA1 . Sixty minutes following bacterial application , resistance of the ΔvvhA-exposed monolayers dropped to 50% initial while ΔvvhA rtxA1::bla , or ΔvvhA ΔrtxA1 monolayers retained TER >90% , as observed in previous independent experiments ( Fig 5 ) . Monolayer cell lysis was measured by sampling LDH release to media in both the apical and basal transwell chambers and expressed relative to LDH release from monolayers treated with 0 . 1% Triton X-100 . LDH release to the basal transwell chamber was never detected ( S2 Fig ) so monolayer lysis was quantified using media sampled from the apical chamber . The absence of LDH in the basolateral chamber may also explain the absence of bacteria in the same compartment , if the lower edge of the monolayer and appendages filling the filter pores do not provide a clear path across a partially lysed monolayer ( S2 Fig ) . Despite the large drop in TER in ΔvvhA-treated monolayers by 60 minutes , ΔvvhA cell lysis at 60 minutes averaged less than 10% and was no greater than the low levels likewise observed in PBS , ΔvvhA rtxA1::bla , or ΔvvhA ΔrtxA1-exposed monolayers ( Fig 5G , 60 minutes ) . Therefore , rapid MARTX-dependent loss of TER by ΔvvhA occurs independent of cell lysis . By contrast , at 180 minutes following bacterial application , ΔvvhA-exposed monolayers retained just 10% initial resistance , ΔvvhA rtxA1::bla-exposed monolayers exhibited 50% initial resistance , and ΔvvhA ΔrtxA1-exposed monolayer resistance had increased to 180% initial . At this late timepoint , ΔvvhA ΔrtxA1 lysis levels remained <20% while both ΔvvhA and ΔvvhA rtxA1::bla induced lysis exceeding 80% of cells in the monolayer . Therefore , T84 cell lysis occurred after prolonged monolayer exposure and corresponds to increased dextran flux and TER loss in both ΔvvhA and ΔvvhA rtxA1::bla-exposed monolayers . Now knowing the important role of the MARTX effector domain repertoire in its entirety , the role of individual domains within the region was explored . Ideally , this experiment would have been conducted using strains with point mutations in the active sites of each MARTX effector domain so as to generate catalytically inactive effector domains in the context of the MARTX holotoxin . However , the catalytic residues of numerous domains have not as yet been identified , and some catalytic point mutants are known to exert intermediate effects when target-binding activity is retained [29 , 33 , 42 , 43] . Therefore , a library of strains was generated in the ΔvvhA background in which each strain harbors an in-frame deletion in the rtxA1 coding region to eliminate a single effector domain from the otherwise functional toxin ( Fig 7A ) . All Δeffector strains were first validated as inducing release of LDH from polarized T84 monolayers at 180 minutes ( Fig 7B ) , demonstrating that the modified rtxA1 genes expressed toxin that retained the cell lysis activity linked to the repeat regions . In addition , Δduf1 , Δrid , Δabh , and Δmcf were confirmed to induce loss of detectable Ras from HeLa cells when co-incubated with the target cells for 1 hour at MOI = 100 ( Fig 7C ) . Ras was detected in cells incubated with the Δrrsp strain as expected due to loss of the RRSP-dependent cleavage of Ras [30] . Cytopathic epithelial cell rounding in response to V . vulnificus has also been previously attributed to the MARTX effector domain repertoire [18] . As shown in Fig 7D , a strain expressing the wild-type MARTX toxin induced rounding of more than 90% of HeLa cells in 120 minutes ( ΔvvhA , Fig 7D ) while a ΔvvhAΔrtxA1 strain does not induce any effect above the background . The Δduf1 , Δabh , Δmcf , and Δrrsp strains similarly induced cell rounding , independently validating these strains are producing functional MARTX toxins . As expected , due to the disruption of its Rho-inactivated domain linked to cytoskeleton disassembly , epithelial cell rounding was significantly reduced in HeLa cells incubated with the Δrid strain ( Fig 7D ) . Thus , all Δeffector strains were validated as retaining the ability to lyse cells and retain or lose activity specifically linked to at least two of the effector domains . These validated strains were next used to define the effector domains essential for rapid loss of TER in polarized T84 monolayers . For this experiment , polarized T84 monolayers were exposed apically to Δeffector strains and TER was measured at 60 minutes , when early loss of TER is most pronounced , and at 120 minutes , when effector-independent barrier dysfunction begins ( Fig 5 ) . Interestingly , all of the Δeffector strains induced loss of TER at levels comparable to the strain expressing the entire effector domain repertoire ( Fig 7E and 7F ) . Therefore , no single MARTX effector is essential for induction of epithelial barrier dysfunction in vitro . In addition , this result indicates that despite its requirement for induction of cell rounding in nonconfluent HeLa epithelial cells , Rho inactivation induced by RID does not alone account for TER loss at 60 minutes . This suggests that MARTX effectors function redundantly or synergistically with regard to epithelial barrier disruption .
The important role of the MARTX toxin product of the rtxA1 gene as a virulence factor during V . vulnificus infection has now been appreciated for nearly a decade [44] . This potent cytotoxin has been linked to induction of multiple forms of cell death in vitro [17 , 22 , 27] . During i . g . infection , fulminant intestinal tissue damage has been observed . Moreover , bacterial dissemination and sepsis are phenotypes intimately linked to lethal infection outcomes [6 , 37–39 , 45] . Together , these data have led to a prevailing model that massive toxin-mediated destruction of the intestinal epithelial barrier is the key mechanism by which V . vulnificus exits the intestine culminating in lethal sepsis . The data presented here indicate a paradigm shift in our conceptual understanding of early V . vulnificus transmigration of the intestinal epithelial barrier . Indeed , the cytopathic activities of the MARTX toxin initiate dissemination earlier than previously appreciated and the early mechanisms involved are far subtler in nature . In studying the contribution of MARTX toxin regions to holotoxin function , we previously identified the MARTX repeat regions as sufficient for formation of pores in the eukaryotic plasma membrane resulting in epithelial cellular necrosis [18] . These bacteria that produce a MARTX toxin able to lyse cells , but absent the effector domain region , were indistinguishable during intestinal infection of mice from bacteria that did not produce the toxin at all ( Fig 1 ) . Thus , it is the entire toxin , inclusive of the effector domains , that confers MARTX-mediated virulence during i . g . V . vulnificus infection . Notably , genetic and biochemical functional characterization of MARTX toxin complexity previously lent hypothetical support to this result . Yet , these data represent the first direct experimental evidence that non-lytic functions previously attributed to the MARTX toxin ( such as cytoskeleton disassembly , induction of apoptosis , inhibition of autophagy , and modulation of stress signaling [27 , 30 , 32 , 33 , 46 , 47] ) must play an important role in pathogenesis . To further understand how the toxin contributes to virulence , a physiologically relevant in vitro model system using polarized T84 cells was newly optimized for use with V . vulnificus . These studies demonstrated that the MARTX toxin induces biphasic intestinal epithelial dysfunction in the form of early-onset increases in paracellular permeability , followed by late-onset cell lysis . The effector domain region of the MARTX toxin is responsible for the rapid loss of barrier function . Yet , studies with bacteria that express toxins lacking individual effector domains reveal that no single effector domain is essential for this rapid intestinal epithelial disruption . Rather there must be an additive or synergistic function from multiple effector domains in loss of epithelial barrier function , a mechanism by which effector domains may contribute to bacterial translocation from the intestine to the liver and spleen following i . g . infection in mice , resulting in V . vulnificus sepsis . The dissociation between MARTX-associated virulence , epithelial barrier breach , and lysis is further reflected in results from study of animal histopathology . The MARTX holotoxin does not induce overt intestinal tissue damage or excess apoptosis during early infection . This finding is consistent with the T84 experiments where rapid loss of TER was not linked to dramatic loss of cytoskeleton structure . Paracellular permeability increase in the absence of major changes to cytoskeletal morphology suggests that more delicate modulation of cytoskeletal dynamics , such as those at intercellular junctions , is occurring . By contrast , the second phase of TER loss was linked to cell damage resulting in release of LDH . This second event likely accounts for the outlier pathology sample that showed observable bacterial staining in the intestinal lumen as well as epithelial damage at villous tips suggest that rare stochastic events leading to early-stage bacterial outgrowth facilitate epithelial damage . However , the absence of significant necrotic or apoptotic phenotypes in all ΔvvhA-infected mice indicates that the MARTX effector domain region is necessary for dissemination not because it causes or induces overt tissue damage , but because it induces early bacterial transit from the intestine via other mechanisms . It was recently observed that the V . vulnificus elastase product of vvpE modulates paracellular permeability by altering tight junction protein dynamics [41] . However here no significant dysfunction is observed in monolayers exposed to ΔvvhA ΔrtxA1 , indicating that vvpE does not increase permeability or intestinal damage in this context . The apparent discrepancy is likely due to differences both in study methods and bacterial MOI . vvpE was shown to increase intestinal permeability in vivo when mice were inoculated i . g . with a high dose of 1 . 1x109 CFU . However , the in vitro MOI used in the present study corresponds to intestinal bacterial loads following inoculation with 5x106 CFU , a nearly 200-fold lower dose . Interestingly , observed rtxA1 and vvhA-independent virulence at high-dose ( Fig 1D ) does indicate a role for either bacterial components , such as LPS , or for minor virulence factors , in the absence of the two major cytotoxins . These combined data suggest that vvpE may play a functional role at the intestinal epithelial barrier , and in virulence , at high bacterial load . A consideration now is how the bacteria are moving across the barrier in vivo . Modulation of paracellular permeability by the MARTX effector domain region may act to directly facilitate paracellular bacterial transit between epithelial enterocytes for subsequent transport to the lymphatics or blood stream in the intestinal lamina propria . Further , the same mechanisms by which the MARTX effector domain repertoire dysregulates enterocytes may also dysregulate other specialized cell populations in the intestinal epithelial monolayer . M cells , responsible for luminal antigen sampling , are present in the epithelial layer that covers lymphoid nodules and Peyer’s patches . The so-called “weak point of the intestinal epithelial barrier” [48] , M cells are known to provide a route for transepithelial migration of viable bacteria , including V . cholerae , from the intestinal lumen to the underlying Peyer’s patches [49 , 50] . While M cell luminal sampling is integral to proper antigen presentation and immune responsiveness , pathogens such as Salmonella , Shigella , and Yersinia exploit the properties of M cells to access the mucosa and spread systemically [50 , 51] . Goblet cells have been implicated in transcytosis of Listeria and thus represent another putative route by which V . vulnificus is breaching the intestinal barrier [52] , potentially facilitating systemic bacterial spread . An important caveat of this study is that events in the intestine are focused only during the early phase of infection , at 6 hr p . i . , in the mouse . At this time point , the number of bacteria in the gut does not yet depend upon rtxA1 , and toxin-mediated epithelial destruction has not yet occurred . Nonetheless , previous studies have demonstrated that these rtxA1-mediated phenomena certainly occur at later time points [16 , 37] . This indicates that when known effects on rtxA1-dependent immune clearance [24 , 25 , 28] are not yet impacting intestinal bacterial load , strains producing the MARTX holotoxin are already detectable in the spleen and liver . It is postulated that within the bloodstream and at tissue sites , rtxA1 and its various regions may also play a critical role in resistance to immune clearance promoting bacterial outgrowth at these sites . In human cases where V . vulnificus infection progresses to septic shock and multi-organ failure , this rise in bacterial loads in the liver and spleen may predict dissemination-associated fatality . However , direct damage to the spleen or liver by the bacteria may not be essential for death , given that mice exhibit no organ damage at 6 hr p . i . yet begin to die by 8 hr p . i . Overall , the results presented in this study support a model in which V . vulnificus bacteria expressing the MARTX holotoxin rapidly induce intestinal epithelial dysfunction in the form of increased paracellular permeability and transmigration ( Fig 8 ) . These early steps are sufficient to facilitate bacterial dissemination and associated virulence potential . Our evidence suggests that initial translocation of bacteria out of the intestine is mediated by the MARTX effector domains and occurs in the absence of overt tissue damage in the intestine , spleen and liver tissues . Subsequent bacterial outgrowth–or a sporatic event resulting in higher bacterial burden–then leads to intestinal tissue necrosis in vivo ( Fig 8 ) . However , in the absence of MARTX effector domain functions , the early breach of the barrier and early arrival of bacteria at distal organs does not occur , resulting in dramatically reduced virulence potential .
V . vulnificus strains ΔvvhA , ΔvvhA rtxA1::bla , ΔvvhA mcf::bla , and ΔvvhA ΔrtxA1 were generated from the Korean clinical isolate V . vulnificus CMCP6rif as described previously [18 , 33] . Escherichia coli strains DH5 αλpir , SM10λpir , and S17λpir [53 , 54] were used for new strain construction . Bacteria were routinely grown in Luria-Burtani ( LB ) broth ( 10 g tryptone , 5 g yeast extract , 5 g NaCl ) containing 50 μg/mL rifampin or 10 μg/ml chloramphenicol as needed . For all experiments , V . vulnificus was streaked from frozen glycerol stocks onto LB plates . The following day , single colonies were grown in 2 mL LB broth overnight at 30°C and then subcultured 1:100 into LB without antibiotic and grown to mid-log phase . Cultures were pelleted and resuspended in sterile phosphate buffered saline ( PBS , 10mM sodium phosphate , 140 mM NaCl , pH 7 . 4 ) to indicated concentrations based on optical density ( A600 ) . The CMCP6 MARTX toxin CPD processing sites have not been precisely mapped although the boundaries of each effector domain have been defined based on extensive sequence alignment [19] . To ensure that processing of neighboring effectors was not negatively impacted by the deletion , the predicted processing sites were preserved along with 15% of the effector domain itself . The designed deletions correspond to the following nucleotides based on the CMCP6 sequence of Kim et al . , [57] ( National Center for Biotechnology Information Reference #NC_004460 . 2 ) : Δduf1 ( Δ5890–6699 ) ; Δrid ( Δ6814–8688 ) ; Δabh ( Δ8794–9279 ) ; Δmcf ( Δ9490–10 , 722 ) ; and Δrrsp ( Δ10 , 753–12 , 252 ) . Fragments corresponding to regions upstream and downstream of the desired deletion were either commercially synthesized ( Integrated DNA Technologies , Coralville , IA ) or amplified from the CMCP6 genome . The two fragments corresponding to each strain were assembled into digested pDS132 [55] either using Gibson Assembly according to the manufacturer’s protocols ( New England Biolabs , Ipswich , MA ) or by standard ligation using T4 DNA ligase . The resulting plasmids were confirmed by sequencing and transformed to SM10λpir [53] , and S17λpir [54] . The Δeffector deletion plasmids were transferred to V . vulnificus ΔvvhA by conjugation followed by selection for double homologous recombination using sucrose counterselection to isolate recombinants as previously described [56] . Deletions in the rtxA1 gene were confirmed by amplification of DNA across the deletion junction . Strains were validated by assessment of LDH release as described below , by western blotting for Ras proteolysis using the pan-Ras RAS10 ( EMD Millipore , 05–516 , 1:500 ) and tubulin ( Sigma-Alrich , T6074 , ( 1:10 , 000 ) monoclonal antibodies as previously described [30] , and by assessment and quantification of HeLa cell rounding as previously described [18] . Bacteria were grown as described above . RNA was extracted using Qiagen RNeasy Kit ( Qiagen , 74104 ) and RNA Protect Bacteria Reagent ( Qiagen , 76506 ) according to the manufacturer’s instructions . Isolated RNA was quantified using a Nano-Drop spectrophotometer . RNA was DNase treated using Turbo DNA Free Kit ( Life Tech , AM1907 ) . RNA was reversed transcribed using random hexamers ( Roche , 11034731001 ) and Superscript III Reverse Transcriptase ( Life Tech 18080–093 ) in the presence of RNasin ( Promega N2611 ) . qRT-PCR was carried out using iQ SYBR Green Supermix ( BioRad 170–8880 ) and the BioRad iQ5 Multicolor RealTime PCR Detection System . Efficiency testing established primer efficiency of 83% and 87% for rtxA1 and 16s rRNA primer pairs , respectively . Primers used were: qRT-RTXF ( 5’AATACCGCTCTTCACAACC3’ ) ; qRT-RTXR ( 5’GCTTTCTGGGTGCTTACC3’ ) ; qRT-16srRNA_F ( 5’CTTGACATCCAGAGAATCTA3’ ) ; qRT-16srRNA_R ( 5’GACTTAACCCAACATTTCAC 3’ ) Three separate qRT-PCR analyses were performed and data pooled following analysis . The 16s rRNA gene served as internal housekeeping control . Fold change was calculated relative to parental CMCP6rif . This study was carried out in strict accordance with the recommendations in the United States Public Health Service ( USPHS ) regulations and applicable federal and local laws . The protocol ( Protocol No . IS00000905 ) was approved by the Northwestern University Institutional Animal Care and Use Committee ( IACUC ) as detailed in methods . All efforts were made to minimize suffering . Female ICR mice were obtained from Charles River at age 32–38 days . Mice were anesthetized via intraperitoneal injection with 100 μL of anaesthetic cocktail containing 60–70 μg/kg ketamine and 12–14 μg/kg xylazine in PBS . Mice were inoculated i . g . using a 1-cm animal feeding needle attached to a 1-mL syringe . Mice were administered 50 μL of 8 . 5% aqueous sodium bicarbonate , followed immediately by 50 μL of bacterial culture containing the CFU indicated for a given experiment . Mice were monitored every 2 hours for the first 28 hours of the experiment and subsequently every 4–8 hours for a total experimental duration of 48 hr p . i . For low dose survival experiments , mice were inoculated with V . vulnificus strains ΔvvhA ( n = 17 ) , ΔvvhA rtxA1::bla ( n = 14 ) and ΔvvhA mcf*::bla ( n = 5 ) For high dose survival experiments , mice were inoculated with ΔvvhA rtxA1::bla ( n = 12 ) and ΔvvhA ΔrtxA1 ( n = 11 ) . For bacterial recovery from organs , 5–6 mice per group were inoculated and then euthanized 6 hr p . i . The whole intestine ( less the cecum ) was excised and homogenized in 5 mL PBS . The liver and spleen were excised and each homogenized in 1 mL PBS . CFU/organ was calculated by plating serially diluted homogenates to LB agar containing rifampin to select for V . vulnificus . For histopathology , 3–4 mice per group were inoculated and euthanized 6 hr p . i . The liver and spleen were dissected . A 1-cm sample was isolated from the proximal end of each segment of the small and large intestine ( duodenum , jejunum , ileum , colon ) for cross-sectional sampling . The remaining portions of each segment were opened along the longitudinal axis , rolled from proximal to distal , and sectioned to obtain samples in a “swiss roll” orientation . After 24–48 hours fixation in 10%-buffered formalin , all tissues were paraffin-embedded , processed , and stained with H&E . For immunohistochemistry , 4 μm of the same embedded tissues were sectioned , mounted on slides , and stained for apoptotic marker cleaved caspase-3 using the CP229C antibody from Biocare Medical , Concord , CA . All pathology sides were viewed and scored by N . T . B . blinded to treatment groups . T84 cells obtained from American Type Culture Collection ( #CCL-248 ) were routinely grown in T84 media ( 1:1 DMEM/F12 Nutrient Mix ( Gibco 11320–033 ) ) supplemented with 10% fetal bovine serum ( FBS ) and 1% penicillin-streptomycin ) to no more than 30 passages . For cell polarization , Costar Transwell Permeable Supports ( 6 . 5mm insert , 24-well plate , 3 . 0 μm polycarbonate membrane , Reference #3415 ) were coated with collagen and dried in a laminar flow hood overnight . Transwells were incubated with T84 media for 1 hour prior to the addition of 106 T84 cells to the apical chamber of the transwell . Media was changed every 2–3 days for 10–14 days until monolayers reached ≥1000Ω/cm2 [58] as measured using an EVOM ( World Precision Instruments ) . A minimum of three monolayers were prepared per assay condition . One hour prior to bacterial co-incubation , monolayers were washed twice with warm Hanks Balanced Salt Solution ( HBSS , Sigma-Aldrich ) and media was replaced with 1:1 phenol-red free T84 media without FBS or antibiotic . Ten μL of PBS or the appropriate concentration of bacteria were applied drop-wise to the apical media . Monolayers were maintained at 37°C using a plate warmer . TER was measured in 15–20 minute intervals or as indicated in legends . For confocal imaging , following 60 minutes of bacterial co-incubation , cells were fixed in 4% paraformaldehyde for 20 minutes . Monolayers were permeabalized using 0 . 1% TritonX-100 . Actin was stained using AlexaFluor 488 phalloidin ( ThermoFisher A12379 ) and nuclei were stained with 4′-6′-diamidino-2-phenylindole ( DAPI , Life D1306 ) , each according to manufacturer’s recommendations . Entire monolayers affixed to membrane were excised , mounted in Pro-Long Gold Antifade ( Life Technologies , P36930 ) under a cover slip , and imaged using a Nikon A1R Spectral microscope . For dextran flux studies , 200 μg fluorescein dextran ( 3-kD , ThermoFisher D3305 ) was added to the apical chamber immediately following application of PBS or bacteria to the apical chamber of transwells . Dextran transit across the monolayer was measured by sampling 20 μL media from the basal transwell chamber , after which the extracted volume was replaced with 20 μL fresh media . Sample fluorescence was measured using a Tecan Safire 2 fluorescence plate reader ( excitation: 494 nm , emission: 521 nm ) and amount of dextran ( in pg ) was calculated against a standard curve , accounting for volume differences due to sampling . Flux rates were reported as μg dextran/hr/cm2 from the slope of the plotted linear curve . For LDH release assays , 100 μL of media was extracted from either the apical ( Fig 5 ) or basal ( S2 Fig ) chambers of PBS , bacterial , or Triton X-100 incubated monolayers . One μL 100 mg/ml gentamicin was added and samples were centrifuged at 15 , 000xg for 1 minute . 50 μL of the resulting supernatant was transferred to a 96-well culture plate . LDH activity was measured using the Promega CytoTox Non-Radioactive Cytotoxicity Assay kit according to manufacturer’s instructions . The apical and basal samples were processed separately with data reported adjusting for volume . Statistical analyses were performed as indicated in figure legends using GraphPad Prism 6 . 0 software .
|
The bacterium Vibrio vulnificus causes highly lethal infections in susceptible individuals exposed to contaminated water or seafood . V . vulnificus produces a MARTX toxin , which plays a significant role in bacterial virulence . The large MARTX toxin has numerous functional regions that have been characterized in vitro , but have not been functionally assessed during infection . In this study , we use a mouse model to show that a region of the MARTX toxin–the effector domain repertoire–is required for bacterial virulence . This region of the toxin promotes disruption of the intestinal epithelial barrier and movement of bacteria from the initial site of infection to the liver and spleen . Surprisingly , bacterial movement from the intestine occurs even in the absence of overt intestinal tissue damage , indicating that more subtle actions of the effector domains are responsible for promoting systemic bacterial spread . Together , our results advance the field of MARTX toxin research by functionally characterizing the toxin during infection . Results indicate that cytolytic effects on epithelial cells observed in vitro do not directly correspond to activities that are critical to bacterial pathogenesis in vivo and should encourage the study of diverse MARTX toxin functions .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"toxins",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"spleen",
"pathogens",
"vibrio",
"cell",
"processes",
"microbiology",
"toxicology",
"toxic",
"agents",
"physiological",
"processes",
"vibrio",
"vulnificus",
"bacteria",
"bacterial",
"pathogens",
"digestive",
"system",
"tissue",
"repair",
"medical",
"microbiology",
"microbial",
"pathogens",
"glucans",
"gastrointestinal",
"tract",
"lysis",
"(medicine)",
"biochemistry",
"polysaccharides",
"dextran",
"anatomy",
"cell",
"biology",
"virulence",
"factors",
"physiology",
"apoptosis",
"biology",
"and",
"life",
"sciences",
"glycobiology",
"organisms"
] |
2017
|
The Effector Domain Region of the Vibrio vulnificus MARTX Toxin Confers Biphasic Epithelial Barrier Disruption and Is Essential for Systemic Spread from the Intestine
|
Spontaneous sub-cellular calcium release events ( SCRE ) are conjectured to promote rapid arrhythmias associated with conditions such as heart failure and atrial fibrillation: they can underlie the emergence of spontaneous action potentials in single cells which can lead to arrhythmogenic triggers in tissue . The multi-scale mechanisms of the development of SCRE into arrhythmia triggers , and their dynamic interaction with the tissue substrate , remain elusive; rigorous and simultaneous study of dynamics from the nanometre to the centimetre scale is a major challenge . The aim of this study was to develop a computational approach to overcome this challenge and study potential bi-directional coupling between sub-cellular and tissue-scale arrhythmia phenomena . A framework comprising a hierarchy of computational models was developed , which includes detailed single-cell models describing spatio-temporal calcium dynamics in 3D , efficient non-spatial cell models , and both idealised and realistic tissue models . A phenomenological approach was implemented to reproduce SCRE morphology and variability in the efficient cell models , comprising the definition of analytical Spontaneous Release Functions ( SRF ) whose parameters may be randomly sampled from appropriate distributions in order to match either the 3D cell models or experimental data . Pro-arrhythmogenic pacing protocols were applied to initiate re-entry and promote calcium overload , leading to the emergence of SCRE . The SRF accurately reproduced the dynamics of SCRE and its dependence on environment variables under multiple different conditions . Sustained re-entrant excitation promoted calcium overload , and led to the emergence of focal excitations after termination . A purely functional mechanism of re-entry and focal activity localisation was demonstrated , related to the unexcited spiral wave core . In conclusion , a novel approach has been developed to dynamically model SCRE at the tissue scale , which facilitates novel , detailed multi-scale mechanistic analysis . It was revealed that complex re-entrant excitation patterns and SCRE may be bi-directionally coupled , promoting novel mechanisms of arrhythmia perpetuation .
Cardiovascular disease is one of the major healthcare problems faced by the developed world , with increasing prevalence associated with aging populations [1–3] . Improved understanding of the mechanisms underlying cardiac arrhythmias , a major component of cardiovascular diseases’ impact on morbidity and mortality , is vital to the effort to improve both quality and duration of life . Rapid arrhythmias , such as tachycardia and fibrillation , are associated with disordered and incomplete contraction , reducing cardiac output and potentially leading to sudden cardiac death . The underlying rapid and irregular electrical activation of cardiac tissue may be mediated by abnormal spontaneous pacing ( focal ectopic activity; “arrhythmia triggers” ) , self-perpetuating re-entrant excitation ( “arrhythmia substrate” ) , or a complex interplay between both mechanisms ( trigger-substrate interactions ) [4 , 5] . Management of these arrhythmias is typically challenging , often requiring invasive procedures such as implanted defibrillators or catheter ablation; even these interventions have limited success rates [6 , 7] . Understanding of the mechanisms underlying the genesis , perpetuation and recurrence of rapid arrhythmias will ultimately lead to the development of improved treatment strategies . Whereas multiple experimental and simulation studies have investigated the tissue substrate for the emergence and perpetuation of re-entrant excitation [8 , 9] , the multi-scale mechanisms of arrhythmia triggers , and their dynamic interaction with the tissue substrate , remain elusive [10] . Malfunction of the intracellular calcium ( Ca2+ ) handling system has been implicated in the development of rapid arrhythmias , linking sub-cellular spontaneous Ca2+ release events ( SCRE ) to pro-arrhythmic triggers in single cell [11–15] . However , translation of these cellular data to assess the mechanisms and importance of SCRE in tissue-scale arrhythmia is a significant challenge [10] , namely because of the dependence of SCRE on stochastic fluctuations at the sub-cellular , nanometre scale which must propagate to the whole-organ , centimetre scale . These complex multi-scale mechanisms are discussed in detail in previous reviews [11 , 12 , 16] , and are briefly summarised below . A feedback mechanism emerging from the structure-function relationships underlying cardiac cellular excitation-contraction coupling presents a potential pro-arrhythmogenic pathway for the propagation of random state-transitions at the microscopic scale ( sub-cellular ) to the macroscopic ( whole-organ ) : ( i ) spontaneous opening of the ryanodine receptors ( RyRs ) , which control release of Ca2+ from the Sarcoplasmic Reticulum ( SR ) into the intracellular space , can trigger spontaneous Ca2+ sparks in restricted nanodomains called dyads; ( ii ) the sub-cellular spatial distribution of dyads presents a substrate for the propagation of Ca2+ sparks as a whole-cell event; ( iii ) this spontaneous whole-cell Ca2+ transient can lead to cellular delayed-after-depolarisations ( DADs ) and triggered action potentials ( TA ) through activation of the sodium-calcium exchanger ( NCX ) ; ( iv ) TA may propagate in tissue as ectopic focal excitation , potentially leading to the initiation of transient or sustained arrhythmia . The dynamics of Ca2+ homeostasis , also strongly dependent on the refilling of the SR through the SR-Ca2+ pump ( SERCA ) , is critical for the emergence of SCRE: cardiac cells typically exhibit a minimum SR-Ca2+ load threshold above which whole-cell SCRE occur . The cytosolic Ca2+ concentration and kinetics of the RyRs ( which may be more prone to spontaneous release in disease [14 , 17] ) also strongly affect SCRE dynamics and vulnerability , which can be reflected in adaptations of this SR-Ca2+ load dependence ( wherein lower thresholds correspond to larger vulnerability to SCRE ) . Computational modelling provides a viable approach for detailed simultaneous multi-scale evaluation of cardiac arrhythmia mechanisms . Nevertheless , simulating SCRE in tissue-scale models is non-trivial due to the contrasting computational requirements of models at these different scales: single-cell models capable of reproducing SCRE are computationally intensive and unsuitable for simulation of the thousands or millions of coupled cells comprising tissue models appropriate for studying arrhythmia mechanisms . New approaches must therefore be developed to realise this goal . Only a few previous studies have attempted such modelling , for example investigating: ( i ) the minimum tissue substrate for the emergence of focal excitations resulting from non-stochastic DADs [18]; ( ii ) the emergence of focal excitation from stochastic SCRE [19–23] and its potential interaction with extracellular matrix remodelling [24]; ( iii ) SCRE as a mechanism for both triggered activity and conduction block [25]; and ( iv ) the potential complex considerations for pharmacological action on both triggers and substrate [22] . In a previous study , a phenomenological approach was introduced to model the synchronisation of cellular triggers in tissue [22] . The present study aims to refine this approach , and extend it to allow dynamic modelling of trigger-substrate interactions . The novel approach , comprising a hierarchy of computational models , is applied to: ( i ) directly translate behaviour observed in detailed cell models to the tissue-scale and ( ii ) study the dynamic interactions of SCRE and re-entrant excitation . The approach is also generalised to allow fully controllable investigations and integrate limited experimental data .
The developed framework consists of a hierarchy of computational models ( Fig 1 ) : To induce prominent full-cell release events through different cellular conditions , representative ( but non-specific ) models were included for isoprenaline ( ISO , sympathetic response which enhances CICR ) and two types of pro-SCRE general disease remodelling mimicking features observed in conditions such as AF and HF ( e . g . , [14 , 38] ) : ( i ) SERCA was up-regulated and NCX was down-regulated ( RSERCA/NCX ) ; ( ii ) the SR-Ca2+ threshold for release was lowered through increased inter-CRU coupling ( RCRU-CRU ) . RSERCA/NCX also involved remodelling of the ion currents , including a reduction of IK1 , to provide a different AP environment coupled to the Ca2+ handling model . Details are provided in S1 Text ( Model Description ) . The purpose of these models , which are not biophysically detailed or representative of specific regulation or diseased conditions , was to induce pro-arrhythmic behaviour which is not observed under control conditions , and test and demonstrate the ability of the 0D approximations to accurately capture variable cellular conditions and translate these to the tissue-scale . A partial Ca2+-clamp ladder protocol ( Fig 2 ) was implemented to both derive and validate the SRF in the 0D model: At each step ( total duration 2s ) , the intracellular- and SR-Ca2+ concentrations were initially clamped to specified values , with the SR-Ca2+ clamped concentration incrementing for each successive step ( Fig 2Aii ) . When spontaneous intracellular release occurs , the Ca2+ concentrations were allowed to dynamically evolve ( clamp constraint removed ) . The membrane potential was allowed to evolve during this protocol , and the conductances of INa and ICaL were set to zero to prevent excitation and resulting interruption of the SCRE by CICR . This protocol illustrates the variety of SCRE and their underlying spatio-temporal dynamics which must be captured in the 0D model ( Fig 2B; S1 Video ) . The phenomenological approach involved the development of SRF which describe whole-cell RyR dynamics associated with SCRE observed in the 3D cell model , based on an extension of previous work [21 , 22] . These functions are waveforms which approximate the range of morphologies for the time series of the whole-cell NRyR_O/NRyR observed during SCRE ( Fig 3A ) . The concept of these SRF was introduced in a previous work [22] in the context of a statically defined model ( corresponding to the Direct Control approach , below ) . This required the parameters describing the SRF to be defined by the user for specific simulations and therefore is not suitable for investigating natural cellular responses to different conditions or coupled dynamic simulations . Here , the functions themselves are refined , a method to derive the function parameters dynamically from model variables is introduced , leading to behaviour directly in-line with the 3D cell models , and the approaches are generalised to be fully controllable and suitable for parameterisation to experimental data . For completeness and context , the SRF are described here in full .
The 0D model implementing the Dynamic Fit SRF model was first validated by comparison of whole-cell SCRE under Ca2+ clamp conditions with a second set of simulations ( i . e . , not those on which the model was derived ) of the 3D cell model , for the control and remodelling conditions ( S2 Text ( Validation and Results ) ) . These simulations highlight the strong agreement for waveform morphology and its variation , the distributions and their summary properties , and the differences between the control and remodelled conditions . Secondly , the 0D and 3D models were compared across the range of ionic models ( see Methods: Action Potential and Tissue Models ) and remodelling/ISO conditions ( see Methods: Pro-arrhythmic conditions ) . Examples of dynamics emerging in conditions close to the SR-Ca2+ threshold ( Fig 6Aa ) and high above it ( Fig 6Ab ) show good agreement between the 3D and 0D models ( compare Fig 6Ai-ii with 6Aiv-v ) , importantly capturing the key features and differences between the conditions . The case for low SR-Ca2+ ( Fig 6Aa ) was intentionally selected as one of those in which the match was the poorest , in order to fully illustrate the quality of the approximation in an upfront manner; even here , the match is reasonable , and the main features of the behaviour are preserved ( DADs with some TA; wide initiation time distribution ) . The distributions of initiation time and the probability of triggered APs across all conditions tested which resulted in notable SCRE also show good agreement ( Fig 6B ) , confirming the ability of the 0D models to dynamically reproduce the SCRE of the 3D cell models and capture cellular and condition dependent behaviour . Note that the differences between the cell-types and conditions is not of primary interest , but , rather , it is the match between 3D and 0D models , including the reproduction of these differences , which is of interest , with a key feature being the probability of TA or DADs . The potential for the models to simulate SCRE at the tissue and organ scale was illustrated through pacing tissue models under an equivalent rapid pacing–quiescent protocol to the single cells ( see Methods: Simulation Protocols ) . Under the right conditions ( i . e . , significant SR-Ca2+ loading and thus large-scale release events ) a triggered action potential emerged from a single focus and propagated throughout the tissue , observed in multiple tissue models ( S2 Text ( Validation and Results ) ) . Multi-focal activations were also observed . The different SCRE dynamics observed across the range of 3D single cells models under pro-SCRE conditions ( Fig 6B ) was accentuated in tissue , wherein focal excitations only emerged under conditions which resulted in significant TA ( Fig 7 ) . Note also the important impact of reduced IK1 , present in the atrial cell models and RSERCA_NCX remodelling conditions , on allowing the emergence of tissue focal excitation . Evaluation of the focal activation emerging in the 2D sheet illustrates the mechanism by which these small scale cellular events result in full tissue excitation and the dual role of electrotonic coupling ( Fig 8; S2 and S3 Videos ) : SCRE occurring in independent cells initially causes only a small depolarisation of the cell membrane , repressed by electrotonic coupling ( Fig 8C , purple trace ) ; however , this depolarisation is also electrotonically spread to surrounding cells ( Fig 8B ) ; due to this depolarised resting potential , SCRE occurring later can act to further depolarise the surrounding tissue ( Fig 8C , blue trace ) ; once local tissue is sufficiently depolarised , the probability of DADs manifesting as TA significantly increases and appropriately-timed firing cells can much more easily initiate a focal excitation ( Fig 8C , orange trace ) . In homogeneous tissue , using a baseline General Dynamic SRF model parameter set ( see Methods: Simulation Protocols ) , the probability curve for ectopic activity was substantially steeper than that for the emergence of TA from DADs in single cell ( Fig 9Ai ) . Reduction of the density of IK1 shifted both the single cell and tissue TA probability curves to the left ( lower SR-Ca2+ ) , towards the curve describing DADs ( Fig 9Aii ) ; however , the steep relationship in tissue remains . With control IK1 density , both heterogeneity models ( involving either 10% or 20% of cells randomly allocated to each of the additional four General Dynamic SRF models with different SR-Ca2+ thresholds ) shifted the SR-Ca2+ relationship to the right ( higher SR-Ca2+ concentrations ) , despite the presence of cells more susceptible to SCRE . With reduced IK1 density , the small heterogeneity model still shifted the relationship to the right , but the large heterogeneity model now shifted it to the left ( Fig 9B ) . In general , the introduction of heterogeneity reduced the steepness of the probability curve at low probabilities ( Fig 9B ) . Heterogeneity models with the less vulnerable cells removed ( higher SR-Ca2+ threshold cells reassigned to baseline ) produced the same overall features: three cases still shifted to the right despite the presence of more pro-SCRE cells and absence of less vulnerable cells . SCRE leading to non-TA inducing DADs was demonstrated to produce unidirectional conduction block during applied pacing in a homogeneous 2D model of the human ventricular wall ( Fig 10A; S4 and S5 Videos ) under simulated sodium channelopathy conditions ( see Methods: Simulation protocols ) , wherein the SCRE induced DADs inactivate the sodium channel and result in non-uniform propagation following a stimulus uniformly applied to one edge ( Fig 10A ) . The potential for SCRE mediated focal excitation to result in unidirectional conduction block due to regional APD heterogeneity was demonstrated in the 2D transmural model of the human ventricular wall . Focal excitations originating in a small temporal window could exhibit a conduction block with the still refractory M-cell region; this was not observed in the homogeneous tissue ( Fig 10B; S6 Video ) . Note that the site of conduction block , unlike in simulations with an applied focal stimulus , is not necessarily clearly at the boundary between the EPI and M cells: in the illustrated case , the focus emerges from the far side of the boundary . Sustained re-entry combined with pro-arrhythmic conditions resulted in significant SR-Ca2+ loading , which promoted the onset of SCRE mediated focal excitations following termination . In the illustrated 2D simulations using the minimal RA AP model , re-entry sustained for ~13 s before self-terminating , loading the SR-Ca2+ to ~ 1 . 25 mM ( Fig 11A and 11B; S7–S10 Videos ) . By implementing a General Dynamic SRF model with the SR-Ca2+ threshold set below , but close to , the maximal SR-Ca2+ observed ( CaSRthreshold = 1 . 125 mM ) , a single , delayed focal excitation was observed following termination ( Fig 11A and 11Bii ) . With a lower threshold ( CaSRthreshold = 1 . 0 mM ) multiple and rapid focal excitations were observed which perpetuated the arrhythmic dynamics for the duration of the simulation ( Fig 11A and 11Biii ) . In general , where the tissue was sufficiently vulnerable for focal excitation to emerge , its origin was localised to the region of the scroll wave core on its final excitation pathway ( Fig 11B; S8 and S9 Videos ) ; focal excitations emerging away from the scroll wave core were also observed closer to the threshold ( Fig 11C ) . Localisation of focal excitation to the scroll wave core is determined by the dynamics of re-entrant excitation: the scroll wave core remains unexcited associated with each re-entrant cycle ( Fig 12A ) , which leads to an island of high SR-Ca2+ in this region ( Fig 12Ac ) . Due to this large SR-Ca2+ and longest recovery time , this region thus presents the earliest SCRE which may manifest as TA . This relationship is clearly illustrated by comparing the spatial distribution of recovery time ( time since last AP which induced CICR ) at the moment before focal excitation occurs with the associated focal activation map: the site of activation can correspond exactly ( Fig 12Bi , iv ) or approximately ( Fig 12Bii-iii , v ) to the location of longest recovery time , in both single- ( Fig 12Bi-iv ) and double- scroll wave ( Fig 12Bv ) simulations; the earlier the focal excitation occurs relative to the final re-entrant scroll wave , the stronger the correlation between location of longest recovery time and activation source ( Fig 12Bvi ) . Under pro-SCRE conditions comprising tight synchronisation ( small width of ti distribution at short-coupled intervals ) and short-duration , large-amplitude releases , focal activity can occur at a rate comparable to re-entry; such pacing can transiently drive the excitation , potentially eventually terminating arrhythmia as the SR-Ca2+ depletes . Furthermore , the wavefront could degenerate back into a sustained re-entrant excitation ( Fig 12C; S11 Video ) . The latter is promoted by the asymmetric conduction patterns emerging from rapid focal excitation interacting with the tail of the previous scroll wave or asymmetric focal excitation . During this mechanism switching , the underlying driving mechanism ( focal or re-entrant ) was not clear in AP traces from randomly selected cells in the 2D tissue , and could only be determined by spatio-temporally high-resolution analysis of the excitation patterns ( Fig 12C ) . SR-Ca2+ loading as a result of re-entry led to different emergent behaviour than that induced by regular rapid pacing with matched average SR-Ca2+ peak concentrations ( Fig 13 ) . The impact of the islands of high SR-Ca2+ with longer waiting times is clear , as focal excitations in re-entry simulations could emerge almost immediately following termination whereas those following regular rapid pacing exhibited a minimum delay ( approximately equal to the sum of SR refilling time and earliest ti; Fig 13Ai ) . However , when SR-Ca2+ is close to the threshold for release , focal excitations were observed following rapid pacing but not following re-entry . This was due to the heterogeneous SR-Ca2+ loading ( and subsequent timing of SCRE ) associated with re-entry , whereas regular rapid pacing led to almost homogeneous SR-Ca2+ loading ( offset by activation time; Fig 13Aii ) . Focal excitations following regular rapid pacing were always plane wave , not occurring early enough ( in electrically homogeneous media ) to interact with the tail of the previous excitation; those following re-entry , however , exhibited asymmetric excitation patterns when occurring sufficiently early to interact with the tail of the previous excitation ( Fig 13B ) . Note that this is due to both the earlier excitation time and the asymmetry of the previous re-entrant excitation compared to the plane wave of regular pacing .
An increased incidence of spontaneous calcium release events ( SCRE ) is frequently observed in isolated cardiomyocytes from the diseased myocardium [11–15] , and their pro-arrhythmic coupling to the membrane potential through activating inward NCX current has led to the hypothesis that these dysfunctional Ca2+ handling phenomena play a role in the initiation and dynamics of complex arrhythmia conduction patterns . However , investigating these multi-scale mechanisms presents a significant challenge , both for experimental and simulation approaches , and thus the precise mechanisms and potential importance of these events manifesting as tissue-scale arrhythmia have yet to be fully described . In this study , a multi-scale computational approach was developed to simulate the dynamics of stochastic SCRE in organ-scale models of cardiac excitation . The computational framework comprises a hierarchy of models ( Fig 1 ) encompassing the microscopic- ( 3D cell model ) , mesoscopic- ( 0D cell model ) and macroscopic-scales ( tissue models ) . Spontaneous Release Functions ( SRF; Fig 3 ) were used to reproduce the morphology of SCRE in the 0D cell and tissue models ( Figs 6 and 7 ) , directly translating the single-cell modelling to the tissue-scale . The approaches were then applied to study arrhythmia mechanisms . Firstly , the dual role of electrotonic coupling in the emergence of SCRE as a full tissue focal excitation was illustrated ( Fig 8 ) . Secondly , the role of cellular variability in SCRE dynamics on the SR-Ca2+—focal excitation relationship was investigated ( Fig 9 ) . Thirdly , two mechanisms of SCRE mediated conduction block were demonstrated ( Fig 10 ) , illustrating mechanisms by which re-entry may be initiated . Finally , the long-term interactions with re-entry were investigated ( Figs 11 and 12 ) , demonstrating that sustained re-entrant excitation can load the SR-Ca2+ and promote SCRE mediated focal excitation , and revealing a purely functional mechanism of localisation . This mechanism also resulted in re-entrant pre-pacing leading to earlier and more asymmetric focal excitations than observed in matched regular pacing scenarios ( Fig 13 ) . The present study presents a novel approach to efficiently simulate stochastic sub-cellular Ca2+ release dynamics through the implementation of simple and controllable analytical wave functions—the SRF . The use of these simple functions allowed direct control over SCRE dynamics as well as models fit to the behaviour of 3D single cells , facilitating both general mechanistic analysis and translation of single-cell investigations to the organ-scale . The phenomenological approach considered whole-cell behaviour only , and was not based on capturing underlying detail . Sub-cellular heterogeneity of factors such as t-tubule density can influence the vulnerability to and dynamics of whole-cell SCRE [36 , 41] . This may be particularly relevant for atrial cells ( which , for large mammals , exhibit variable size and t-tubule density , and for small mammals in general exhibit a lack of t-tubules [42 , 43] ) and disease conditions ( in which t-tubule density is often observed to decrease [44–46] ) . This top-down approach facilitates direct incorporation of the dynamics of these heterogeneous conditions , which may be essential for accurately modelling focal excitation in the atria and disease states . In general , adaptations to the SR-Ca2+ threshold for SCRE , frequently observed to shift towards lower concentrations in pathophysiological conditions such as heart failure [47 , 48] ( possibly as a result of sub-cellular structural remodelling [36] ) , can be easily accounted for using this approach . Analysis of the impact of cellular variability in the dynamics of SCRE on the probability of the emergence of an ectopic beat highlighted the complex considerations surrounding the amplitude and timing of SCRE , and its modulation by IK1: introducing a proportion of cells more vulnerable to SCRE in combination with reduced IK1 led to a negative shift the SR-Ca2+ relationship ( i . e . , towards lower SR-Ca2+ values ) as well as the emergence of lower probability events , contrasting to the almost step-function response in homogeneous tissue; the same SCRE variability without reduced IK1 right-shifted the probability curve , despite the presence of larger amplitude SCRE compared to the homogeneous condition . This important role of reduced IK1 , as previously highlighted [19 , 20 , 25 , 49] , is due to a combination of it resulting in ( i ) resting potentials closer to the INa activation threshold and ( ii ) smaller repolarising current opposing the DAD , and may be of particular relevance to arrhythmia associated with heart failure remodelling , in which IK1 is generally reduced . Preliminary analysis was also performed to assess the criticality of parameter distributions for the emergence of focal excitation ( S2 Text ( Validation and Results ) –section 3 , Figure S4 ) . These simulations revealed clear constraints determining thresholds between distributions which did and did not result in focal activity , and the modulation of these constraints by IK1; however , they also revealed the challenges of predicting emergent behaviour within the threshold region . The mechanisms by which a suitably timed focal excitation can result in conduction block and the onset of self-perpetuating re-entrant excitation have been extensively studied previously ( e . g . [27] ) . The present study also demonstrates a novel feedback mechanism by which re-entrant excitation promotes SCRE: Combining simulations of SCRE with those of sustained re-entrant excitation demonstrated that the resulting SR-Ca2+ loading can promote the emergence of focal excitation following termination of re-entry , perpetuating the arrhythmic conduction patterns . The simulations revealed a purely functional mechanism of localised ectopic and re-entrant excitation , without the requirement for a specifically vulnerable region , determined by the additional relaxation time associated with the unexcited core of a scroll wave . This localization combined with rapid focal activity resulted in excitation patterns which were highly asymmetric and almost indistinguishable from the re-entry , due to conduction block with the tail of the previous re-entrant or focal excitation . Furthermore , these asymmetric conduction patterns could result in a complete re-entrant circuit and the re-initiation of sustained re-entry . This mechanism switching may not clearly present in mapping or ECG measurements , but may have significant implications for pharmacological intervention and provide one explanation for variable and limited success . In the first instance , the Dynamic Fit SRF models are naturally generalised within the range of modifications to the AP model which do not include any differences in the underlying Ca2+ handling system , i . e . , remodelling and regulation of the sodium and potassium currents and ICaL: these currents do not directly affect the probability of spontaneous Ca2+ sparks or their propagation as whole-cell events in the model , and so the influence of their modulation on SCRE is only through their effect on SR-Ca2+ loading and electrotonic load . This can be further generalised if small modifications were made to the Ca2+ handling system in relation to Jup , Jleak and to a lesser extent INCX , which have only a small effect on Ca2+ spark propagation ( in the model ) and do not significantly alter the SR-dependence . This was demonstrated in the present study by the ISO model ( which had no effect on RyR or NCX , but did affect Jup and LTCC open channel availability ) : the SRF did not require an ISO-dependent parameter in order for the 0D model to reproduce the 3D model behaviour under ISO conditions ( Fig 6 ) . Further modifications to the Ca2+ handling system , which do significantly alter the SR-dependence , required the SRF models to be rederived , as was demonstrated by the two remodelling models . Whereas time-consuming , such an approach could be used to reproduce different dynamics emerging from any individual or combination of regulation and/or remodelling conditions . The process and approaches presented in this study can be further generalised to be incorporated into other cell models . Theoretically , any “standard” non-spatial cell model which contains a rigorous model of CICR and includes RyR open state dynamics can directly include the analytical SRF . As an example , the Grandi et al . , 2011 human atrial cell model [50] was selected as this includes one of the most physiological descriptions of RyR dynamics in a non-spatial cell model . Incorporation of the SRF required only a few additional lines of code ( pertaining to the implementation algorithm ) further to the SRF themselves , with the magnitude rescaled to reflect the maximum open-state occupancy observed in that model ( S3 Text ( Generalisation of approaches ) ) . This demonstrates the potential suitability for direct integration with available contemporary , non-spatial AP models , without the requirement to replace the native intracellular Ca2+ handling system or develop a spatial model equivalent . The General Dynamic approach also facilitates parameterisation to experimental data , even with limited information . To demonstrate this functionality , the data provided by Workman et al . , 2012 [34] were selected ( S3 Text ( Generalisation of approaches ) ) , as these data pertain only to measurement of membrane potential ( and not direct measurement of Ca2+ dynamics ) and are therefore not an ideal dataset for parameterisation of SCRE . Even under these limited conditions , it was possible to broadly reproduce the emergence of DADs and TA observed experimentally ( S3 Text ( Generalisation of approaches ) ) . Work from only a few research groups has attempted to simulate stochastic SCRE at the organ-scale [19–25] . These independent studies used alternative phenomenological approaches to overcome the inherent challenges of this multi-scale simulation as presented in this manuscript . Further to providing an independent approach which forms a complementary tool to the previous models , which is of particular importance for theoretical investigation of systems with many unknowns and highly non-linear behaviour , the present approach differs from those previously namely in: ( i ) the motivation to reproduce SCRE in tissue models in-line with that observed in specific ( and variable ) 3D cell models , for direct translation of single cell modelling studies to the tissue-scale; ( ii ) the ability to directly control waveform parameters and relate observed behaviour to these parameters; ( iii ) an approach which readily allows direct incorporation of both limited and detailed experimental data; and ( iv ) the presentation of an open-source computational framework for congruent investigation of SCRE at single cell- and tissue-scales ( S1 Code ) . Only a few computational studies have attempted to dissect the multi-scale mechanisms involved in SCRE mediated arrhythmia . Initially , the minimum tissue substrate for the emergence of focal excitations resulting from non-stochastic EADs and DADs was investigated [18] , followed by demonstration of independent cellular events emerging as a focal excitation [19–25] and SCRE as a mechanism for both triggered activity and conduction block [25] . Potential interaction with extracellular matrix remodelling was demonstrated [24] , as well as the potential non-linear considerations for pharmacological action on both triggers and substrate [22] . Important features observed in these previous studies are in agreement with that of the present study , i . e . in relation to the mechanism of synchronisation overcoming electrotonic load , the steep SR-Ca2+ relationship of ectopic activity in homogeneous tissue , the importance of IK1 in governing the vulnerability to ectopic activity , and the potential role of non-TA inducing DADs to cause conduction abnormalities [19 , 20 , 25 , 49]; this study therefore provides independent validation of these features . The present study also provides novel analyses and mechanistic insight , pertaining to: ( i ) the analysis of SCRE vulnerability variability on the SR-Ca2+-TA relationship , and the quantification of parameter distribution thresholds to predict focal excitation; ( ii ) investigation of the potentially pro-arrhythmic bi-directional coupling between SCRE and re-entry; and ( iii ) demonstration of a mechanism of localisation of these phenomena which can also lead to focal-re-entrant mechanism switching . Due to the substantial components of this paper , full limitations associated with the models and approaches are discussed in detail in the S4 Text ( Limitations ) and the reader is referred there for full description of the applicability of the models in present form; here , key model limitations and those regarding the novel analysis are discussed . The method to derive the SRF in order to reproduce behaviour of the single cell model ( i . e . , not the general implementations ) required a large volume of computationally intensive simulations to be performed ( ~ 5 000 hours of computation time per condition ) , although these do not need to be repeated once the parameters have been derived . Alternative approaches were also considered which have potential advantages . For example , deriving a similar iterative-map approach to that presented in [51] , parameterized to the spatial cell model dynamics , would require significantly less intensive simulations and perhaps provide a more robust underlying dynamic system . A purely mathematical derivation would completely circumvent this computationally intensive requirement , and was considered in the early stages of this research . Such a derivation remains an attractive prospect for a truly rigorous and portable efficient model of SCRE . However , the analytical waveform approach presented also has advantages: based on whole-cell behaviour , parameter sets could be derived to describe underlying spatial models in limitless different conditions including sub-cellular heterogeneity and its variability , which may be significantly more challenging to reproduce with an underlying dynamical system . The General Dynamic implementation also negates the requirement for simulations on which to derive the model , and permits both comprehensive mechanistic analysis and direct parameterization to experimental data . Simulations of SCRE emerging as tissue-scale arrhythmia in general required model parameters and conditions towards the extreme of physiologically observed behaviour ( i . e . , corresponding to highly diseased myocardium ) , and therefore do not provide a complete picture of the role of SCRE across the spectrum of patients presenting arrhythmia . Further analysis and new methodological approaches will be required to practically simulate lower probability events and fully assess the role of SCRE in cardiac arrhythmia . In the wider context , these simulations represent only initial detailed in silico analysis of the impact of SCRE at the organ scale: experimental validation , integration with biophysically detailed species and disease specific models , and parameterisation to experimentally measured SCRE statistics , are essential to translate the approaches and mechanistic insight of the present study to arrhythmia in patients . It is intended that these approaches will be further developed and incorporated with sophisticated biophysically detailed cell models and experimentally validated simulation of cellular SCRE in multiple cardiac conditions , in order to suggest new experiments and contribute to detailed analysis of the role of SCRE in cardiac arrhythmia . Demonstration of the generalisation potential of the approaches is hoped to encourage those interested researchers in the community to integrate the presented framework with their cell models and simulation studies; open-source C++ code of the entire framework and detailed documentation is therefore provided in S1 Code and available on the GitHub repository ( https://github . com/michaelcolman/MSCSF ) . The multi-scale cardiac modelling approaches described in this manuscript and accompanying model code present the possibility to model the impact of stochastic , sub-cellular calcium dynamics on organ scale arrhythmic excitation patterns with congruent detailed cellular and tissue simulations or parameterised to specific experimental datasets . Such approaches revealed multi-scale coupling between SCRE and re-entrant excitation and a purely functional mechanism for their localisation . The mechanistic insight gained from the application of these approaches may help to improve understanding and management of cardiac arrhythmia .
|
A loss of the regular rhythm of the beating heart , called arrhythmia , can inhibit its pumping function and even lead to sudden death . Understanding the processes by which normal rhythm is interrupted presents a major yet critical research problem . One challenge is the inherent multi-scale dependence of the electrical activity of the heart: behaviour at the microscopic scales ( single proteins ) can propagate to the macroscopic ( whole-heart ) . Simultaneously studying phenomena at both of these scales is difficult , if not impossible , to perform experimentally . Developing mathematical models of the heart in order to perform variable and controlled simulations of its electrical activity provides the possibility to undertake such multi-scale analysis . However , new approaches are required to simulate the potential role of “spontaneous calcium release” , which occurs at the sub-cellular scale , in triggering arrhythmia events at the whole-heart scale . This study develops an approach to perform such simulations and applies it to study the long-term interactions of these sub-cellular triggers with the complex electrical activity of the most dangerous arrhythmias–tachycardia and fibrillation . The new mechanistic insight has implications for both diagnosis and treatment of the associated disorders .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"action",
"potentials",
"membrane",
"potential",
"electrophysiology",
"neuroscience",
"simulation",
"and",
"modeling",
"probability",
"distribution",
"mathematics",
"cardiac",
"pacing",
"cellular",
"structures",
"and",
"organelles",
"cardiology",
"research",
"and",
"analysis",
"methods",
"arrhythmia",
"biophysics",
"cell",
"membranes",
"probability",
"theory",
"physics",
"intracellular",
"membranes",
"cell",
"biology",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"neurophysiology",
"biophysical",
"simulations"
] |
2019
|
Arrhythmia mechanisms and spontaneous calcium release: Bi-directional coupling between re-entrant and focal excitation
|
Motor control requires the generation of a precise temporal sequence of control signals sent to the skeletal musculature . We describe an experiment that , for good performance , requires human subjects to plan movements taking into account uncertainty in their movement duration and the increase in that uncertainty with increasing movement duration . We do this by rewarding movements performed within a specified time window , and penalizing slower movements in some conditions and faster movements in others . Our results indicate that subjects compensated for their natural duration-dependent temporal uncertainty as well as an overall increase in temporal uncertainty that was imposed experimentally . Their compensation for temporal uncertainty , both the natural duration-dependent and imposed overall components , was nearly optimal in the sense of maximizing expected gain in the task . The motor system is able to model its temporal uncertainty and compensate for that uncertainty so as to optimize the consequences of movement .
In the execution of any movement , there is always timing uncertainty . This uncertainty has two major consequences . First , it limits performance on any task for which there are costs associated with temporal imprecision . Second , it has implications for how the motor system should plan movements when the costs of temporal imprecision are asymmetric . In hurrying to catch a subway train , for example , the cost of arriving early is usually small compared to the cost of arriving late and missing the train . An optimal movement planner must take into account temporal reward asymmetries in forming movement plans . The complexity of movement planning under risk is further increased because temporal uncertainty in the motor system changes constantly . Two major sources of variation in temporal uncertainty occur over different time courses and have different properties: One is a uniform , global shift in temporal uncertainty possibly due to aging , fatigue , injury or disease [1]–[9] . The second is a linear increase in the standard deviation of movement duration with increases in mean movement duration [10] . Here we use a model of optimal temporal movement planning to investigate the control of movement duration in the face of these two types of temporal uncertainty while human subjects attempted to touch a computer screen within a specified temporal window . We introduced asymmetries in the penalties imposed for early vs . late movement timing ( Figure 1A ) , while at the same time increasing subjects' temporal uncertainty by adding Gaussian noise with 25 ms standard deviation ( see Methods ) . As in all models of motor planning and motor control based on decision theory , we are concerned with the interplay of three elements: possible decisions ( here planned movement time , τ ) , uncertainty in the mapping of motor decisions to motor outcomes ( represented by the family of probability distributions p[t|τ] ) , and the costs/benefits resulting from those motor outcomes , G ( t ) . The mathematical models considered here are part of a growing literature on Bayesian decision models of motor phenomena , such as models of motor adaptation [11]–[13] and motor planning/control e . g . , [14]–[21] , including the use of prior information in spatial [16] , [18] and temporal [17] motor planning , the use of asymmetric cost functions in spatial motor planning [14]–[15] , [19] and when selecting a speed-accuracy tradeoff [20]–[21] . The neural computation of decision variables such as those considered here and in previous work has also begun to be investigated [22]–[25] . Figure 1B illustrates the computations needed to maximize expected gain with temporally asymmetric penalties . When discussing movement duration , we must distinguish between the planned arrival time , denoted τ , and the actual arrival time , t . When movements are executed , the actual arrival time will be unpredictably earlier or later than τ . In Figure 1B we show four possible choices of τ and outline the calculation of expected gain for each . Note that the optimal planned arrival time need not fall within the temporal reward window . Human performance will be optimal if the CNS learns its linear temporal uncertainty function , ( 1 ) as it relates to planned movement time ( τ ) , and uses this information ( ασ and βσ ) to plan reach times that maximize expected gain . Human performance in our task could be sub-optimal in several ways , each depending on the type of information the CNS maintains about Equation 1 . We consider 5 such sub-optimal models , denoted M1 , … , M5 . In the first three of these , subjects fail to take account of ασ , βσ , or both when planning reaches . In model M1 , subjects fail to compensate for the experimentally imposed static increase in temporal uncertainty due to the added Gaussian noise ( SD = 25 ms ) ; in M2 subjects fail to compensate for the linear increase in temporal uncertainty with increasing reach duration; and in M3 subjects fail in both respects ( for details , see Methods: Data Analysis and Model Comparison ) . Models M4 and M5 were analogous to models M2 and M3 , respectively , but assumed the offset or slope were unknown and hence not fixed to match the training data or added 25 ms timing uncertainty . We compare subjects' performance to each of these sub-optimal movement strategies , and to the optimal strategy ( M0 ) that results in maximum expected gain .
During training trials , subjects attempted to produce reaches with an experimenter-specified temporal duration; no rewards or penalties were imposed . In Figure 2A , we plot the mean movement duration as a function of the target duration for subject HT . The points lie near the identity line , indicating that the subject could accurately produce a wide range of movement times on command . Figure 2B shows the temporal uncertainty function ( the standard deviation of arrival times as a function of target duration , with and without the added noise ) measured during training for the same subject . As expected , unperturbed standard deviations ( dot-dashed line , open symbols ) increase linearly across this range . Estimated Weber-noise parameters ( ασ ) for all subjects' temporal uncertainty functions , and verification of the stationarity of those functions ( across the training trials and the subsequent main experiment ) , are provided in Figure 3 . Note that fitted functions obtained from training data ( lines ) and the standard deviations measured during main-experiment reaches ( filled diamonds ) were well-matched , consistent with the idea that subject performance did not change during the experimental reaches . Each of the models makes predictions of reach durations that are based on the aspects of the temporal uncertainty function it incorporates . Because the optimal model ( M0 ) incorporates both components of the temporal uncertainty function , it can take account of the temporal noise actually experienced by each subject when planning reaches , in turn allowing it to predict optimal movement times . Three of the sub-optimal models ( M1–M3 ) each specify only a portion of the actual temporal noise experienced by subjects . Because these models cannot account for the full temporal uncertainty function , their predicted ‘best’ movement times are sub-optimal . For each subject and model , we derived predictions of the mean duration in each of the four conditions that would maximize expected gain in the task given that temporal uncertainty function ( see Methods: Model Predictions; Figure 4 illustrates these calculations for an example subject ) . These predictions allow us to compare observed performance in the task to the theoretical performance of subjects who maximize expected gain under the constraints imposed by each of the four models . In addition to these four models , we considered two sub-optimal models that did not have fixed parameters ( M4 and M5 ) . In models of this type , the model likelihood ( see Method: Data Analysis and Model Comparison ) is calculated by integrating over the possible values of the unknown parameters ( e . g . , overall noise level ) . The results of a Bayesian comparison of the performance of the four models ( see Methods: Data Analysis and Model Comparison ) favored the optimal model M0 over the sub-optimal models; yielding 11 . 5 dB in favor of M0 , but −60 . 5 dB , −11 . 5 dB and −41 . 4 dB of evidence for M1 , M2 and M3 , respectively . Models M4 and M5 are less constrained , resulting in evidence below −100 dB . Negative evidence is evidence against a model relative to the other possible models . In our previous work [26] we have used 3 dB evidence , corresponding to odds of nearly 2∶1 , as a minimal guideline for inferring an advantage for a model over its competitors . The 11 . 5 dB evidence for M0 is strong , corresponding to nearly 15∶1 odds in favor of the optimal model over the set of alternatives . To assess inter-subject variability , we recomputed the evidence values for 5 subgroups of subjects , with each subgroup consisting of all subjects but one . The change in evidence that occurred as we left each subject out is a measure of how much the conclusions we draw are based on one subject alone . While the evidence decreases somewhat when each subject is removed ( and it should since we are basing our conclusion on fewer data ) , it always favored M0 , and always by at least 7 . 5 dB , consistent with the conclusions based on all subjects taken together . We note , in particular , that removing the non-naive subject who was an author ( TEH ) still resulted in evidence of 9 dB in favor of M0 . In addition , we plotted , for all subjects and conditions , the mean observed movement duration as a function of the duration predicted by each of the four models ( Figure 5 plots the deviations of the actual from the predicted movement times ) . In such a plot , consistency of the data with the model corresponds to the data falling along the identity line . We computed linear regressions of observed mean duration as a function of predicted mean duration for each of the four models . Only M0 had a best-fit slope and intercept whose confidence intervals contained those of the identity line ( Table 1 ) , corroborating the result of the Bayesian model comparison . We conclude that the evidence favoring M0 over any of the competing models is overwhelming , implying that subjects compensated for their increased uncertainty at longer durations and also for the 25 ms added uncertainty imposed experimentally . To investigate how the suboptimal models fail , we present differences between observed average temporal endpoints and model predictions for each of the four models ( Figure 5 ) . For each of the sub-optimal models , we describe how the pattern would appear if data were fit with that model . Model M1 compensates for increased temporal uncertainty with increased movement duration but fails to compensate for the σ = 25 ms temporal noise added experimentally . Subjects conforming to this model will have temporal aim points closer to the center of the target region than they should be since they are based on an erroneously small estimate of temporal uncertainty . That is , compared with the optimal model ( M0 ) , model M1 predicts longer durations for predictions of durations shorter than the target duration ( 650 ms ) , and shorter durations for predictions longer than the target duration . Thus , we predict the left-hand cloud of residuals to move down and right and the right-hand cloud to move up and left , which is precisely what happened ( upper-right panel , Figure 5 ) . Subjects employing model M2 ( lower-left panel , Figure 5 ) would fail to take duration-dependent noise into account , but compensate for the s = 25 ms temporal noise added experimentally . Such subjects overestimate noise for short durations and underestimate it for long durations . Intuitively , the residuals should move up and left . This is true of most data points , but not all . The intuitive pattern is occasionally broken due to the complex , nonlinear calculation of expected gain ( Figure 1B ) and the switch from the veridical uncertainty function ( M0 ) to an incorrect , flat function ( M2 ) . As expected , the predictions of M3 combine the shifts of the other two suboptimal models . In summary , based on the comparison of the optimal and three suboptimal models , we conclude that subjects delayed or advanced their temporal endpoints in accordance with the calculated optimal times defined by M0 . The Bayesian model comparison employed is novel and correct for comparison of non-nested models ( see Method: Data Analysis and Model Comparison ) . We also carried out a set of statistical tests based on linear regression of actual versus predicted times . The conclusions based on these regressions tests are identical to those just reported: we reject models M1 , M2 and M3 but not M0 ( Table 1 ) . The gains earned by subjects potentially provide an additional dimension for testing the models . We have compared actual gains to expected gains predicted by each of the models . However , the gain functions are flat relative to the sampling variability of observed points earned , so that this analysis does not serve to differentiate the models . To investigate the possibility that subjects used a hill-climbing strategy during the main experiment , instead of maximizing expected gain by taking account of their own temporal uncertainty function and experimentally imposed gain function , we performed a hill-climbing simulation using each subject's temporal uncertainty function . In the simulation , intended duration was moved away from the penalty region by 3Δt ms after each penalty and towards the center of the target region by Δt ms for each miss of the target that occurred on the opposite side from the penalty ( corresponding to the 3∶1 ratio of penalty to reward ) . The value of Δt was initially set to be relatively large . With each change of direction of step , Δt was reduced by 25% to a minimum step size of 1 . 5 ms . While this simulation approximately reproduced the final average reach times observed experimentally , it does not provide a good model of subject performance . First , there were significant autocorrelations of reach durations beyond lag zero in the simulation data but not in the experimental data . Second , a learning algorithm would be expected to produce substantially higher σ values during test than those observed during training . This is what we found with our hill-climbing simulation . Using subjects' training σ values to produce the simulated data , the simulation produced 17 out of 20 main-experiment σ values that were above the training values , whereas our subjects' main-experiment σ values ( Figure 3 ) were entirely consistent with temporal uncertainty functions measured during training .
To move accurately , an organism's motor system must generate an intricate series of precisely timed neural commands . The exact nature of these commands is not known . Whatever the format of the command signals [27]–[32] , movement controlled by any physical controller-actuator system , including biological motor systems , will always exhibit some motor uncertainty . Nevertheless , it is possible to plan movements that will maximize expected gain in the face of that uncertainty . To do so , an organism must be capable of assessing both the probabilities of possible movement outcomes and their consequences . One of the most thoroughly studied cases in which humans integrate the probabilities of possible movement outcomes and their consequences is the tradeoff between movement speed and spatial accuracy [20]–[21] , [33]–[34] . However , in our experiment we were concerned with temporal accuracy , and faster movements are typically more temporally accurate ( the opposite of the spatial speed-accuracy tradeoff ) . By imposing costs for early/late arrivals , we were able to determine whether the motor system is capable of picking movement times that maximize expected gain , taking into account temporal uncertainty . We conclude that , in the timing task we examined , the motor system estimates and compensates almost perfectly for its own temporal uncertainty and correctly anticipates how that uncertainty interacts with the asymmetric reward structure of the environment . This outcome is plausible given the close neurophysiological links between motor timing and the assessment of probabilities and consequences [22]–[25] , [35]–[37] . We note however that it has been argued that a representation of time plays no role in one of the most basic forms of motor learning: motor adaptation [38] . The current study provides evidence that the motor system is capable of using a representation of time in at least some circumstances where the consequences of the movement are unambiguously linked to the timing of the movement , and in addition that it does so optimally . Several models of spatio-temporal movement control are based on optimizing an internal cost function that either includes or predicts movement timing . One such model of trajectory formation , the minimum variance model [39] , assumes that the CNS selects a spatio-temporal reach trajectory by optimizing a cost function based on the movement's endpoint variance . In particular , the minimum variance model selects “…the temporal profile of the neural command … so as to minimize the final positional variance for a specified movement duration…” [39] , p . 782 . More recently the minimum-time model of trajectory formation has been proposed [40] based on the assumption that , subject to a constraint on movement accuracy , the CNS attempts to minimize movement duration . In both models , the speed-accuracy tradeoff is modeled by scaling the spatial variance of the reach with the amplitude of the motor control signal; that is , they assume signal-dependent spatial motor noise . In the absence of signal-dependent noise , both models would predict a ‘bang-bang’ control scheme , where the control signal takes first a maximum positive and then maximum negative value producing alternating maximum forward and reverse accelerations leading to maximum movement speed and hence minimum duration . However , bang-bang control predicts trajectories that are inconsistent with typical motor behavior . By modeling spatial noise as signal-dependent , it is possible to predict a range of important behavioral results with both the minimum-variance and minimum-time models , such as the smooth variation in spatial and temporal reach profiles e . g . , [41]–[42] , Fitts' law [33] , and the spatio-temporal details of saccadic trajectories [43] . Unlike these previous studies , here the emphasis is on accuracy of movement duration . This results in a reverse speed-accuracy tradeoff; slower movements have lower temporal accuracy ( even though they have higher spatial accuracy ) . We show that , in a task where spatial uncertainty ( and therefore signal-dependent spatial noise ) plays essentially no role , reach durations are selected to nearly maximize expected gain in the presence of duration-dependent temporal uncertainty . Duration-dependent temporal uncertainty constitutes a constraint on the temporal aspects of movement planning that is similar in many respects to the planning constraint imposed by signal-dependent spatial noise . Simultaneously minimizing temporal and spatial noise provides a method of solving the underconstrained problem of trajectory selection . Although several previous studies have proposed multiply-constrained models of movement planning [44]–[45] and the duration-dependence of temporal uncertainty is well known e . g . , [10]; [46]–[47] , we provide the first demonstration of the CNS making use of its own temporal uncertainty in movement planning . While selecting the movement trajectory that minimizes spatial and/or temporal noise is a possible method of movement planning , the optimal movement planner carefully separates the constraints imposed on spatial and temporal accuracy ( duration-dependent temporal noise and signal-dependent spatial noise ) with the costs of spatial and temporal errors , which we discuss next . In both the minimum-time and minimum-variance models [39]–[40] , a trajectory is selected so as to optimize an internal cost for spatial variance or movement duration ( respectively ) in the presence of signal-dependent spatial noise . The cost is internal in the sense that it does not make reference to any externally imposed costs on movement errors , such as monetary rewards and penalties that may be imposed due to one's spatial precision or movement duration . There have been a large number of models of movement based on the optimization of internal cost functions that identify movement cost with an invariant kinematic or dynamic variable ( time [48] , spatial precision [39] , torque-change [49]–[50] , jerk [51] , etc . ) . However , there are pitfalls inherent in identifying movement cost with an aspect of the movement itself , despite the current movement goals . For example , the minimum-variance model always chooses a movement with the best possible spatial precision , even when that level of precision is unnecessary for the task . Similarly , the minimum-time model always chooses the shortest duration movement that satisfies the constraint on spatial precision even when , as in some conditions of the current study , an external temporal cost function rewards longer-duration movements . Recent models of optimal movement planning e . g . , [14] , [18] , [26] , [44] approach the problem somewhat differently . In these models , which have previously been used to predict spatial movement endpoints [14] , [18] and movement trajectories [44] , the difference between a constraint on movement planning and a cost incurred from movement error must be recognized . While duration-dependent temporal noise , signal-dependent spatial noise , energy consumption , biomechanics , etc . constitute constraints on movement planning and control , they are not properly costs . A cost essentially imposes a weighting on the available constraints , and is task dependent . By experimentally imposing costs [14]–[15] , [18]–[21] , [26] on spatial or temporal inaccuracy , it is possible to predict flexible movement strategies that incorporate task-relevant constraints ( e . g . , duration-dependent temporal uncertainty ) while effectively ignoring ( down-weighting ) constraints that are not as important to the task at hand ( signal-dependent spatial uncertainty ) . In the present study , we manipulated the temporal cost function by imposing penalties on too-short reach durations in some conditions , and too-long durations in other conditions , and determined whether subjects responded appropriately to these different cost functions . We have modeled movement planning as minimizing an external gain function in the presence of task-relevant internal temporal noise . By identifying the to-be-minimized cost with the movement goal we have separated fixed kinematic/dynamic variables from the purpose of the movement . This allows us to predict flexible movement plans that may minimize spatial or temporal uncertainty , but only when that is relevant to the task at hand . A deeper understanding of movement planning and execution will result from models that similarly separate cost functions from fixed sets of kinematic/dynamic variables while simultaneously taking account of task-relevant spatial and/or temporal uncertainty .
Subjects were first given a training session in which temporal targets ( width: 3 ms , no adjacent penalty region ) were presented at six target durations ( 565 , 595 , 625 , 655 , 685 and 715 ms; 8 repetitions each , in separate blocks , followed by 50 repetitions each , in separate blocks ) spanning the range of temporal aim points observed during pilot work . Although this window was too narrow for subjects to reliably hit , subjects were not scored during training , and were told simply to time their reaches as closely to each target time as possible . This session allowed us to estimate the standard deviation of each subject's movement durations for a set of precisely known target durations , and also allowed subjects to learn their own ( noise-added ) temporal uncertainties in the task . Standard deviations at each target time ( Figures 2B and 3 ) were measured from the final 40 repetitions to avoid possible initial practice effects . Immediately following training , subjects were given a temporal target centered at 650 ms , with a half-width of 0 . 6σ650 , where σ650 was the estimated SD of movement duration for a mean duration of 650 ms . In this way , we equated the difficulty of the task across subjects based on their training performance . Subjects were paid a bonus for touching the spatial target within the temporal target window ( Figure 1A , green , cross-hatched bars ) , and penalized for touching the spatial target within a temporal penalty window ( Figure 1A , red , striped bars ) or for failing to touch the spatial target . Four blocked conditions were employed ( Figure 1A ) , two early temporal penalty conditions and two late penalty conditions ( 64 trials each ) . The two early temporal penalty regions began at 0 ms and ended either 0 . 6σ650 or 1 . 35σ650 ms prior to 650 ms . The two late temporal penalty regions began either 0 . 6σ650 or 1 . 35σ650 ms following 650 ms , and were open-ended . The outcome of each trial was signaled by distinct auditory tones notifying the subject that a reward was earned or a penalty assessed . The possible reward earned on any trial was $0 . 12 and the penalty was −$0 . 36 ( or −$0 . 60 for missed spatial targets ) . Note that the ratio of penalty to bonus magnitudes was 3∶1 . Trials in which the spatial target was not touched were re-run ( fewer than 1% of all trials ) to equate the number of touched-target trials in each condition . The untouched-target trials were not analyzed . Subjects were four students at New York University who were not aware of the purpose of the experiment and one author ( TEH ) . All subjects gave informed consent before the experiment . The experimental protocol had been approved by the Institutional Review Board at New York University . As described in the Introduction , decision theoretic models of motor behavior are concerned with the interplay of three elements: movement strategy , uncertainty , and the gain or loss from possible movement outcomes . The interplay of these three elements is represented graphically in Figure 1B for the optimal model , M0 . Calculation of the temporal endpoints predicted by each of the models to be considered required that the expected gain , in terms of average bonus earned per reach , be computed based on the constraints supplied by the hypothetical system . For example , the optimal neuromotor controller would make use of information concerning both Weber-like increases in temporal uncertainty with increasing reach time , and the experimentally increased overall temporal uncertainty . A given motor strategy or plan , s , determines the critical states of the system . Although motor plans are complex sequences of control signals in time , the only consequence of the choice of motor plan in our task is to select an expected temporal endpoint , τs . The expected gain from s is then given by ( Figure 1B ) : ( 2 ) where G ( t ) describes the gain or loss associated with a particular temporal endpoint ( Figure 1A and Figure 1B , middle panel ) . The term p ( t | τs ) describes the probability density of temporal endpoints expected from any chosen movement strategy s . Note that these are planned durations , not reaction times , and hence we have no a priori expectation that these distributions will be skewed . We model the duration distribution as a Gaussian with mean arrival time τs and a standard deviation σ ( τs ) ( 3 ) ( QQ plots of these distributions confirm that the Gaussian distribution models the data well ) . The temporal uncertainty function , σ ( τs ) is able to capture the well-known Weber-like scaling of temporal standard deviation with mean arrival time τs ( Figure 1B , top panel ) . We used values estimated from each subject's training data to compute individual σ ( τs ) functions for models M0–M3 . In Figure 1B ( bottom panel ) , for the rightmost choice of τ , the probability of arrival in the penalty zone is nearly as high as that of arrival in the reward zone . This choice of τ is likely to lead to nearly as many penalties as rewards . Given that the penalty/reward ratio was 3∶1 , expected gain is negative for this choice of τ . The distribution associated with the leftmost choice of τ is primarily in the uncolored time zone where the subject earns nothing . This choice of τ is likely to lead to rare rewards and extremely rare penalties , resulting in only a small total reward across many trials . Interestingly , a third choice of τ , centered on the temporal reward region , earns even less than the previous choice of τ because of a combination of its proximity to the temporal penalty , the magnitude of temporal movement noise , and the ratio of the reward to penalty magnitudes . The best of the four choices shown is therefore the τ located at the left edge of the rewarded temporal region . Of the four shown , it makes the best compromise between the width of the probability distribution for t and its distance from the centers of the reward and penalty regions , given the widths of those regions and the ratio of gains to losses . Of course , there are infinitely many possible choices of τ . The lower panel shows the expected gain as a function of τ , with the maximum expected gain ( MEG ) point indicated with a circle at the peak of the expected gain function . If observers select this value τopt , they maximize their expected gain . We computed τopt for each of the four penalty conditions and each subject based on an estimated temporal uncertainty function σ ( τs ) that was specific to each subject . In all cases the optimal ( maximum expected gain ) value of τs was shifted away from the penalty region . The optimal Bayesian model ( M0 ) makes full use of the temporal uncertainty function σ ( τs ) from each subject's training session . The five sub-optimal models use less information . M1 uses the σ ( τs ) calculated from each subject's training data without the experimentally added σ = 25 ms noise . M2 uses each subject's constant σ for all τs that includes the overall added σ = 25 ms noise; it uses the square root of the average of perturbed variances about the target durations measured during training . M3 uses the subject's constant σ without the experimentally added noise . M4 and M5 use a constant offset and constant offset and slope , respectively , but assume that the values of these parameters are unknown . Of course , some subjects are more accurate than others but this is explicitly taken account of in our analysis . Each model's predictions are defined in terms of performance relative to an individual's temporal uncertainty function . Subjects who are inherently poorer timers are being compared to a standard ( defined by each model ) that is tailored to ( defined in terms of ) the limits of that subject's abilities . So while there are in fact individual differences between subjects , these were removed in the design and analysis of the experiment . Because we equated subjects in this way we could analyze group data . The predicted movement strategy , s , is therefore a function of the type ( s ) of temporal uncertainty information incorporated by each model Mm , the reward structure defined by the jth experimental condition ( j = 1 to 4 ) , and the temporal uncertainties measured during training for the kth subject ( k = 1 to 5 ) . Let denote the value of τ predicted by model Mm based on an estimate of timing uncertainty calculated from the assumptions of each model . For convenience , we denote the temporal uncertainty for an attempt to produce a movement duration of ( using the full temporal uncertainty function based on the training trials ) , , as . The models we considered are not all nested and consequently we chose a method of model comparison for non-nested models [52]–[54] that we describe next . Let denote the ith arrival time ( of the 64 trials per condition ) in condition j for the kth subject . The likelihood of model Mm is given by: ( 4 ) where . ( 5 ) Note however that for M4 and M5 , the model likelihood must be calculated by integrating over the unknown parameters: the constant offset , , and constant offset and slope , , of the temporal uncertainty function , respectively , where the prior probability distributions over the parameters are taken to be bounded Jeffreys ( uninformative ) priors [55] . Let π ( Mm ) denote the prior probability of the mth model . Then the posterior probability of the mth model given the data is ( 6 ) and ( 7 ) is a comparison of the posterior probability of the optimal model M0 to the combined posterior probabilities of sub-optimal models: it is a measure of evidence [53] favoring the optimal model ( the factor of 10 allows us to express evidence in decibels , denoted dB ) . A similar evidence measure can be computed for each of the sub-optimal models using the odds ratio of the probability of each sub-optimal model to the combined probability for the remaining five models ( four sub-optimal and one optimal ) . We set the prior probabilities of the six models to be equal and computed these evidence measures .
|
Many recent models of motor planning are based on the idea that the CNS plans movements to minimize “costs” intrinsic to motor performance . A minimum variance model would predict that the motor system plans movements that minimize motor error ( as measured by the variance in movement ) subject to the constraint that the movement be completed within a specified time limit . A complementary model would predict that the motor system minimizes movement time subject to the constraint that movement variance not exceed a certain fixed threshold . But neither of these models is adequate to predict performance in everyday tasks that include external costs imposed by the environment where good performance requires that the motor system select a tradeoff between speed and accuracy . In driving to the airport to catch a plane , for example , there are very real costs associated with driving too fast and also with being just a bit too late . But the “optimal” tradeoff depends on road conditions and also on how important it is to catch the plane . We examine motor performance in analogous experimental tasks where we impose arbitrary monetary costs on movements that are “late” or “early” and show that humans systematically trade off risk and reward so as to maximize their expected monetary gain .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"neuroscience/behavioral",
"neuroscience",
"neuroscience/sensory",
"systems",
"neuroscience/motor",
"systems",
"computational",
"biology/computational",
"neuroscience",
"neuroscience/theoretical",
"neuroscience"
] |
2008
|
Optimal Compensation for Temporal Uncertainty in Movement Planning
|
The mechanisms by which Regulatory T cells suppress IL-2 production of effector CD4+ T cells in pathological conditions are unclear . A subpopulation of human Treg expresses the ectoenzyme CD39 , which in association with CD73 converts ATP/ADP/AMP to adenosine . We show here that Treg/CD39+ suppress IL-2 expression of activated CD4+ T-cells more efficiently than Treg/CD39− . This inhibition is due to the demethylation of an essential CpG site of the il-2 gene promoter , which was reversed by an anti-CD39 mAb . By recapitulating the events downstream CD39/adenosine receptor ( A2AR ) axis , we show that A2AR agonist and soluble cAMP inhibit CpG site demethylation of the il-2 gene promoter . A high frequency of Treg/CD39+ is associated with a low clinical outcome in HIV infection . We show here that CD4+ T-cells from HIV-1 infected individuals express high levels of A2AR and intracellular cAMP . Following in vitro stimulation , these cells exhibit a lower degree of demethylation of il-2 gene promoter associated with a lower expression of IL-2 , compared to healthy individuals . These results extend previous data on the role of Treg in HIV infection by filling the gap between expansion of Treg/CD39+ in HIV infection and the suppression of CD4+ T-cell function through inhibition of IL-2 production .
Regulatory T cells ( Treg ) play a dominant role in self-tolerance , control of autoimmune diseases and control of chronic infections by suppressing effector T cells activation , proliferation and functions [1] . Natural Treg derive from the thymus and are characterized by high levels of IL-2 receptor ( CD25 ) and transcription factor FoxP3 and low levels of IL-7 receptor alpha ( CD127 ) [2]–[5] . Induced Treg are heterogeneous and their phenotype and frequency vary across different disease states . They include interleukin-10 ( IL-10 ) producing Tr1 , transforming growth factor ( TGF-β-expressing Th3 cells ) [6] , [7] and also Foxp3+CD39+ effector/memory Tregs [8] . The imbalance of T cell responses in favor of Treg can hamper efficient effector T cell responses as it has been observed in cancer and certain chronic infections [9] . In acute and chronic phases of HIV infection , a dual role for Treg has been reported due to their expansion [10]–[12] . Treg can suppress anti-HIV specific CD4+ and CD8 T cell responses by inhibiting cytokine production and cell proliferation [13] , [14] . Increased Treg frequency at the mucosal site is accompanied by increased immune activation and decreased HIV-specific T-cell responses [15] . However , Treg can have a beneficial role by protecting HIV infected patients either at the primary or chronic phase of infection from the deleterious effects of HIV-induced chronic immune activation [11] , [16] , [17] . In HIV controllers , low frequencies of Treg have been associated with effective adaptive immune responses , but also with generalized immune activation and CD4 depletion [18] . Numerous mechanisms of Treg suppression have been reported [1] . These include secretion of inhibitory cytokines ( IL-10 , TGF-ß or IL-35 ) , induction of apoptosis by IL-2 deprivation , perforin/Granzyme B or by CTLA-4 and GITR interactions pathways [1] , [19] . Treg also use CD39 ( nucleoside triphosphate diphosphorylase-1 ) and CD73 ( ecto-5′-nucleotidase ) for their suppressive activity . These ecto-enzymes hydrolyse extra-cellular pools of inflammatory ATP into adenosine diphosphate ( ADP ) and/or adenosine monophosphate ( AMP ) to adenosine [20]–[25] . Extracellular adenosine is known to be an important physiological regulator of the immune response [26] , [27] by inhibiting T cell proliferation and IFN-γ/IL-2 production [28] and these effects are mediated through the adenosine-receptor A2A ( A2AR ) by stimulating the generation of intracellular cyclic AMP ( cAMP ) [28] . It has been recently shown that Treg inhibit HIV replication in conventional T cells through cAMP-dependent mechanisms [29] . We have recently evaluated the impact of CD39/adenosine pathway in HIV pathogenesis and reported that expanded Treg/CD39+ in infected patients correlate with immune activation and CD4+ cell depletion [30] . Importantly , we showed that these Treg exerted a strong suppressive effect on effector CD8 T cell functions and these inhibitory effects were relieved by using an anti-CD39 monoclonal antibody [30] . Here we explored the molecular mechanisms used by Treg/CD39+ cells to mediate their suppressive activity on CD4+ T cell function during HIV infection . Using co-culture experiments , we show that Treg/CD39+ cells inhibit IL-2 mRNA expression in activated effector CD4+ T cells . Importantly , this inhibition was partly reversed when CD39 enzymatic activity was blocked by an anti-CD39 mAb . We reasoned that this effect could be mediated through an epigenetic regulation of IL-2 expression involving cAMP-dependent mechanisms . We found that IL-2 inhibition mediated by Treg/CD39+ was correlated with a decrease in CpG demethylation of the il-2 gene promoter . This effect was reproduced by using A2AR agonist as well as soluble cAMP . We also found that CD4+ T cells from HIV infected patients express high level of cytoplasmic cAMP and exhibit a lower frequency of demethylated CpG site of the il-2 gene promoter following in vitro activation through the T cell receptor , when compared to healthy donors . Accordingly , A2AR expression was higher in ex vivo CD4+ T cells from HIV+ patients as compared to healthy controls . All together , these results make the link between Treg/CD39+ expansion and epigenetic mechanisms of IL-2 regulation .
We first investigated whether Treg inhibited the expression of IL-2 in autologous anti-CD3 stimulated naive CD4+ T cells in co-culture experiments ( Figure 1A ) . Naive CD4+ T cells were labeled with CFSE before co-culture with Treg populations and activated overnight with anti-CD3/28 mAbs . CD4+CFSEhigh non-dividing cells were then FACS-sorted and IL-2 mRNAs were quantified by qRT-PCR . As shown in Figures 1B and 1C , Treg/CD39+ and Treg/CD39− inhibited dramatically mRNA IL-2 expression of anti-CD3/28 activated CD4+ T cells , but this effect was more pronounced , in the presence of Treg/CD39+ as compared to Treg/CD39− . Interestingly , in the presence of blocking anti-CD39 mAbs [29] , [31] , the suppressive function of Treg/CD39+ was decreased by 25±4% ( P<0 . 05 ) . As expected , these antibodies have no effect in co-cultures with Treg/CD39− . These results show that Treg/CD39+ inhibit , at least partially , the expression of IL-2 through the enzymatic activity of CD39 . Moreover , we directly assessed CD39 enzymatic activity and measured ATP catalysis into ADP and AMP by HPLC and we have also quantified inorganic phosphate , which derives ATP hydrolysis . The results demonstrate the catalysis of exogenous ATP into ADP and AMP in the presence of purified Treg/CD39+ but not Treg/CD39− . Interestingly this catalysis was inhibited when an anti-CD39 mAb was added to the cells ( Figure S1 ) . We also evaluated the production of Adenosine via CD73 in our co-culture model . A significant increase in extracellular CD73 expression was observed in both naive CD4 T cells and Tregs upon overnight anti-CD3/28 mAbs stimulation ( 7±7 . 3 vs . 22 . 6±8 . 8% and 6±5 . 5 vs . 20 . 7±5 . 3 , respectively , P<0 . 05 , Figure S2 ) . In line with these data , AMP was converted to adenosine in a specific manner as this conversion was totally inhibited by an inhibitor of CD73 enzymatic activity ( 6 . 3±6 . 4 vs . 0 . 13±0 . 3 µM , P<0 . 05; Figure S3 ) . Demethylation of the unique specific CpG site 1 in the il-2 promoter is essential for inducing IL-2 production by TCR activation [32] . We therefore analyzed the methylation status of the CpG site 1 by the bisulfite genomic sequencing method on DNA extracted from anti-CD3/28 activated CD4+ T cells cultured or not with Treg/CD39+ . Sequencing of il-2 promoter gene region of 25–30 clones of each experimental condition was performed . As shown in Figure 1D , in vitro activation of CD4+CD45RA+CD25low naive cells led to a higher frequency of demethylated CpG site 1 in the il-2 gene promoter as compared to non-activated cells ( 43 vs . 21%; P = 0 . 01 ) whereas this effect was inhibited significantly when Treg/CD39+ were added to the co-cultures in the presence of an irrelevant IgG control ( 26% , P = 0 . 02 ) . However , this inhibitory effect of Treg/CD39+ was partially reversed in the presence of an anti-CD39 mAb ( 32% , P>0 . 05 ) . All together these data suggest that Treg/CD39+ suppress IL-2 expression in activated cells at the il-2 gene promoter level . To assess whether CD39/adenosine pathway is involved in the Treg/CD39+ mediated inhibition of IL-2 expression , we evaluated the effects of the A2AR agonist CGS21680 and A2AR antagonist ZM241385 , on IL-2 expression of anti-CD3/CD28 stimulated-CD4+ T cells . We found that CGS inhibited significantly the expression of IL-2 ( 69±11 . 5% as compared to DMSO control condition ) . This effect was partially relieved when the A2AR antagonist ZM was added to cultures of activated CD4+ T cells in the presence of CGS ( 33±4%; P = 0 . 04 for comparison of CGS and CGS+ZM conditions ) ( Figure 2A ) . Of note , ZM alone did not alter the expression of IL-2 transcripts of activated CD4+ T cells . Next we looked at the effects of A2AR agonist at the DNA level ( Figure 2B ) . The frequency of demethylated CpG site 1 of the il-2 promoter gene in the presence of anti-CD3/28 was 76% and became 53% in the presence of CGS which corresponded to 30% inhibition of CpG demethylation ( Figure 2B ) . Addition of ZM before adding CGS to activated CD4+ T cells restored the frequency of demethylated CpG at the same level than activated CD4+ T cells ( 70% and 75% , respectively; P = NS ) . No effect of the DMSO as control of the vehicle of CGS and ZM was observed ( percentage of demethylation around 75% in activated CD4+ T cells ) . These results show that the adenosine pathway is involved in the epigenetic regulation of the expression of the IL-2 gene in activated CD4+ T cells . In CD39/Adenosine enzymatic cascade , A2AR signaling activates intracellular adenyl cyclase enzyme , which increases intracellular levels of cAMP in conventional T cells [29] . To assess whether adenyl cyclase was involved in CD39-induced inhibition of IL-2 expression , we pre-incubated CD4+ naive T cells with adenyl cyclase inhibitor ddADA or adenyl cyclase activator forskolin , 30 minutes before anti-CD3/CD28 stimulation . As shown in Figure 3A , forskolin inhibited dramatically the expression of IL-2 transcripts in stimulated cells ( 97±2% inhibition ) . In contrast , inactivation of adenyl cyclase by ddADA favored IL-2 expression . Accordingly , we found that the frequency of CpG site demethylation of il-2 gene promoter was 42% in ddADA conditions ( P = 0 . 02 for comparison of ddADA and non-activated conditions ) while it remains close to non-activated cells in the presence of forskolin ( 21% and 27% , respectively , P = 0 . 33; Figure 3B ) . To assess the impact of cAMP on CD4+ T cell proliferation , IL-2 mRNA expression and CpG site demethylation , CD4+ naive T cells were CFSE labeled and activated with anti-CD3/CD28 mAbs . As shown in a representative experiment performed in triplicate ( Figure 4A ) , cAMP inhibited CD4+ T cell proliferation in a dose dependent manner ( 16±16% and 75±5% for 100 and 1000 µM respectively; Figure 4B ) . Similarly , cAMP inhibited IL-2 mRNA expression in activated CD4+ T cells a dose dependent manner ( 39±16% and 67±15% inhibition for 100 and 1000 µM respectively; Figure 4C ) . As shown in Figure 4D and as compared to non-activated CD4+ T cells , anti CD3/CD28 mAbs led to an increase of 26% in the percentages of demethylated CpG site 1 ( P = 0 . 012 ) , while no changes were observed when cAMP ( 1000 µM ) was added to the culture ( 1% changes from non-stimulated conditions ) . Next , we investigated whether the effects of cAMP on IL-2 mRNA expression translated to a decrease in the production of IL-2 by activated CD4+ T cells and if this effect was restricted only to naive CD4+ T cells . For this , naïve ( N ) , central ( CM ) , effector memory ( EM ) as well as terminally differentiated effectors ( TE ) ( Figure 5A ) were stimulated with anti-CD3/CD28 mAbs with or without different doses of cAMP . As shown in Figure 5B and C ( for one representative experiment and pooled data , respectively ) , at the highest dose , cAMP inhibited by up to 75% the frequency of N and CM IL-2 producing cells as assessed by ICS assay ( Figure 5C ) . This effect was also notable but less dramatic when cAMP was added to EM or TE ( Figure 5C ) . Interestingly , the frequency of IL-2 producing cells following anti-CD3/CD28 stimulation within these latter subsets was higher than those of N and CM . Accordingly , ex vivo analysis of EM and TE FACS-sorted cells showed that the frequency of CpG site 1 demethylation reached 92–100% ( data not shown ) . It is well known that in chronic HIV infection , T-cell dysfunction is characterized by reduced IL-2 production [33] , [34] . The mechanisms leading to this defect remain unclear . Given our previous results showing that chronically infected HIV patients exhibit high levels of Treg/CD39+ [30] , we investigated the potential implication of the CD39/A2AR/cAMP pathway in the regulation of IL-2 expression in CD4+ T cells purified from HIV-1 infected patients . Figure 6 shows that both naive and memory CD4+ T cells from HIV+ patients express significant higher levels of A2AR mRNA as compared to healthy controls ( P<0 . 05 and P<0 . 01 respectively; Figure 6A ) . Ex vivo CD4+ T cells from HIV+ patients ( n = 6 ) exhibit also higher levels of intra-cytoplasmic cAMP ( mean 548 . 3±9 . 1 fmol/million cells ) as compared to controls ( n = 6; mean 761 . 6±29 . 4 fmol/million cells ) ( P = 0 . 002; Figure 6B ) . Therefore , we quantified IL-2 mRNA expression in purified naive CD4+ T cells from HIV+ART- patients ( n = 6 ) and healthy controls ( n = 8 ) following stimulation in the presence of high doses of anti-CD3/CD28 mAbs ( 5 µg/ml ) . As shown in Figure 6C , IL-2 mRNA levels were significantly lower in stimulated CD4+ T cells from HIV+ patients as compared to healthy controls ( P = 0 . 004 ) . In another set of experiments performed on a DNA pool obtained from ex vivo and anti-CD3/CD28 activated CD4+ T cells from HIV+ART- patients and healthy controls ( n = 3 per group ) , the analysis of a large number of molecular clones ( 43 to 67 clones analyzed in each experimental condition ) showed that the frequencies of demethylated CpG site 1 in non-stimulated cells were identical in HIV+ patients and healthy controls ( P = 0 . 45 , Figure 6D ) . Anti-CD3/CD28 activation increased significantly CpG demethylation in HIV− subjects ( P = 0 . 006 for the comparison between ex vivo and activated conditions ) but not in HIV+ patients ( P = 0 . 67 ) . Importantly , we showed that the status of patients ( HIV− and HIV+ ) is significantly correlated with CpG demethylation of the il-2 promoter gene upon anti-CD3/CD28 activation ( P = 0 . 02 ) . All together these results demonstrate a constitutive high expression of A2AR and cAMP resulting in a clear inhibitory effect on CpG demethylation accompanied by the lack of IL-2 production in HIV+ART- upon anti-CD3/CD28 activation .
The mechanisms used for Treg's immunosuppressive function during the course of HIV infection are not completely elucidated . Recently we , and others , have shown an increased CD39 expression by Treg in HIV infected progressors compared to healthy controls [30] , [35] . Moreover , we have shown that blocking CD39 enzymatic activity increased the production of cytokines by HIV-specific T cells . Genetic analysis of several cohorts of HIV-infected individuals showed a relative protection against the development of AIDS associated with CD39 genetic polymorphism [30] . All together these studies strongly suggest that the CD39/Adenosine pathway may play a detrimental role contributing to T cell dysfunction in HIV infection . Data presented here extend our previous results by demonstrating that Treg/CD39+ are potent suppressor of IL-2 production by effector T cells as compared to Treg/CD39− . By recapitulating the steps involved downstream of pericellular adenosine signals , we show that the involvement of the CD39/adenosine/cAMP pathway impacts on il-2 gene promoter by inhibiting the demethylation of the unique specific CpG site 1 in il-2 gene promoter , a seminal event for IL-2 expression in activated CD4+ T cells [32] . Furthermore we show that CD4+ T cells from HIV infected individuals are excessively sensitive to this pathway . Our data demonstrate that CD4+ T cells from HIV infected individuals are partially resistant to demethylation of CpG site 1 following an in vitro stimulation with anti-CD3/CD28 mAbs as compared to cells from healthy controls . Likely , this defect is associated with a lower expression of IL-2 by activated CD4+ T cells . Of note , as the majority of CpG site 1 in memory CD4+ T cells is already demethylated ( more than 80% ) , in our in vitro anti-CD3/CD28 mAbs stimulation experiments , we used naïve CD4+ T cells to be able to study the induced modifications in CpG site . Treg can induce cAMP in effector T cells by increasing adenosine levels in the microenvironment through CD39 and CD73 ectoenzyme pathways [8] , [22] . In contrast to CD39 which is expressed on both human and murine Treg , CD73 is found only at the surface of murine Treg and this molecule is mostly absent on human Tregs membrane [8] , [22] , [36] . However , it is found in intra cellular compartment of human Tregs [36] . A rapid export of pre-formed CD73 to the surface of T cells due to the activation and its removal from the cell surface by an enzymatic cleavage has been reported [37]–[39] . Within human T cells , CD73 is mostly expressed by effector CD4 and CD8 cells [40] and might also be present in soluble form in the microenvironment [41] . It has been also shown an up regulation of CD73 mRNA in activated T cells [42] . In line with this observation , overnight co-culture experiments showed an increase CD73 expression at the surface of both naïve CD4+ T cells and Tregs upon anti-CD3/CD28 mAbs activation together with the conversion of both ATP and AMP into Adenosine . This molecule may be up regulated in inflammatory conditions and in cancerous tissues accompanied by high enzymatic activity [42]–[44] . Several studies have shown that adenosine plays an important non-redundant role in the regulation of T-cell activation via its specific A2AR [26] , [45]–[47] . Signals induced by agonists of A2AR have an inhibitory effect on INF-γ and IL-2 production by effector T cells [28] . We confirm here our previous data [30] showing that both naive and memory CD4+ T cells from HIV-infected individuals express high levels of A2AR compared to healthy controls . Importantly , we show that by using an A2AR agonist , there was a specific and significant decrease in CpG site 1 demethylation of the il-2 gene promoter followed by a decrease in IL-2 mRNA expression . These results help to make the link between pericellular adenosine signals through the purinergic receptor A2AR and dysfunction of Treg target cells in HIV infection . Signals induced by A2AR agonists increase intracellular levels of cAMP [28] via activation of intracellular adenyl cyclase [48] . cAMP is known as an inhibitor of several cellular functions and immune responses such as T cell proliferation [49] and IL-2 production [50] . It has been shown in a murine AIDS model [51] and in ex vivo studies that T cells from HIV+ patients [52] exhibit higher levels of intracellular cAMP , resulting in a higher sensitivity of these cells to inhibition by cAMP analogues as compared to uninfected T cells [51] , [52] . In accordance with this , we show significant increased intracellular cAMP levels in CD4+ T cells from HIV+ patients compared to healthy controls . Our data strongly suggest that the increased expression of cAMP in CD4+ T cells from HIV infected patients impairs IL-2 epigenetics regulation . First , we show that by using soluble cAMP with anti-CD3/28 stimulated CD4+ T cells we prevented both CpG site 1 demethylation of the il-2 gene promoter as well as IL-2 mRNA and protein expressions . In line with our results , it has recently been shown in systemic lupus erythematosus , that cAMP has suppressive activity on IL-2 and IL-7 production through epigenetic modifications in IL-2 and IL-7 promoter genes [53] , [54] . Increased levels of cAMP [28] are mediated by intracellular adenyl cyclase activity [48] in effector T cells [29] , [55] , [56] . cAMP has been described as a key component of Treg mediated suppression [57] and this suppression can be reversed by inhibition of adenylate cyclase activity [38] . In accordance with these studies , our results demonstrate that the adenyl cyclase activator forskolin , inhibits totally IL-2 mRNA expression and il-2 specific CpG site 1 demethylation in activated T-cells . In contrast , adenyl cyclase inhibitor ddADA favors IL-2 production . These data suggest that Treg suppress IL-2 production through cAMP-dependent mechanism which directly impacts on CpG site 1 demethylation in the promoter regions . The reason why CD4+ T cells from HIV infected individuals express higher levels of cAMP could be related to indirect or direct pathways . Likely , in the context of HIV infection several pathogenic pathways could favor the generation of adenosine and cAMP such as the generation of extracellular high ATP levels related to chronic activation and inflammation and the increased frequency of Treg/CD39+ in HIV [30] . Moreover , CD4+ T cells exhibit an increase in ATPase activity , a result that was associated with a higher percentage of cells expressing CD39+ [58] . Recently , in a model of acute SIV infection , a high expression of CD39 on CD8+FOXP3+CD25+ T cells was shown in the gut mucosa , a site of intense viral replication and inflammation [59] . On the other hand , Treg can also increase intracellular cAMP in effector T cells using an alternative mechanism , via gap junction , as it has recently been reported [29] , [57] . These junctions allow intercellular communication between adjacent cells and the passage of ions and other molecules . It has been shown that resting T cells exhibit a low density of these channels . Whether chronically activated CD4+ T cells from HIV infected patients exhibit higher levels these channels warrants further studies . From a physio-pathological standpoint , cAMP may play a dual role: a deleterious role by reducing HIV-specific antiviral immune responses [56] and T cell dysfunction as shown here and also a protective effect by limiting viral replication in infected cells and decreasing viral entry . It has been recently reported that increased cAMP levels through in vitro adenylate cyclase activation with forskolin diminished viral transcription and levels of HIV-p24Gag protein in activated T cells [60] , [61] . Moreover , it is well known that during HIV infection there is an important decrease in CD4+ T cell proliferation and IL-2 production in viremic patients [34] , but the mechanisms leading to this anergy remain unclear . Our data clearly show that even at high dose of anti-CD3/CD28 conditions , CD4+ T cells from chronically infected and untreated HIV patients , were not able to induce CpG site 1 demethylation of the il-2 gene promoter which consequently impairs the production of IL-2 . Further studies are needed to determine the role of CD39/adenosine/cAMP pathway in HIV acute infection but also in HIV infected patients under antiretroviral therapy , in order to evaluate whether these defects could be restored after treatment . It would be also interesting to evaluate whether CD39 mediated ATP hydrolysis as well as intra-cellular levels of cAMP differ according to the stage of HIV infection and disease progression notably in rapid progressors and elite controllers . Moreover , it will be also interesting to assess whether the different transcription factors necessary for an effective IL-2 expression , such as Oct-1 [32] , which are recruited at the promoter level upon cell stimulation , are the same in HIV-infected patients with different clinical outcomes compared to healthy individuals . These studies will provide novel findings , which could help explain the transcriptional repression of the il-2 gene in chronically infected HIV patients . Altogether , our data strongly suggest that in viremic HIV+ patients , the decrease in T-cell proliferation and IL-2 expression is due in part to the inability of CpG site 1 to demethylate upon T cell stimulation . This defect is caused by increased intracellular cAMP , due in part to increased hydrolysis of inflammatory ATP by both expanded Treg/CD39+ and increased A2AR expression levels . Thus , our study establishes the link between Treg/CD39+ expansion and epigenetic mechanisms of IL-2 regulation in progressive HIV infection .
Blood samples from antiretroviral therapy ( ART ) naive HIV-infected patients and HIV-negative healthy donors were collected at The Clinical Immunology Department of Henri Mondor Hospital and the Regional Blood Transfusion Centre , Creteil , France . Ethical committee approval and written informed consent from all subjects were obtained before study initiation . Total CD4+ naive and memory T cells were purified using negative isolations kits from Miltenyi Biotec ( Bergisch-Gladbach , Germany ) according to the manufacturer's instructions . Treg/CD39+ and Treg/CD39− populations were FACS sorted using a moFlow cell sorter ( Beckman-Coulter ) . The purity of sorted populations was >95% . Treg cells were defined by CD4+CD25highFoxP3+CD127low T cells as we have previously reported [30] . Facs-sorted CD4+CD45RA+CD25− naive T cells were stained with 0 . 5 µM CFSE ( Molecular probes , Eugene OR , US ) and co-cultured with Facs-sorted Treg/CD39+ or Treg/CD39− at 1∶2 ratio , in the presence of 1 µg/mL anti-CD3 and anti-CD28 mAbs ( Beckman Coulter , Villepinte , France ) . Total cell concentration was 3×105/well ( 96-well plate ) in a final volume of 200 µl . In some experimental conditions anti-CD39 mAb ( 10 µg/ml , clone A1 , BioLegend , San Diego , LA ) or IgG control was added to the cultures . After 18H , CFSE+ activated but non-divided CD4+CD45RA+CD25− T cells were Facs-sorted , then RNA and DNA extracted ( Figure 1A ) . Treg/CD39+ or Treg/CD39− cells ( 5×104 cells/well ) were co-cultured with effector CD4+ T cells ( 105 cells/well ) in the presence or absence of anti-CD39 mAb or control IgG1 mAb ( 10 µg/mL ) for 2 h . The cells were then washed with a phosphate-free reaction buffer ( containing 0 . 5 mM CaCl2 , 120 mM NaCl , 5 mM KCl , 60 mM glucose , and 50 mM Tris –HCl buffer , pH = 8 ) and ATPase activity was initiated by the addition of ATP or AMP at a concentration 100 µM in 200 µl of reaction buffer for 120 and 45 min respectively at 37°C . To block the internalization of Adenosine , the cells were pre-incubated for 15 min with the adenosine transporter inhibitor Dipyridamole ( Sigma-Aldrich ) , at a concentration of 10 µM , prior to the addition of ATP or AMP . In some experiments a CD73 inhibitor , adenosine 5′- ( α , β-methylene ) diphosphate ( Sigma-Aldrich ) , was added at a concentration of 100 µM 15 min prior to the addition of ATP or AMP . The released inorganic phosphate by hydrolysis of ATP was measured using the malachite green phosphate detection kit ( R&D System , Minneapolis , USA ) according to the manufacturer's instructions . In some experiments the supernatants were frozen ( −80°C ) until analysis by HPLC . This was done with either an Ultimate 3000 Thermofisher HPLC coupled with a UV detector on a reverse-phase column ( Lichrospher 100-5 RP18 Macherey-Nagel ) using a mobile phase gradient from 0 to 20% acetonitrile/50 mM KH2PO4 ( pH = 6 ) containing 10 ml of Tetrabutylammonium phosphate 5 mM [62] , or with a Beckman Coulter System Gold HPLC coupled with a UV system gold 168 detector on a reverse-phase column ( Phenomenex Luna 3u C18 ( 2 ) 100A , 150 mm×4 . 6 mm ) using a mobile phase composed of 25 mM TBA , 5 mM EDTA , 100 mM KH2PO4/K2HPO4 , pH 7 . 0 and 2% methanol ( v/v ) , at a flow rate of 1 ml/min [63] . Different concentrations of cAMP ( 8-Bromoadenosine 3′ , 5′-cyclic monophosphate sodium salt , Sigma-Aldrich , Lyon , France ) were pre-incubated for 30 min with CFSE labeled naïve CD4+CD45RA+CD25− T-cells . Cells were then cultured in 96-well plates and stimulated with 2 µg/ml of coated anti-CD3 and soluble anti-CD28 mAbs for 5 days . At day 2 of culture , cAMP was added in identical concentrations as day 0 . The effect of cAMP on proliferation was evaluated measuring the percentage of CFSElow dividing cells . In some experiments , CD4+CD45RA+CD25− cells were pre-incubated with different reagents: 10 µM adenosine receptor agonist CGS 21680 ( Sigma-Aldrich , Lyon , France ) or 2 µM adenosine receptor antagonist ZM 241385 ( Tocris bioscience , Bristol , UK ) or with 200 µM of 2′ , 5′-DiDeoxyadenosine ( ddADA ) ( Sigma-Aldrich ) or 2 µM of Forskolin , an adenyl cyclase activator ( Sigma-Aldrich ) or DMSO for 30 minutes before activation with 2 µg/mL anti-CD3/CD28 mAbs . Total RNA was isolated from naive and total CD4+ T cells . qRT-PCR was performed using an ABI Prism 7500 Sequence Detection System ( Applied Biosystems , Courtaboeuf , France ) in 50 µL reaction with Platinum SYBR Green qPCR SuperMix-UDG w/ROX ( Invitrogen ) and 0 . 2 µM of each primer . S14 mRNA was used as a control to normalize each sample . Sequences of the IL-2- , A2AR- and S14-specific primers were forward: CGAGGGCTAAGGGCATCATTG , reverse: CTCCTTTGGCTGACCGCAGTT , forward: GGCAGACCGAGATGAATCCTCA , reverse: CAGGTCCAGGGGTCTTGGTCC and forward: GAATCCCAAACTCACCAGGA , reverse: TCAGTTCTGTGGCCTTCTTG respectively . The relative levels of IL-2 and A2AR mRNA were calculated using the 2−ΔΔCTmethod . CD4 T cells from HIV ART naive patients and healthy controls were isolated using negative isolations kits ( Miltenyi Biotec ) . Intracellular cAMP levels were quantified in cell lysates ( 2×105 cells/subject ) using a commercially available assay ( cAMP Direct Biotrak EIA , GE , Healthcare Biosciences , Pittsburgh , PA ) according to the manufacturer's instructions . Anti-CD39-APC ( clone TU66 ) , -CD25-PE , -CD4-FITC , -CD3-Pacific Blue , -IL-2-PE-Cy7 , -CD28-Percp-CY5 . 5 and -CD127-Biot/strepta-APCCy5 . 5 were from BD Biosciences ( Le Pont de Claix , France ) . Anti-CD45RA-ECD was obtained from Beckman Coulter ( Villepinte , France ) and -FoxP3-Alexa 488 was obtained from ebiosciences ( Montrouge , France ) . Cells were analysed on an LSR II ( BD Immunocytometry systems ) . Total DNA was isolated from naive CD4+ T cells using DNeasy Kit DNA extraction kit ( Qiagen , Duesseldorf , Germany ) . Genomic DNA was bisulfite converted using EpiTect Kit ( Qiagen , Duesseldorf , Germany ) according to the manufacturer's instructions . The unique essential CpG site ( site 1 ) in the il-2 gene promoter [32] was amplified by a PCR ( forward GGAAAAATTGTTTTATATAGAAGG , reverse: TTCCTCTTCTAATAACTCTTTAA ) followed by a nested-PCR ( forward GGAAAAATTGTTTTATATAGAAGG , reverse: ATAAATATAAATAAAATCCCTCT ) . A clonal assay was performed for each experimental condition . Briefly , nested-PCR products were used for cloning into a pCR4-TOPO TA plasmid kit ( Invitrogen , Carlsbad , CA , USA ) and transfected in Escherichia coli according to the manufacturer's instructions . Colonies were grown on Luria-Bertani ( LB ) plates overnight at 37°C . Clones screening was done using PureLink Quick Plasmid Miniprep Kit ( Invitrogen , Carlsbad , CA , USA ) and plasmid DNA was prepared for sequencing analysis . Sequence analyses were done with SeqScape software ( Applied Biosystems , Foster City , CA , USA ) on at least 25–30 clones from each sample . The non-parametric Mann-Whitney U , Fisher's exact and paired T tests were used for statistical analyses ( GraphPad Prism 5 . 0 statistical software ) . A P-value <0 . 05 was considered as significant .
|
Regulatory T cells ( Treg ) represent a subset of T lymphocytes and have a pivotal role in chronic viral infections and cancer by limiting immune activation . It has been shown that Treg are expanded in chronic HIV infected patients . However , the mechanisms of Treg immune-modulator functions are not clearly known . CD39 is an ectonucleotidase which converts the proinflammatory ATP signal into AMP and the immunosuppressive adenosine in concert with another ecto-enzyme CD73 . We have previously reported that CD39/adenosine pathway is involved in AIDS progression . However , the mechanism of Treg immunosuppression through CD39 and its involvement in HIV pathogenesis remains unclear . We report here that Treg/CD39+ inhibits the production of IL-2 , a cytokine that stimulates the growth of T lymphocytes , via CD39/Adenosine/cAMP enzymatic pathway . The signals induced by adenosine specific receptor A2AR , increase the intra cellular levels of cAMP . We show that cAMP inhibits CpG site demethylation of the il-2 gene promoter . We found that T cells from HIV patients have a higher expression on A2AR as well as intra-cellular cAMP and a lesser capacity to produce IL-2 upon stimulation than healthy subjects . Our results contribute to elucidate the mechanisms by which Treg suppression occurs during HIV infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"immunopathology",
"medicine",
"infectious",
"diseases",
"immune",
"cells",
"immunity",
"hiv",
"immune",
"activation",
"t",
"cells",
"immunity",
"to",
"infections",
"immunology",
"retrovirology",
"and",
"hiv",
"immunopathogenesis",
"biology",
"viral",
"diseases"
] |
2013
|
Regulatory T Cells Negatively Affect IL-2 Production of Effector T Cells through CD39/Adenosine Pathway in HIV Infection
|
It is widely known that prion strains can mutate in response to modification of the replication environment and we have recently reported that prion mutations can occur in vitro during amplification of vole-adapted prions by Protein Misfolding Cyclic Amplification on bank vole substrate ( bvPMCA ) . Here we exploited the high efficiency of prion replication by bvPMCA to study the in vitro propagation of natural scrapie isolates . Although in vitro vole-adapted PrPSc conformers were usually similar to the sheep counterpart , we repeatedly isolated a PrPSc mutant exclusively when starting from extremely diluted seeds of a single sheep isolate . The mutant and faithful PrPSc conformers showed to be efficiently autocatalytic in vitro and were characterized by different PrP protease resistant cores , spanning aa ∼155–231 and ∼80–231 respectively , and by different conformational stabilities . The two conformers could thus be seen as different bona fide PrPSc types , putatively accounting for prion populations with different biological properties . Indeed , once inoculated in bank vole the faithful conformer was competent for in vivo replication while the mutant was unable to infect voles , de facto behaving like a defective prion mutant . Overall , our findings confirm that prions can adapt and evolve in the new replication environments and that the starting population size can affect their evolutionary landscape , at least in vitro . Furthermore , we report the first example of “authentic” defective prion mutant , composed of brain-derived PrPC and originating from a natural scrapie isolate . Our results clearly indicate that the defective mutant lacks of some structural characteristics , that presumably involve the central region ∼90–155 , critical for infectivity but not for in vitro replication . Finally , we propose a molecular mechanism able to account for the discordant in vitro and in vivo behavior , suggesting possible new paths for investigating the molecular bases of prion infectivity .
Transmissible spongiform encephalopathies ( TSEs ) are progressive and fatal neurodegenerative disorders affecting animals and humans , with the most common forms being scrapie of sheep and goats , bovine spongiform encephalopathy ( BSE ) of cattle , chronic wasting disease ( CWD ) of cervids , and Creutzfeldt-Jakob disease ( CJD ) in humans . TSEs are caused by the misfolding of the host-encoded prion protein monomers ( PrPC ) into autocatalytic self-replicating aggregates ( PrPSc ) . The protein-only hypothesis postulates that the agent responsible for these pathologies , the prion , is exclusively composed of PrPSc [1] . Prion replication can be modelled as a template-based mechanism , where PrPC misfolds into PrPSc and acquires the ability to recruit other PrPC molecules and to trigger their misfolding , through an autocatalytic template-based mechanism [1] . This model is supported by protein misfolding cyclic amplification ( PMCA ) [2] , a technique that mimics in vitro the PrPC-to-PrPSc autocatalytic conversion , leading to the generation of a huge amount of infectious prions in healthy brain homogenates seeded with minute amounts of PrPSc and subjected to multiple cycles of sonication and incubation [3] . Although devoid of a nucleic acid genome , prions exist as strains . Increasing evidence supports the view that prion strains are encoded by distinct PrPSc conformers [4–6] . Strain features are usually kept unchanged on serial passages in the same host; however , seminal studies reported that prion strains can mutate when crossing the species barrier or even on sub-passage in the same host species [7] . The phenomenon of prion mutation has been replicated in ex vivo [8 , 9] and in vitro studies , mimicking inter- [10 , 11] and intra-species [12 , 13] transmission , thus highlighting the dynamic nature of prions in response to modifications in the replication environment . Furthermore , we have recently observed the emergence of mutant PrPSc conformers during in vitro amplifications even in the absence of PrP sequence mismatches , PrPC modifications , RNA-depletion or treatments with drugs [14] , supporting the idea of the quasispecies nature of prions [6 , 15] . Bank voles have been shown to be highly sensitive to sheep scrapie [16 , 17] . Accordingly , bank vole brain homogenates were an efficient substrate for the in vitro amplification of scrapie by PMCA [18] . This prompted us to investigate the sensitivity of bank vole PMCA to scrapie by limiting dilution experiments [17] . During these studies , we found that the propagation at limiting dilutions of a scrapie isolate resulted in strain mutation . Here we report the emergence , the isolation , the biochemical features and the biological characterization of the prion mutant identified . Interestingly , the prion mutant was characterized by an extremely C-terminal PK resistant core ( PrPres ) and arose exclusively from extremely diluted scrapie seeds . Moreover , it could be denoted as a defective prion mutant , since it was efficiently autocatalytic in vitro , but unable to propagate in vivo . These findings indicate that the defective prion mutant , although derived from a natural scrapie isolate , lacks of some essential structural characteristics for being infectious . We propose a molecular mechanism able to account for this discordant in vitro and in vivo behaviour , so as to highlight new paths for investigating the molecular underpinnings of prion infectivity .
To test the sensitivity of detection of serial bank vole PMCA ( bvPMCA ) on natural scrapie isolates , serial ten-fold dilutions of 3 sheep scrapie isolates ( 198/9 , ES47/10/2 and ES47/10/3 ) were subjected to 10 rounds of PMCA ( Fig 1A ) . The sensitivity observed with the whole set of inocula resembled that reported by Chianini et al . [17] , showing a limit of detection of 10−6/10−7 after 8 bvPMCA rounds ( Fig 1A ) . Sixteen unseeded samples were used as negative controls and remained negative up to the 10th round , showing that neither cross-contamination nor spontaneous appearance of PrPSc occurred under these experimental conditions , as previously reported [18] . Moreover , with the only exception of dilution 10−7 of 198/9 ( see below ) , the electrophoretic pattern of the PK-resistant core of the amplified PrPSc was similar to that in the original inocula , as previously observed with other scrapie isolates [17] . We will refer to this pattern as 18K , according to the apparent MW of the unglycosylated PrPres . Unexpectedly , a different PrPSc type was obtained from the dilution 10−7 of 198/9 after 6 PMCA rounds . This atypical PrPSc type , characterized by a PrPres of ~14 kDa ( molecular weight of unglycosylated band ) and thus named 14K , emerged only from the last detectable dilution of sample 198/9 and preserved its biochemical features for two further bvPMCA rounds ( Fig 1A ) . The experiment was repeated two more times , starting from the same three scrapie seeds . Again , this atypical PrPSc type emerged only from the last detectable dilutions of 198/9 , respectively 10−5 after 7 rounds of PMCA and 10−7 after 4 rounds . In order to exclude that the atypical PrPSc could have arisen by the exposure of bank vole substrates to diluted sheep brain seeds , serial ten-fold dilutions , from 10−5 to 10−7 , of brain homogenates from 3 healthy sheep were subjected to 7 rounds of PMCA . None of these negative seeds induced the amplification of 18K or 14K PrPSc ( S1 Fig ) . We selected two of the 14K samples obtained from the isolate 198/9 derived from the dilution 10−7 after 8 bvPMCA rounds ( named 14K/1 ) and the dilution 10−5 after 11 rounds ( named 14K/2 ) , as well as an 18K sample derived from the dilution 10−5 after 6 bvPMCA rounds from the same scrapie isolate . The three samples were subjected to serial bvPMCA in order to produce sufficient amount of material for further analyses , and they all confirmed to be autocatalytic in vitro . However , while 18K and 14K/2 preserved faithfully their biochemical characteristics , 14K/1 gradually shifted to an 18K profile ( Fig 1B ) , being composed of both 14K and 18K PrPres after 2 bvPMCA rounds and of only 18K PrPres after 2 other rounds , as revealed by deglycosylation studies ( Fig 1B ) . The reappearance of 18K from 14K/1 could be explained by the presence of minimal amounts of 18K in the sample , which outcompeted 14K during successive PMCA rounds . To test this hypothesis , we reasoned that we could get rid of any remaining 18K in 14K/1 by a dilution approach . Indeed , serial 10-fold dilutions of 14K/1 subjected to 4 serial rounds of bvPMCA showed that 18K emerged only from the lowest dilutions , while 14K was preserved from dilution 10−5 to 10−7 ( Fig 2A ) . Moreover , 14K further preserved its signature for other 4 rounds , confirming that 14K/1 was composed of a mixture of a huge quantity of 14K and minute amounts of 18K , that could be selectively removed by dilution . As our aim was to obtain a 14K as pure as possible , we focused our attention on sample 14K/2 , which did not shift to 18K in previous experiments . Five more rounds of bvPMCA did not revealed any 18K profile from sample 14K/2 . To get rid of any possible 18K we then subjected 14K/2 to the same dilution approach used for 14K/1 ( Fig 2B ) . The two last positive dilutions of the 3rd round ( 10−5 and 10−6 ) were thus pooled , diluted 1:100 in vole brain homogenate , split in 28 tubes and subjected to 2 rounds of PMCA , in order to produce a large amount of a pure 14K/2 population . Thus , after a total of 21 rounds of PMCA and a cumulative dilution factor of 10−31 , a “cloned” 14K was obtained from the scrapie isolate 198/9 . We then tested the efficiency of in vitro replication of 18K and 14K/2 , subjecting a serial ten-fold dilution curve of the two samples to bvPMCA amplification . 18K dilutions up to 10−6 were positive after a single round , while the 10−5 dilution of 14K/2 was still negative ( Fig 2C ) . The finding that 18K was ~100 times more efficient than 14K/2 further corroborates our previous hypothesis that in 14K/1 a minimal 18K component outcompeted 14K during serial PMCA . Deglycosylation experiments showed that both 18K and 14K/2 PK-resistant cores were composed of C-terminal , fully glycosylated PrP fragments . Epitope mapping of 18K and 14K/2 showed that 18K was similar to vole-adapted scrapie and it was recognized by the antibodies with epitopes from SAF32 ( directed to the octarepeat region of PrP ) to the C-terminus ( Fig 2D and S2 Fig ) indicating to be composed of the expected ~80–231 PrPres fragment , similar to the PrPres previously observed in sheep and voles with classical scrapie [19 , 20] . In contrast , only the C-terminal antibody SAF84 ( aa 163–169 ) was able to recognize 14K/2 ( Fig 2D and S2 Fig ) . The absence of the Sha31 epitope ( aa 145–152 ) in the PK-resistant core of 14K/2 indicated that PrPSc was cleaved by proteinase K between the aa 152 and 163 , thus spanning aa ~155–231 . Deglycosylation of 18K and 14K/2 confirmed that both PrPres types were composed of single variably glycosylated C-terminal PrP fragments ( S2 Fig ) . We then assessed the PrPSc conformational stability of 18K and 14K/2 by denaturation with increasing GdnHCl concentrations . Again , the two bvPMCA-derived PrPSc types had distinct features , as 14K/2 showed a conformational stability higher than 18K , with GdnHCl values of 2 , 1 and 1 , 3 M respectively ( Fig 2E and 2F ) . Overall , based on the different PK-cleavage site and conformational stability observed , these results imply that PrPSc types with PrPres cores of 18K and 14K are self-replicating bona fide PrPSc aggregates with distinct conformations . We tested the infectivity and the strain properties of bvPMCA-derived PrPSc 18K , 14K/1 and 14K/2 by bioassay in vole , in comparison with the original scrapie isolate 198/9 . The scrapie isolate 198/9 transmitted to 100% of voles with a mean survival time of 167 days post infection ( dpi ) . The survival time shortened to 94 dpi on sub-passage , indicating the existence of a transmission barrier for adaptation of sheep scrapie to voles ( Table 1 ) . The neuropathological phenotype ( Fig 3A ) was indistinguishable from that previously observed in other ARQ/ARQ scrapie isolates from Italy and UK [16] , and all voles showed the expected 18K PrPres profile with no evidence of shorter PrPres fragments ( Fig 3A ) . In previous studies we showed that PrPSc from several Italian classical scrapie isolates displayed a uniform conformational stability , with GdnHCl1/2 values of ~2 M [19 , 20] , which was preserved after transmission in voles [19 , 21] . Accordingly , the conformational stabilities of 198/9 and vole-adapted 198/9 were similar , with GdnHCl1/2 of ~2 M ( Fig 3D ) . BvPMCA-derived 18K was even more efficient than the original scrapie isolate , giving a shorter survival time in the first passage , 117 dpi , probably due to its in vitro adaptation . Indeed , on second passage the survival time showed only a minor shortening and converged to that observed for vole-adapted 198/9 ( Table 1 ) . The neuropathological phenotype matched that observed in voles infected with the scrapie isolate 198/9 , and the 18K PrPres profile was preserved upon in vivo propagation ( Fig 3B ) . In contrast , the conformational stability of the 18K inoculum was not preserved in vivo , as vole-adapted 18K showed a rightward shift in the denaturation curve and a GdnHCl1/2 value of ~2 M ( Fig 3E ) , thus fully converging with vole-adapted 198/9 ( Fig 3D ) . These findings indicate that the same prion strain was isolated by both in vitro and in vivo adaptation of natural scrapie to voles , despite the divergent conformational stabilities of in vitro and in vivo derived PrPSc types . BvPMCA-derived mutant 14K/1 , which supposedly contained a mixture of 14K and 18K , efficiently transmitted in voles too , with a survival time of 154 dpi ( Table 1 ) . The survival time and the neuropathological phenotype at second passage were similar to those observed with 198/9 and 18K ( Fig 3C ) . Furthermore , as previously observed in vitro , the PrPres profile of 14K/1 shifted to 18K during in vivo replication ( Fig 3C ) . In contrast to all other samples , 14K/2 was unable to cause disease in voles , as none of the inoculated animals showed clinical signs up to 697 dpi ( Table 1 ) . All voles sacrificed for intercurrent disease or found dead were negative by neuropathological assessment and by WB for brain PrPSc . The inability of 14K/2 to infect voles was in sharp contrast with its efficient in vitro propagation . We thus determined the kinetics of in vivo PrPSc clearance and replication of intracerebrally inoculated 14K/2 and 18K . Two groups of 8 voles were inoculated with 14K/2 or 18K and 2 voles for each group were sacrificed at different time points ( 0 , 3 , 14 and 52 dpi ) ; brains of voles were homogenized and used as seed for PMCA reactions ( Fig 3F ) . Both 18K and 14K/2 were easily detectable at 0 dpi , barely detectable at 3 dpi and become undetectable at 14 dpi , indicating a similar trend of clearance for both PrPSc types . However , at 52 dpi 18K became strongly positive , indicating an active replication in vole brain , while 14K/2 remained undetectable . We further attempted to detect 14K/2 seeding activity at later time points , by using the brain of 9 voles inoculated with 14K/2 and sacrificed for intercurrent disease between 168 and 697 dpi , as well as the spleen of 2 voles sacrificed at 588 and 697 dpi . Again all tissues were negative when tested in a 3-rounds serial bvPMCA able to detect very low levels of PrPSc . These findings imply that 14K/2 , containing the “cloned” 14K PrPSc , not only did not induce disease in voles , but was unable to replicate in vivo at levels sufficient to self-sustain in brain and spleen tissues .
In agreement with previous observations [17] , our findings show that sheep PrPSc is able to efficiently propagate on vole PrPC by PMCA , even when seeded at extremely high dilutions ( up to 10−7 ) , allowing to derive in vitro vole-adapted prion populations starting from a seemingly low number of replicative units . As expected , after several rounds of PMCA we usually recovered a faithful vole-adapted scrapie PrPSc , characterized by a PrPres of 18 kDa; however , from high dilutions of a single sheep isolate , 198/9 , we repeatedly derived a PrPSc with a shorter PK-resistant core of 14 kDa . We selected and studied two prion populations , named 18K and 14K/2 , both derived in vitro by serial bvPMCA , starting from highly diluted seeds of the same scrapie isolate . PrPres from 18K and 14K/2 populations had different PK-cleavage site and different conformational stability , so that they could be seen as different bona fide PrPSc conformers , potentially encoding for different biological properties . Indeed , 18K was competent for in vivo replication and resulted in a pathological phenotype indistinguishable from vole-adapted 198/9 while , unexpectedly , 14K/2 was unable to infect voles . These results suggest that the in vitro derived prion population containing 18 kDa PrPres aggregates encoded for the faithful scrapie strain , while that exclusively composed of 14K kDa PrPres aggregates behaved like a defective mutant , being unable to replicate in live animals . It is of note , however , that even 18K did not preserve 100% fidelity compared to its in vivo counterpart , sheep and vole-adapted scrapie; indeed , the conformational stability of sheep and vole-adapted PrPSc was higher than that of 18K , which however reverted to the original phenotype upon in vivo propagation . Several experimental approaches have been used to investigate the phenomenon of prion mutation . The derived observations gave rise to two different theories , not necessarily mutually exclusive: the “deformed templating model” by Ilia Baskakov and colleagues which postulates that the templating process is imperfect and changes in the replication environment play an active role in the de novo generation of new PrPSc variants [22–24] , and the “cloud” model postulated by Charles Weissman and John Collinge [6 , 9 , 15 , 25 , 26] which proposes that even cloned strains consist of a cloud of different PrPSc conformers and that changes in the replication environment give selective advantage to the “fittest” among pre-existing variants . The first hypothesis would suggest that the 14K mutant could have been somehow induced by the in vitro replication environment of PMCA , while the latter hypothesis implies that the original scrapie isolate was already composed of a cloud of PrPSc conformers , then subjected to evolutionary constraints in the new in vitro replication environment . Whatever the molecular mechanism that underlies prions mutation and evolution , it looks undeniable that prions can adapt and evolve when propagated under particular selection regimes . Our findings are in line with previous observations and corroborate the notion that an agent that apparently lacks any nucleic acid information is still able to adapt in response to changes in the replication environment . Moreover , the fact that the mutant PrPSc emerged exclusively and repeatedly from the highest dilution of the isolate 198/9 further validates the hypothesis that the starting population size may have a profound impact on the evolutionary landscape of scrapie , at least in vitro [14] . These observations can be supportive of the “cloud” hypothesis , i . e . that prion populations behave as quasispecies . Indeed , in viral quasispecies , successive bottleneck passages mediate the surfacing of minority components present or generated in the mutant spectra of the original population . This is because mutant spectra are not mere distributions of neutral mutants , but they can hide components that in isolation would display dissimilar biological properties [27] . In this scenario , the starting population size may affect the evolutionary outcome of a given prion population under a new selective environment , such as a new host or any other in vivo or in vitro evolutionary constraint , which has important implications for understanding the inter- and intra-host biological variability of prions . In previous studies , ex vivo and in vitro selected prion mutants had the tendency to revert to the parental strain upon in vivo propagation [9 , 28] . In the present study , one mutant population , 14K/1 , reverted to the parental strain when propagated in vivo , while the other , 14K/2 , did not and was unable to propagate at all . Overall , our and previous studies suggest that most in vitro selected prion mutants may have a lower fitness than their parental strains for in vivo propagation . As far as mutants still contain minor populations of the parental strains , these would then easily reemerge in vivo . Our results strongly support this hypothesis , by showing i ) that minor populations of 18K were actually present in the mutant 14K/1 , which reverted to the parental strain in vivo , and ii ) that 14K populations propagated in vitro under conditions able to get rid of any 18K ( Fig 2 ) were indeed defective for in vivo propagation . It seems difficult to reconcile the findings that 14K PrPSc was suitably competent for autocatalytic self-propagation in vitro , but lacked any infectivity in vivo . In vitro studies with recombinant PrP have shown the spontaneous emergence of autocatalytic but noninfectious recombinant PrP fibrils characterized by ~160–231 C-terminal PK-resistant core [29–31] . Compelling evidences suggested that PrP fibrils with more extended C-terminal PK-resistant core is required for fully competent in vivo replication [28 , 29] . The present findings show that a similar phenomenon could be observed with natural scrapie and suggest that autocatalytic but noninfectious PrPSc is not merely an artificial product of recombinant PrP , but can derived from infectious PrPSc under experimental conditions able to replicate faithful and infectious prions . Usually , full infectious PrPSc contains a PK-resistant core spanning residues ~ 90–231 , which characterizes most known natural TSEs . We have recently found that PrPSc with PK-resistant core as short as that spanning residues ~ 100–145 [21] may still be highly infectious [32] . In contrast , all the above mentioned “defective” prions have in common a shorter than usual C-terminal PrPres , which do not include the central PrP domain . This PrP domain contains the polybasic lysine cluster ( residues 101–110 , vole PrP numbering ) , which has been suggested as a key structural modulator in the conversion of PrPC to PrPSc , either by binding to anionic cofactors that promote prion replication [31] or by mediating proper PrPC/PrPSc interaction [33] . Interestingly , PrPSc aggregates that lack the central polybasic domain showed to be autocatalytic in vitro but barely infectious in vivo [33] . The dominant PrPC processing event , α-cleavage , occurs at the start of the hydrophobic core region , at approximately K110↓H111 ( vole PrP numbering ) [34 , 35] , producing the membrane anchored C-terminal C1 fragment and releasing the corresponding N-terminal N1 fragment , which contains the polybasic domain . This cleavage event is prevented in infectious PrPSc , as the α-cleavage site is in the tightly packed PK-resistant core and is solvent excluded , but it is allowed in PrPSc with less extended ~ 155–231 PK-resistant core , in which residues 110–111 are solvent exposed and available to the endoproteolytic processing . Interestingly , noninfectious PrPSc with ~ 155–231 PK-resistant core have only been observed in vitro under conditions in which α-cleavage do not occur , i . e . in experiments involving purified preparations of recombinant PrP or in brain homogenates supplemented with protease inhibitors , that are routinely added in PMCA reactions to prevent PrPC proteolysis . This observation prompted us to hypothesize that preserving PrPSc from α-cleavage could be indispensable for prion replication . In Fig 4 we describe a molecular mechanism based on this hypothesis , which is able to reconcile the contrasting in vitro and in vivo behavior of 14K PrPres . The model proposes that the short 100–110 polybasic peptide containing the lysine cluster plays a crucial role in prion replication and that the infectivity of 14K PrPSc was prevented by the removal of the polybasic domain through in vivo α-endoproteolysis , which is structurally inhibited in infectious 18K PrPSc . Overall , taking into account the hypothesis of the quasispecies nature of prion strains , we might suppose that even defective PrPSc conformations such as 14K could be continuously generated and eliminated by the host endoproteolitic cleavage , i . e . the in vivo negative selection . Once the negative selection is abrogated , as occurs in PMCA experiments in which brain homogenates are supplemented with protease inhibitors , such innate continuous generation of PrPSc conformational variants might eventually result in the selective emergence of defective mutants . In conclusion , several studies have been conducted to understand what PrPSc needs to be a prion; however , most , if not all , of these studies were built on recombinant PrP aggregates and , as such , have to deal with the community skepticism towards synthetic prions . To our knowledge , our findings provide the first “authentic” defective prion mutant , composed of brain-derived PrPC and originating from a natural scrapie isolate . Our findings could provide new insights for dissecting the molecular mechanisms that differentiate autocatalytic PrPSc amplification from in vivo prion replication .
Bank voles carrying methionine at codon 109 were obtained from the breeding colony at the Istituto Superiore di Sanità ( ISS ) . Experiments involving animals adhered to the guidelines contained in the Italian Legislative Decree 116/92 , which transposed the European Directive 86/609/EEC on Laboratory Animal Protection , and then in the Legislative Decree 26/2014 , which transposed the European Directive 2010/63/UE on Laboratory Animal Protection . The research protocol was performed under the supervision of the Service for Biotechnology and Animal Welfare of the ISS , and was approved by the Italian Ministry of Health ( decree number 84/12 . B ) . Substrates were prepared according to the protocol previously reported [18] . Brain tissues from three ARQ/ARQ scrapie infected sheep ( 198/9 , ES47/10/2 and ES47/10/3 ) , from three ARQ/ARQ healthy sheep and from voles inoculated with 18K or 14K/2 , or spleen tissues from voles inoculated with 14K/2 , were homogenized in PBS ( 10% w/v ) containing Complete Protease Inibitor Cocktail ( Roche ) , divided into small aliquots and stored at -20°C . Serial ten-fold dilutions of the homogenates ( a single tube per dilution from 10−2 to 10−8 , where 10−2 means 1% w/v of infected brain homogenate ) were prepared in PMCA substrate and were subjected to serial PMCA . PMCA was performed using the Misonix S3000 sonicator , following the procedures described in Cosseddu et al . [18] . The in vitro amplification was performed in a total volume of 50 μl for each dilution , and the sonication program consisted of 20 seconds sonication pulses every 30 minutes for 48 hours , at a constant temperature of 37°C . At the end of each round , 5 μl of each reaction mix were diluted 1:10 in fresh substrate for a new amplification round . PMCA reactions were added with an equal volume of Tris-HCl sarcosyl 4% and then digested with 100 μg/ml of proteinase K ( Sigma-Aldrich ) for diagnosis and 200 μg/ml for epitope mapping . Samples were shaken at 750 rpm for one hour at 55°C . The digestion was then stopped by adding 3 mM PMFS ( Sigma ) . For epitope mapping , PK digested samples were added with an equal volume of isopropanol/butanol ( 1:1 v/v ) and centrifuged at 20 , 000 g for 5 min . Supernatants were discarded and the pellets were suspended in denaturing sample buffer ( NuPAGE LDS Sample Buffer , Invitrogen ) and heated for 10 min at 90°C . Electrophoresis and Western blotting were performed as previously described [21] . The membranes were then analyzed with anti-PrP monoclonal antibody SAF84 ( aa 167–173; 1 , 2 μg/ml ) , Sha31 ( aa 146–152; 0 , 6 μg/ml ) , 9A2 ( aa 99–101; 1 μg/ml ) , and SAF32 ( octarepeat; 4 , 8 μg/ml ) . Following incubation with horseradish peroxidase-conjugated anti-mouse immunoglobulin ( Pierce Biotechnology ) at 1:20000 , the PrP bands were detected by enhanced chemiluminescent substrate ( SuperSignal Femto , Pierce ) and VersaDoc imaging system ( Bio-Rad ) . The chemiluminescence signal was quantified by QuantityOne software ( Bio-Rad ) . Deglycosylation was performed by adding 18 μl of 0 . 2 M sodium phosphate buffer ( pH 7 . 4 ) containing 0 . 8% Nonidet P40 ( Roche ) and 2 μl ( 80 U/ml ) di N-Glycosidase F ( Roche ) to 5 μl of PK-digested and denaturated samples and by incubating overnight at 37°C with gentle shaking . Samples were then analysed by Western blotting as described above . To normalize the buffer condition of all the samples ( in vivo and in vitro produced ) subjected to CSA , brains from voles inoculated with 198/9 and 18K and from sheep 198/9 were homogenized in conversion buffer ( PBS 1x , pH 7 , 4; 0 , 15 M NaCl; 1% Triton X with the Roche Complete Protease Inibitor Cocktail ) . The CSA was performed as described [19] . Briefly , brain homogenates and PMCA-derived products ( both 6% w/v ) were added with an equal volume of TrisHCl 100 mM ( pH 7 . 4 ) containing sarkosyl 4% and incubated for 1h at 37°C with gentle shaking . Aliquots of 25 μl were added with 25 μl of GdnHCl to give a final concentration ranging from 0 to 4 . 0 M . After 1 h of incubation at 37°C all samples were diluted to a final concentration of 0 . 4 M GdnHCl and then PK digested ( 50 mg/ml PK final concentration ) for 1 hour at 55°C and gentle shaking . The reaction was stopped with 3 mM PMSF ( Sigma ) . Aliquots of samples were added with an equal volume of isopropanol/butanol ( 1 1 v/v ) and centrifuged at 20000 g for 10 min . Pellets were re-suspended in NuPage LDS Sample Buffer ( Invitrogen ) and were analysed by Western Blotting as described above . Brain tissues were homogenized at 10% ( w/v ) in phosphate buffered saline ( PBS ) and stored at -80°C . PMCA-derived inocula were diluted with PBS for a final concentration of 1% ( w/v ) and stored at -80°C . Groups of eight-week-old voles were inoculated intracerebrally with 20 μl of homogenate into the left cerebral hemisphere , under ketamine anaesthesia ( ketamine 0 . 1 μg/g ) . All animals were individually identified by a passive integrated transponder . The animals were examined twice a week until neurological signs appeared , after which they were examined daily . Diseased animals were culled with carbon dioxide at the terminal stage of the disease , but before neurological impairment was such as to compromise their welfare , in particular their ability to drink and feed adequately . Survival time was calculated as the interval between inoculation and culling or death . After sacrifice , the brain from each animal was removed and cut sagittally into two parts: one stored at −80°C and one fixed in formalin . Histological assessment was performed on formalin-fixed tissues as previously described [36] . Briefly , brains were trimmed at standard coronal levels , embedded in paraffin wax , cut at 6 μm and stained with haematoxylin and eosin . Vacuolar changes were scored in nine grey-matter areas of the brain on haematoxylin and eosin-stained sections , as previously described [36] . Vacuolation scores are derived from at least six individual voles per group and are reported as means ± standard error of the mean .
|
Prions are unique infectious agents , consisting of PrPSc , a self-propagating aggregated conformer of the host-encoded prion protein PrPC . Despite the absence of any nucleic acid information , prions exist as distinct strains that share the same amino acid sequence but differ in their conformation . Moreover , prions can mutate and are thus heterogeneous populations able to evolve and adapt to new replication environments . During in vitro amplification of sheep scrapie , we found that a prion mutant could be obtained from one natural isolate . The prion mutant identified was characterized in vivo and in vitro , showing unusual biochemical and biological features: a smaller than usual C-terminal proteinase resistant core of PrPSc , which spans aa ∼155–231 , and the inability to propagate in vivo despite an efficient autocatalytic replication in vitro . With such a signature , we denoted the mutant as a “defective” prion mutant . We thus postulate a new hypothesis for the discrepancy between the in vitro and in vivo behavior of the defective mutant and suggest that the central PrPSc domain ∼90–160 might have a key role in prion replication . This work provides important new insights into the mechanism underpinning prion replication and has numerous implications for understanding the molecular requirements indispensable for prion infectivity .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
] |
[
"livestock",
"organismal",
"evolution",
"animal",
"diseases",
"medicine",
"and",
"health",
"sciences",
"ruminants",
"enzymes",
"italian",
"people",
"epitope",
"mapping",
"enzymology",
"vertebrates",
"microbiology",
"animals",
"mammals",
"ethnicities",
"animal",
"prion",
"diseases",
"microbial",
"evolution",
"molecular",
"biology",
"techniques",
"zoology",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"zoonoses",
"sheep",
"gene",
"mapping",
"proteins",
"molecular",
"biology",
"voles",
"viral",
"evolution",
"agriculture",
"people",
"and",
"places",
"biochemistry",
"rodents",
"virology",
"biology",
"and",
"life",
"sciences",
"proteases",
"scrapie",
"population",
"groupings",
"evolutionary",
"biology",
"amniotes",
"organisms",
"prion",
"diseases"
] |
2016
|
Isolation of a Defective Prion Mutant from Natural Scrapie
|
Schnyder corneal dystrophy ( SCD ) is a rare genetic eye disease characterized by corneal opacification resulted from deposition of excess free cholesterol . UbiA prenyltransferase domain-containing protein-1 ( UBIAD1 ) is an enzyme catalyzing biosynthesis of coenzyme Q10 and vitamin K2 . More than 20 UBIAD1 mutations have been found to associate with human SCD . How these mutants contribute to SCD development is not fully understood . Here , we identified HMGCR as a binding partner of UBIAD1 using mass spectrometry . In contrast to the Golgi localization of wild-type UBIAD1 , SCD-associated mutants mainly resided in the endoplasmic reticulum ( ER ) and competed with Insig-1 for HMGCR binding , thereby preventing HMGCR from degradation and increasing cholesterol biosynthesis . The heterozygous Ubiad1 G184R knock-in ( Ubiad1G184R/+ ) mice expressed elevated levels of HMGCR protein in various tissues . The aged Ubiad1G184R/+ mice exhibited corneal opacification and free cholesterol accumulation , phenocopying clinical manifestations of SCD patients . In summary , these results demonstrate that SCD-associated mutations of UBIAD1 impair its ER-to-Golgi transportation and enhance its interaction with HMGCR . The stabilization of HMGCR by UBIAD1 increases cholesterol biosynthesis and eventually causes cholesterol accumulation in the cornea .
Schnyder corneal dystrophy ( SCD ) is a rare autosomal dominant genetic eye disease [1] . It is characterized by free cholesterol accumulation in the cornea that causes progressive corneal opacification with aging [1] [2] . Genetics studies have linked SCD to mutations in UbiA prenyltransferase domain-containing protein-1 ( UBIAD1 ) , also known as transitional epithelial response gene 1 ( TERE1 ) [3–5] . Until now , 25 missense mutations altering 21 amino acids , including N102S and G186R ( equivalent to N100S and G184R in mouse Ubiad1 , respectively ) , have been identified in about 50 SCD families [6–8] . However , the causal relationship between UBIAD1 mutations and SCD development has not been proved until very recently [9] . The UBIAD1 protein belongs to the UbiA superfamily of intramembrane aromatic prenyltransferases , which catalyze the biosynthesis of a variety of lipophilic molecules such as ubiquinones , vitamin K , vitamin E , chlorophylls , hemes and archaeal tetraether lipids [10] . UBIAD1 has been identified as a vitamin K2 biosynthesis enzyme in humans and mice , and is essential for mouse embryonic development [11 , 12] . In Drosophila , the UBIAD1 homolog was reported to generate vitamin K2 that functions as an electron carrier for sustaining mitochondrial function [13] . In zebrafish , UBIAD1 was proposed to synthesize ubiquinone coenzyme Q10 that protects against oxidative damages through regulating nitric oxide activity , thereby maintaining vascular endothelial cell survival [14] . However , how UBIAD1 mutations cause cholesterol accumulation in the cornea is not fully understood . The endoplasmic reticulum ( ER ) -localized 3-hydroxy-3-methyglutaryl coenzyme A reductase ( HMG-CoA reductase , HMGCR ) is a rate-limiting enzyme of the cholesterol biosynthetic pathway catalyzing the synthesis of mevalonate [15] . As an important intermediate , mevalonate gives rise to not only cholesterol , but also nonsterol isoprenoids such as farnesyl pyrophosphate and geranylgeranyl pyrophosphate ( GGPP ) that further generates ubiquinone , dolichol and hemes [15] . Interestingly , these mevalonate-derived molecules are also the final products of UBIAD1 superfamily members , suggesting a potential link between HMGCR and UBIAD1 . Theses sterols and nonsterol isoprenoids coordinate to accelerate degradation of HMGCR to prevent over-accumulation of cholesterol [16] . Accumulating sterols induce HMGCR binding to ER-anchored Insig-1 and Insig-2 , which bring ubiquitin ligases gp78 , TRC8 and RNF145 together with other cofactors to ubiquitinate HMGCR , resulting in its degradation in proteasome [17–22] . Elucidating the molecular pathway of HMGCR degradation is of potential clinic significance . Statins as competitive inhibitors of HMGCR can decrease the synthesis of sterols and nonsterol isoprenoids and dramatically increase HMGCR in the liver [23 , 24] . The lanosterol analog HMG499 ( also named Cmpd 81 ) can potentiate the cholesterol-lowering effect of statins through inducing HMGCR degradation [25] . Recently , Ubiad1 was found to be a HMGCR-binding protein through proximity-dependent biotinylation using HMGCR as the bait [26] . The homozygous Ubiad1 N100S knock-in ( Ubiad1N100S/N100S ) mice exhibit corneal opacification and HMGCR accumulation in tissues [9] . In this study , we identified HMGCR as a UBIAD1-associated protein through UBIAD1 immunoprecipitation coupled with mass spectrometry . We demonstrated that UBIAD1 competed with Insig-1 for binding HMGCR , preventing the latter from ubiquitination and degradation . All known SCD-associated UBIAD1 mutants localized in the ER and stabilized HMGCR protein . More importantly , we generated a different SCD-associated knock-in mouse line carrying Ubiad1 G184R mutation . The Ubiad1G184R/+ mice exhibited excess HMGCR protein in tissues and striking corneal opacifications with free cholesterol deposition . These phenotypes recapitulate clinical manifestations of human SCD patients , suggesting that Ubiad1G184R/+ mouse is an ideal model to study human SCD .
To explore the underlying connections between UBIAD1 and cholesterol metabolism , we performed a tandem affinity purification ( TAP ) coupled to mass spectrometry to identify UBIAD1-associated proteins , using HEK-293 cells stably expressing human wild-type ( WT ) or G186R mutant form of UBIAD1 fused with a TAP tag at the C terminus ( S1A Fig ) . The G186R mutation is a mutation found in an early-onset SCD family [27] . Besides , according to the determined structure of archaeal UbiA , the G186 residue was proposed to locate in the surface-exposed loop , and the G186R mutation may affect the interactions with other proteins [10 , 28] . We found three proteins among the top of the list: HMGCR , VCP/p97 and SEL1L ( S1B and S1C Fig , S1 Table ) . HMGCR is the rate-limiting enzyme converting HMG-CoA to mevalonate in the cholesterol biosynthetic pathway [15] . VCP/p97 and SEL1L have been known to be involved in the degradation of HMGCR and other ER proteins [29–31] . Therefore , we focused on the effects of UBIAD1 on sterol-induced degradation of HMGCR protein . The cells stably expressing WT UBIAD1 or G186R mutant were treated with increasing concentrations of 25-hydroxycholesterol ( 25-HC ) for 5 h . The endogenous HMGCR was degraded in a concentration-dependent manner in cells stably expressing UBIAD1 ( WT ) . However , 25-HC failed to induce HMGCR degradation in the UBIAD1 ( G186R ) -expressing stable cells ( Fig 1A ) . In addition , the UBIAD1 ( G186R ) mutation had no obvious effect on SREBP-2 processing ( Fig 1A ) . We next validated these findings using co-transfection experiments . Co-expression of HMGCR with Insig-1 conferred sterol-regulated degradation of endogenous HMGCR , as Insig-1 is a rate-limiting co-factor for multiple E3s including gp78 , TRC8 and RNF145 [18–22] ( Fig 1B , lanes 1–2 ) . The WT form of mouse Ubiad1 had little effect on HMGCR degradation , whereas Ubiad1 ( G184R ) almost completely blocked the degradation ( Fig 1B , lanes 3–6 ) . The K89 and K248 are two ubiquitination sites of HMGCR [16] , and their mutations should abolish sterol-induced degradation of HMGCR . Neither Ubiad1 ( WT ) nor Ubiad1 ( G184R ) increased the amount of HMGCR ( K89R , K248R ) ( Fig 1B , lanes 7–12 ) . The amount of Ubiad1 ( G184R ) was about 2-fold that of Ubiad1 ( WT ) when equivalent amounts of plasmids were used . We next analyzed the effect of Ubiad1 on HMGCR ubiquitination . 25-HC triggered pronounced ubiquitination of immunoprecipitated HMGCR in the presence of proteasome inhibitor MG-132 , which , however , was markedly reduced by WT Ubiad1 ( Fig 1C , second panel , compare lane 2 and 4 ) . The Ubiad1 ( G184R ) further inhibited the ubiquitination of HMGCR ( Fig 1C , second panel , compare lane 4 and 6 ) . The total amount of HMGCR from input was more abundant in Ubiad1 ( G184R ) -expressing cells than those expressing Ubiad1 ( WT ) ( Fig 1C , third panel , lanes 3–6 ) , consistent with results in Fig 1A and 1B . We then used co-immunoprecipitation experiments to address potential interaction between HMGCR , Insig-1 and Ubiad1 . The results showed that HMGCR pulled down more amounts of Ubiad1 ( G184R ) than Ubiad1 ( WT ) ( Fig 1D ) . Meanwhile , Insig-1 association with HMGCR was reduced ( Fig 1D ) . Together , these results suggest that SCD-associated mutant Ubiad1 competes with Insigs to bind HMGCR , thereby blocking Insig-mediated ubiquitination and degradation of HMGCR . We then analyzed whether SCD-associated UBIAD1 ( G186R ) had any effect on cellular cholesterol level . Using the colorimetric method , we measured the amount of total cholesterol in cells stably expressing WT or G186R form of UBIAD1 . UBIAD1 ( G186R ) -expressing cells indeed had more cholesterol than control ( Fig 2A ) . However , no difference in the amount of nonesterified fatty acid ( NEFA ) was detected between these two cell lines ( Fig 2B ) . We next sought to determine whether the G186R mutation increased de novo synthesis of cholesterol by using [14C]-acetate to label newly synthesized cholesterol and fatty acid . [14C]-cholesterol was markedly increased in UBIAD1 ( G186R ) -expressing cells ( Fig 2C ) , whereas [14C]-fatty acid remained comparable between WT- and UBIAD1 ( G186R ) -expressing cells ( Fig 2D ) . These results together suggest that the UBIAD1 ( G186R ) mutation enhances synthesis of cholesterol . Human UBIAD1 contains 338 amino acids with 8 predicated transmembrane helices , and 21 SCD-associated UBIAD1 nucleotide mutations in the coding sequence that altered amino acids at 19 positions are shown in Fig 3A . Protein sequence analysis showed that amino acids that are mutated in SCD are evolutionarily conserved from fly to human ( S2 Fig ) . As UBIAD1 ( G186R ) dramatically stabilized HMGCR and increased cellular cholesterol level ( Fig 1 , Fig 2 ) , we next examined the effect of other SCD-associated UBIAD1 mutations on sterol-induced degradation of HMGCR . Each of the 21 SCD-associated missense mutations of UBIAD1 were co-transfected with HMGCR and Insig-1 expression plasmids individually followed by 25-HC treatment for 5 h . HMGCR protein was reduced by 80% in cells expressing the WT form of human UBIAD1 after receiving 25-HC treatment . However , 25-HC-induced HMGCR degradation was only reduced by 10% to 50% in cells expressing any of the 21 UBIAD1 mutations ( Fig 3B–3F , top panel ) , indicating that HMGCR protein can be stabilized by SCD mutants of UBIAD1 . Interestingly , the protein levels of SCD-associated UBIAD1 mutants were all substantially increased without being affected by 25-HC , even though the cells were transfected with the same amount of plasmids ( 100 ng per dish ) ( Fig 3B–3F , second panel ) . The protein levels of UBIAD1 harboring SCD-associated mutations were 1 . 9 to 9 . 5-fold more than WT UBIAD1 ( Fig 3B–3F , second panel ) , while the Insig-1 protein had no changes ( Fig 3B–3F , second panel ) . These results suggest that the SCD mutations of UBIAD1 render HMGCR proteins more resistance to sterol-induced degradation . We next examined the cellular localization of WT UBIAD1 and UBIAD1 harboring SCD mutations in CHO-K1 cells . Immunofluorescence experiments revealed that the ectopically expressed WT UBIAD1 preferentially co-localized with the Golgi marker GM130 ( Fig 4 ) . However , the SCD-associated UBIAD1 mutants had a diffused distribution and co-localized with the ER marker calnexin ( Fig 4 and S3 Fig ) . We further analyzed the distribution of WT and G186R UBIAD1 under different conditions . S4 Fig showed that WT UBIAD1 presented in Golgi in FCS condition and in ER in sterol-depletion condition . Nonsterol isoprenoid geranylgeraniol ( GGOH ) , but not cholesterol or 25-HC , relocated WT UBIAD1 to Golgi ( S4A and S4C Fig ) . However , the G186R mutant primarily located in the ER under different conditions , and seems to be trafficking-deficient ( S4B and S4D Fig ) . We next analyzed the effects of 25-HC and GGOH on the association between HMGCR and UBIAD1 . Results of S5A Fig showed that GGOH could further degrade HMGCR with 25-HC in WT Ubiad1 expressing cells , and had minimal effect in G184R mutant cells ( S5A Fig ) . Co-immunoprecipitation results in S5B Fig showed that 25-HC stimulated HMGCR to bind more WT Ubiad1 , and GGOH reduced the interaction between HMGCR and WT Ubiad1 . However , the association of HMGCR and G184R mutated Ubiad1 was stronger than WT Ubiad1 , and did not response to 25-HC and GGOH treatments ( S5B Fig ) . Therefore , the SCD-associated mutations may impair ER-to-Golgi transportation of UBIAD1 and enhance UBIAD1-HMGCR interaction to prevent Insig-mediated degradation of HMGCR . Although many missense mutations of UBIAD1 have been found in SCD patients , the causal relationship between these two remains to be proved . To further investigate the consequence of SCD-associated mutation of UBIAD1 in vivo , we generated Ubiad1 G184R ( corresponding to G186R in human ) knock-in mice using a knockout-first conditional ready strategy as shown in Fig 5A . The targeted allele was knocked out and conditional ready , and the mice were first crossed with Flp recombinase-expressing strain to generate the floxed allele that expressed WT Ubiad1 . The mice were then crossed with EIIA-Cre recombinase transgenic mice to get the whole-body knock-in mice ( Fig 5A ) . The WT mice had a single band of 400 bp , while the heterozygous knock-in mice ( Ubiad1G184R/+ ) exhibited bands at both 400 bp and 500 bp ( Fig 5B ) . Further sequencing of these two bands confirmed that the mutated band indeed carried the GGA-to-AGA mutation ( Fig 5C ) . No homozygous knock-in ( Ubiad1G184R/G184R ) mice were generated when the heterozygotes were intercrossed ( Fig 5D ) . These results are not surprising as Ubiad1 knockout mice defective in vitamin K2 synthesis are embryonic lethal as well [12] . The WT and heterozygous knock-in mice ( Ubiad1G184R/+ ) were born at an expected Mendelian ratio ( 34% vs 66% ) ( Fig 5D ) . Ubiad1G184R/+ knock-in mice appeared indistinguishable from WT littermates , and both had similar body weights even at the average age of 105-week-old ( Fig 5E ) . Given that systemic hypercholesterolemia has been reported in some but not all SCD patients [1] , we next sought to evaluate cholesterol levels of WT and Ubiad1G184R/+ mice . The serum levels of total cholesterol ( TC ) , high-density-lipoprotein ( HDL ) and low-density-lipoprotein ( LDL ) cholesterol in Ubiad1G184R/+ knock-in mice were similar to those of WT mice ( Fig 5F , 5G and 5H ) . The serum triglyceride ( TG ) level of Ubiad1G184R/+ mice was slightly , but not significantly , decreased relative to WT mice ( Fig 5I ) . Collectively , these results suggest that the G184R mutation of Ubiad1 does not affect systemic cholesterol and triglyceride levels . Since all SCD-associated point mutations including G186R impaired sterol-induced degradation of HMGCR and caused accumulation of HMGCR ( Fig 1 , Fig 3 ) , we prepared mouse embryonic fibroblast ( MEF ) cells from Ubiad1G184R/+ mice and their WT littermates and analyzed HMGCR degradation . Compared with WT MEF cells , endogenous HMGCR in Ubiad1G184R/+ MEFs was partially resistant to ubiquitination and degradation induced by 25-HC ( Fig 6A , S6 Fig ) , similar to the overexpression results ( Fig 1A–1C ) . We next measured whether the G184R mutation would alter HMGCR protein levels in different mouse tissues . Strikingly , the amount of HMGCR was dramatically higher in the liver ( 2 . 5-fold ) , pancreas ( 12 . 1-fold ) , lung ( 4 . 5-fold ) , spleen ( 4 . 2-fold ) and cornea ( 5 . 7-fold ) of Ubiad1G184R/+ mice than WT mice ( Fig 6B–6F ) . Together , these results indicate that Ubiad1 ( G184R ) protects HMGCR from degradation and increases HMGCR in various tissues . Next , we examined mouse eyes using stereomicroscope . Prominent signs of corneal opacifications were found in 64% ( 21/33 ) of both male ( 11/17 ) and female ( 10/16 ) Ubiad1G184R/+ mice at 102-to-108 weeks of age ( about 2 years , equivalent to the human age of 70 ) ( Fig 7A , bottom panel ) . There was no difference between male and female mice . These corneal opacifications were haze-like and quite similar to human SCD ( Fig 7A , bottom panel ) . As controls , none of the 17 aged WT littermates showed corneal opacification ( Fig 7A , top panel ) . To further characterize this corneal opacification , the corneal section from these aged mice were stained with Filipin , a specific antibiotic binding to free cholesterol [32] . Free cholesterol in the WT cornea mainly localized to the epithelial cells and sporadically to the stromal cells as well ( Fig 7B , top panel ) . However , in Ubiad1G184R/+ cornea the Filipin signals infiltrated throughout the anterior stroma underneath the epithelium in puncta or patches ( Fig 7B , bottom panel ) , which correspond to the cholesterol crystals observed in the corneas of human SCD patients . Biochemical analysis showed that Ubiad1G184R/+ mice had higher levels of total and free cholesterol in the cornea than WT littermates , although both showed similar corneal TG levels ( Fig 7C–7E ) . In contrast , the levels of total cholesterol , free cholesterol and TG levels in the liver , pancreas , lung and spleen were similar between WT and Ubiad1G184R/+ mice ( S7A–S7L Fig ) . Collectively , these results demonstrate that the G184R mutation of Ubiad1 specifically cause free cholesterol accumulation in the anterior stroma of cornea , phenocopying human SCD .
Based on the above findings , we propose a working model depicting how UBIAD1 mutations cause cholesterol accumulation in the cornea ( Fig 8 ) . Under normal conditions in which the WT form of UBIAD1 mainly localizes in the Golgi , an elevation in sterols triggers HMGCR binding to Insig-1 or Insig-2 , which recruits E3 ubiquitin ligases including gp78 , TRC8 and RNF145 for ubiquitination and proteasomal degradation of HMGCR [18–22] . SCD-associated mutations of UBIAD1 impair its transportation from the ER to Golgi , resulting in a more stable protein in the ER that competes with Insig-1 for binding to HMGCR . As a consequence , sterol-induced ubiquitination and degradation of HMGCR is blocked . The increased cholesterol biosynthesis eventually causes cholesterol accumulation in the cornea in aged mice and humans . The association of UBIAD1 with HMGCR was recently reported by Debose-Boyd and coworkers using HMGCR as a bait [26] . They showed that the SCD-associated mutants of UBIAD1 were sequestered in the ER and protected HMGCR from degradation , leading to the accumulation of HMGCR and cholesterol [26 , 33] . They also found cholesterol accumulation in the cornea of aged Ubiad1N100S/N100S mice , which is another SCD model [9] . Our results and theirs are largely consistent and both support a role of HMGCR-UBIAD1 interaction in HMGCR stabilization . However , we employed an unbiased mass spectrometry analysis of immunoprecipitated UBIAD1 and constructed a different knock-in mouse line ( Ubiad1G184R/+ ) . Considering that all SCD patients are heterozygotes of UBIAD1 mutation , our Ubiad1G184R/+ mouse model may mimic human situation more closely . Indeed , Ubiad1G184R/+ mice displayed HMGCR accumulation in multiple tissues and elderly ones had free cholesterol deposition in the cornea . Notably , corneal opacification and free cholesterol accumulation , albeit prominent in two-year-old Ubiad1G184R/+ heterozygote mice , were barely detected in the 3-month-old mice ( data not shown ) . Consistently , SCD patients carrying heterozygous UBIAD1 mutations display slow progression of corneal opacification with aging . In addition , the homozygous Ubiad1G184R/G184R knock-in mice are embryonic lethal , while Ubiad1N100S/N100S mice are grossly normal . Because Ubiad1 knockout mice fail to survive through development owing to the defects in vitamin K2 synthesis [12] , it is reasonable to speculate that the Ubiad1 G184R mutation may affect vitamin K2 synthesis more profoundly than Ubiad1 N100S . Another intriguing phenomenon is that no obvious differences were detected in the serum and tissue levels of cholesterol and triglyceride between the WT and Ubiad1G184R/+ mice ( Fig 5 , Fig 7 , S7 Fig ) , though HMGCR protein levels in all examined tissues were increased ( Fig 6 ) . Cholesterol accumulation was only detected in the cornea ( Fig 7 ) . Such phenomena are likely attributed to the unique anatomic structure of cornea , which lacks blood-vascular system and is separated from the systemic circulation [34] . Other tissues such as the liver , pancreas , lung and spleen , can intensively exchange lipids with the systemic circulation . In addition , compensatory mechanisms , such as the cholesterol esterification and efflux , may balance the cholesterol level in these tissues . According to the BioGPS expression database , the expression of acyl-CoA: cholesterol acyltransferases ( ACAT ) -1 and ACAT-2 , which convert free cholesterol to cholesteryl ester for storage or secretion as lipoproteins [35] , and ABCA1 , which is an essential transporter for cholesterol efflux to HDL [36] , is very low in the cornea [37] . Thus , the cornea cannot efficiently remove excess cholesterol by converting to cholesteryl ester or pumping out of the cell , leading to the accumulation of free cholesterol once HMGCR is stabilized by UBIAD1 mutations . Interestingly , familial lecithin-cholesterol acyltransferase ( LCAT ) deficiency and fish eye disease caused by partial loss-of-function of LCAT also exhibit evident corneal opacification due to free cholesterol accumulation [38] . These diseases highlight the importance of cholesterol homeostasis in maintaining corneal function . Currently , corneal transplant surgery is the only way to restore vision in SCD patients [1] . It is urgent to develop other treatments in the future . As corneal overproduction of cholesterol is caused by HMGCR accumulation , it would be interesting to test whether local application of statins , the well-known inhibitors of HMGCR and are widely used for lowering cholesterol [39] , in the form of eyedrops is effective for reducing corneal opacification . In addition , 2-hydroxypropyl-β-cyclodextrin ( HPβCD ) has an excellent ability to solubilize cholesterol and has been used to treat Niemann-Pick type C ( NPC ) disease caused by lysosomal cholesterol accumulation [40] . The local application of HPβCD would be another promising strategy to reduce corneal cholesterol . Our recently identified HMG499 ( also named Cmpd 81 ) might also be used to treat SCD since HMG499 is a potent HMGCR degrader [25] and SCD is caused by HMGCR stabilization . Collectively , we have found that the SCD-associated mutants of UBIAD1 bind and stabilize HMGCR , thereby increasing cellular cholesterol level . We have also generated and characterized a mouse model ( Ubiad1G184R/+ ) for SCD disease that will be valuable for studying the underlying mechanism of SCD and developing therapeutic strategies as well .
All procedures and care of animals were carried out in accordance with the guidelines and protocols approved by the Institutional Animal Care and Use Committee at the Wuhan University under protocol number WDSKY0201408 . We obtained lovastatin ( PHR1285 ) , mevalonate ( 41288 ) , 25-hydroxycholesterol ( H1015 ) , MG-132 ( M8699 ) , Filipin ( F9765 ) , geranylgeraniol ( G3278 ) , protease inhibitor cocktail ( P8340 ) , N-Ethylmaleimide ( E3876 ) , and paraformaldehyde ( PFA ) ( P6148 ) from Sigma; [14C]-acetic acid sodium salt ( NEC084H001 ) from Perkin Elmer; and FuGENE HD transfection reagent ( E2312 ) from Promega; G418 ( 345810 ) , digitonin ( 300410 ) , henylmethylsulfonyl fluoride ( PMSF ) ( 52332 ) , leupeptin ( 108975 ) , pepstatin A ( 516481 ) and ALLN ( 208719 ) from Merck; Hoechst 33342 ( H1399 ) from Invitrogen . Lipoprotein-deficient serum ( LPDS ) [41] and delipidated-fetal calf serum ( FCS ) [42] was prepared from FCS ( S1580 , Biowest ) by ultracentrifugation in our laboratory [25] . Primary antibodies used for immunoblotting were as follows: mouse monoclonal anti-T7 ( 69522 , Merck , 1 μg/ml ) , mouse monoclonal anti-Flag ( F3165 , Sigma-Aldrich , 1:1000 ) , mouse monoclonal anti-Myc IgG-9E10 ( CRL-1729 , ATCC , 1 μg/ml ) , mouse monoclonal anti-ubiquitin IgG-P4D1 ( sc-8017 , Santa Cruz Biotechnology , 1:1000 ) , mouse monoclonal anti-hamster HMGCR IgG-A9 ( CRL-1811 , ATCC , 2 μg/ml ) , mouse monoclonal anti-clathrin heavy chain ( 610500 , BD Biosciences , 1:1000 ) , mouse monoclonal anti-Actin ( A3853 , Sigma , 1:5000 ) , polyclonal anti-HMGCR antibody ( H2 ) was raised against a C-terminal sequence ( aa410-aa888 ) of human HMGCR in our laboratory [43] . Horseradish peroxidase-conjugated donkey anti-mouse ( 715-005-151 , 1:5000 ) and anti-rabbit ( 711-005-152 , 1:5000 ) secondary antibodies were obtained from Jackson ImmunoResearch Laboratories . Primary antibodies used for immunofluorescence staining were as follows: rabbit polyclonal anti-GM130 ( G7295 , Sigma , 1:300 ) , rabbit polyclonal to calnexin ( ab22595 , Abcam , 1:300 ) . Alexa Fluor 488-labeled donkey anti-mouse IgG ( A-21202 , 1:500 ) and Alexa Fluor 555-labeled donkey anti-rabbit IgG ( A-31572 , 1:500 ) secondary antibodies were obtained from Invitrogen . The following plasmids pCMV-Insig-1-Myc , pCMV-HMGCR-T7 , pCMV-HMGCR-T7-K89R/K248R , pEF1a-HA-Ubiquitin were constructed in our laboratory [25] . The coding regions of UBIAD1 was amplified from human cell cDNA or mouse liver cDNA using standard PCR and cloned into cloned into pcDNA3-C-5xMyc vector . The plasmids encoding the variants of UBIAD1 were generated by site-directed mutagenesis based on full-length human or mouse UBIAD1 . All the constructs were verified by DNA sequencing . The primer sequences are listed in S2 Table . HEK-293 cells were grown in monolayer at 37°C in 5% CO2 in medium containing 10% FCS , Dulbecco’s modified Eagle’s medium ( DMEM ) , 100 units/ml penicillin and 100 μg/ml streptomycin sulfate . CHO-K1 cells were grown in monolayer at 37°C in 5% CO2 in medium containing 5% FCS , Ham’s F-12 and DMEM ( 1:1 ) , 100 units/ml penicillin and 100 μg/ml streptomycin sulfate . On day 0 , CHO-K1 cells were set up for experiments at 4×105 cells per 60-mm dish . On day 2 , the cells were transfected with the indicated plasmids by using FuGENE HD reagent . The total amount of DNA in each transfection was adjusted to 2 μg per dish by addition of pcDNA3 mock vector . For whole cell lysate , the cells were harvested and suspended in 120 μl of RIPA buffer ( 50 mM Tris-HCl , pH 8 . 0 , 150 mM NaCl , 0 . 1% SDS , 1 . 5% NP40 , 0 . 5% deoxycholate , 2 mM MgCl2 ) containing protease inhibitors ( protease inhibitor cocktail 1:500 , 10 μg/ml leupeptin , 5 μg/ml pepstatin A , 25 μg/ml ALLN , 1 mM PMSF and 10 μM MG-132 ) . Protein concentrations of the extracts were determined with Pierce BCA protein assay kit ( ThermoFisher , 23225 ) , then mixed with 4×SDS loading buffer ( 150 mM Tris-HCl , pH 6 . 8 , 12% SDS , 30% glycerol , 0 . 02% bromophenol blue , 6% β-mercaptoethanol ) . Proteins were separated with SDS-PAGE , transferred to nitrocellulose filters ( GE Healthcare ) . Immunoblots were blocked by 5% non-fat milk/TBST and probed with indicated primary and HRP-conjugated secondary antibodies . Images were captured with Amersham Imager 680 ( GE Healthcare ) , and intensities of each band were quantified by Image-Pro Plus 6 software ( Media Cybernetics ) . About 50 mg tissues excised from mouse were suspended in 500 μl of buffer A ( 10 mM HEPES/KOH , pH 7 . 6 , 1 . 5 mM MgCl2 , 10 mM KCl , 5 mM EDTA , 5 mM EGTA , 250 mM sucrose ) containing protein protease inhibitors , and homogenized by Precellys 24 ( Bertin ) . The homogenized suspensions were then passed through a #7 needle for 30 times and centrifuged at 1000 g at 4°C for 7 min . The supernatant from the 1 , 000 g spin was centrifuged at 1 x 105 g for 30 min at 4°C . The pellets were the membrane fractions and were dissolved in 0 . 1 ml of SDS lysis buffer ( 10 mM Tris-HCl , pH6 . 8 , 100 mM NaCl , 1% SDS , 1 mM EDTA , 1 mM EGTA ) . HEK-293 cells were transfected with pCMV-Hs-UBIAD1-WT-TAP or pCMV-Hs-UBIAD1-G186R-TAP . One day later , cells were changed to medium containing 800 μg/ml G418 . Fresh medium was exchanged every 2–3 days until colonies formed after about 2 weeks . Individual colonies were isolated with cloning cylinders , and the expression levels of UBIAD1 protein were verified by immunoblot . HEK-293/Hs-UBIAD1-WT-TAP and HEK-293/Hs-UBIAD1-G186R-TAP stable cells were set up in 100-mm dish . Cells were harvested and lysed in immunoprecipitation ( IP ) buffer ( 1% digitonin , PBS , 5 mM EDTA , 5 mM EGTA ) with protein protease inhibitors . Then cells were needled 15 times and centrifuged at 13 , 200 rpm for 20 min . The supernatant was pre-cleared with protein A/G agarose ( sc-2003 , Santa Cruz Biotechnology ) , then immunoprecipitated with rabbit IgG coupled agarose ( A2909 , Sigma ) at 4°C for 2 hr . Agarose-captured proteins were released with TEV Protease ( P8112S , New England Biolabs ) in 500 μl IP buffer at RT for 1 hr . The released proteins were immunoprecipitated with anti-Flag M2 agarose beads ( A2220 , Sigma ) at 4°C for 2 hr , and eluted with 3x Flag peptide at 4°C for 30 min . The eluted fractions were resolved with SDS-PAGE and stained with Coomassie blue . The identifies of proteins from eluted fractions were determined by tandem mass spectrometry . The cells were harvested and suspended in 600 μl of IP buffer ( 1% digitonin , PBS , 5 mM EDTA , 5 mM EGTA , and 10 mM N-Ethylmaleimide ) containing protein protease inhibitors , and passed through #7 needle 15 times . The cell lysates were centrifuged at 13 , 200 rpm at 4°C for 10 min , immunoprecipitated with anti-T7 antibody coupled agaroses ( 69026 , Merck ) , and eluted with 2x SDS loading buffer . Whole cell lysates and pellets were subjected to SDS-PAGE and immunoblotted with indicated antibodies . Cells were fixed with 4% PFA in PBS , permeabilized with 0 . 2% Triton X-100 ( T8787 , Sigma ) for 8 minutes , blocked with 2% BSA for 1 hr and incubated with indicated primary antibodies for 1 hr at room temperature . Fluorescence-labeled secondary antibodies were used at concentration of 4 μg/ml in PBS containing 0 . 2% BSA for 45 minutes . Nuclei was stained with Hoechst 33342 . Immunofluorescence images were captured by Leica TCS SP2 confocal microscope . Images were acquired at identical laser output , gain , and offset [44] . The protein sequences of UBIAD1 from difference species were aligned using ClustalW algorithm in MEGA X software [45] . The accession numbers of these protein sequences were used as followings: human ( NP_037451 . 1 ) , chimpanzee ( JAA33188 . 1 ) , rhesus ( NP_001247708 . 1 ) , tree shrew ( XP_006145395 . 1 ) , sheep ( XP_004013772 . 1 ) , horse ( XP_008520311 . 1 ) , elephant ( XP_003413515 . 1 ) , dog ( XP_544571 . 1 ) , rabbit ( ETE69346 . 1 ) , mouse ( NP_082149 . 1 ) , chicken ( NP_001026050 . 1 ) , snake ( ETE69346 . 1 ) , frog ( NP_001016538 . 1 ) , zebrafish ( NP_001186655 . 1 ) , sea urchin ( XP_011664743 . 1 ) and fly ( NP_523581 . 1 ) . Corneas were fixed by 4% PFA and sectioned with frozen cryostat ( Leica CM3050 S ) at 7 μm . Frozen sectioned slides were washed twice with PBS , stained with 50 μg/ml filipin in 10% FBS/PBS for 1 hr at room temperature , washed three times with PBS and mounted . Images were captured by fluorescence microscopy ( Zeiss Axio Imager Z2 ) using a UV filter set , and the intensity of mercury lamp was turned to 10% of the maximal strength . Images were acquired at identical output , gain , and offset [46] . Cells were incubated in medium containing 10% delipidated-FCS , 1 μM lovastatin and 50 μM mevalonate for 16 hr . After depletion , cells were washed to remove lovastatin , and change to medium with 10% delipidated -FCS for 3 hr . Then [14C]-acetate ( 36 μCi/100-mm dish ) were added and cells were treated for additional 2 hr . Cells were washed twice with PBS , dissolved by 0 . 5 ml 0 . 1 N NaOH , and saponified with ethanol and 75% potassium hydroxide for 2 hr . Then the nonpolar lipids ( cholesterol ) were extracted in petroleum ether and evaporated to dryness with N2 . Following addition of concentrated HCl , polar lipids ( fatty acids ) were extracted in petroleum ether and evaporated to dryness with N2 . The lipids were resolved by thin-layer chromatography ( 1 . 05582 . 0001 , Merck ) . Radioactive signals were visualized with phosphoimager , and images were captured with Typhoon FLA 9000 ( GE Healthcare ) . The knockout first conditional ready targeting vector , containing a gene-trap cassette , was electroporated into ES cell line from C57BL/6J mice . Positive ES clones were verified by PCR and Southern blot , and three positive ES clones were injected to BALB/c blastocysts to generate chimeric mice . Male mice with high percentage chimeras were bred with female C57BL/6J mice ( Shanghai Laboratory Animal Company , China ) to get the heterozygous Ubiad1 knockout mice . To get a conditional knockin allele , these heterozygous knockout mice were bred with Flp recombinase deleter transgene ( Jackson Laboratory , USA ) to remove the Frt-flanked cassette . Subsequently , these mice were crossed with EIIA-Cre recombinase transgenic mice ( Jackson Laboratory , USA ) to remove Loxp-flanked cassette , thus getting the whole tissue Ubiad1 G184R knockin mice . The genotypes were identified by PCR using primers P1 ( GCAAGCTGTATTTTGCCTTG ) , and P2 ( CGAAAGTGATGAGGATGACGAGGT ) . The male and female Ubiad1G184R/+ were intercrossed to get Ubiad1+/+ and Ubiad1G184R/+ mice littermates for experiments . All mice were maintained on a 12-h light/dark schedule , and fed ad libitum access to water and standard chow diet ( Shanghai Laboratory Animal Company , China ) . Blood were collected from anesthetized mice , and serum was prepared from blood by centrifuging at 1500 × g for 10 min . Tissues were excised and weighted , then homogenized in chloroform/methanol ( 2:1 ) with Precellys 24 ( Bertin ) . The lipids were extracted and dried under N2 , and dissolved in ethanol . Total cholesterol and free cholesterol levels in serum and tissues were determined with Amplex Red Cholesterol Assay Kit ( A12216 , Invitrogen ) . HDL cholesterol , LDL cholesterol and triglyceride levels in the serum and liver were measured with HDL cholesterol , LDL cholesterol and triglyceride Assay Kit ( Shanghai Kehua Bio-engineering , China ) , respectively . Female Ubiad1G184R/+ mice were mated with male Ubiad1G184R/+ mice . Pregnant mice on the 13-day of conception were sacrificed by brief exposure to CO2 . The uterine horns were excised into Petri dish and all embryos were carefully dissected . Each embryo was transferred into a new dish; the heads were used for genotyping . The rest of embryo was minced thoroughly using sharp Iris scissors , and digested with trypsin/EDTA for 20 minutes . Trypsin was neutralized , and the MEFs were collected by centrifuging . MEFs were cultured in medium with 10% FCS/DMEM , 100 units/ml penicillin and 100 μg/ml streptomycin sulfate . Mice were first anaesthetized with 1% pentobarbital sodium in PBS ( 10 μl/g ) , then corneal opacifications were observed and images were captured with stereoscopic microscope ( Olympus SZX16 ) . The corneas were excised from eyes and carefully removed other tissues under stereomicroscope , then used for protein and lipid analyses . The statistical analyses were carried out using GraphPad Prism 6 software . Data were presented as means ± SD and analyzed by unpaired two-tailed Student’s t-test . Statistical significance was set at p < 0 . 05 .
|
Schnyder corneal dystrophy ( SCD ) is a rare genetic eye disease caused by deposition of free cholesterol in the cornea . It is closely correlated with mutations in the UbiA prenyltransferase domain-containing protein-1 ( UBIAD1 ) gene , which encodes an enzyme catalyzing biosynthesis of coenzyme Q10 and vitamin K2 . The underlying mechanism by which UBIAD1 mutations result in SCD development is unclear . Here , we report that SCD-associated mutations trap UBIAD1 in the ER and block Insig-1 mediated HMGCR degradation . We also generated a heterozygous mouse model ( Ubiad1G184R/+ ) that mimics human SCD . We conclude that SCD-associated UBIAD1 mutations decrease HMGCR degradation and subsequently increase cholesterol biosynthesis in the cornea .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"ocular",
"anatomy",
"animal",
"models",
"model",
"organisms",
"immunoprecipitation",
"experimental",
"organism",
"systems",
"research",
"and",
"analysis",
"methods",
"lipids",
"animal",
"studies",
"proteins",
"ubiquitination",
"mouse",
"models",
"sterols",
"precipitation",
"techniques",
"cholesterol",
"biochemistry",
"anatomy",
"post-translational",
"modification",
"cornea",
"biology",
"and",
"life",
"sciences",
"ocular",
"system",
"biosynthesis"
] |
2019
|
Schnyder corneal dystrophy-associated UBIAD1 mutations cause corneal cholesterol accumulation by stabilizing HMG-CoA reductase
|
Chimeric Antigen Receptor ( CAR ) T-cells have emerged as a powerful immunotherapy for various forms of cancer and show promise in treating HIV-1 infection . However , significant limitations are persistence and whether peripheral T cell-based products can respond to malignant or infected cells that may reappear months or years after treatment remains unclear . Hematopoietic Stem/Progenitor Cells ( HSPCs ) are capable of long-term engraftment and have the potential to overcome these limitations . Here , we report the use of a protective CD4 chimeric antigen receptor ( C46CD4CAR ) to redirect HSPC-derived T-cells against simian/human immunodeficiency virus ( SHIV ) infection in pigtail macaques . CAR-containing cells persisted for more than 2 years without any measurable toxicity and were capable of multilineage engraftment . Combination antiretroviral therapy ( cART ) treatment followed by cART withdrawal resulted in lower viral rebound in CAR animals relative to controls , and demonstrated an immune memory-like response . We found CAR-expressing cells in multiple lymphoid tissues , decreased tissue-associated SHIV RNA levels , and substantially higher CD4/CD8 ratios in the gut as compared to controls . These results show that HSPC-derived CAR T-cells are capable of long-term engraftment and immune surveillance . This study demonstrates for the first time the safety and feasibility of HSPC-based CAR therapy in a large animal preclinical model .
HIV-1 specific cytotoxic T lymphocytes mount a key immune response to HIV and are crucial for the control of viremia and the elimination of HIV infected cells . Previous studies have shown that a chimeric antigen receptor containing the CD4 molecule linked to the signaling domain of the T cell receptor ζ chain ( CD4CAR ) can be used to redirect peripheral T cells to target HIV infected cells[1] . CD4 CAR modified T cells can recognize and respond to HIV gp120 envelope protein on infected cells and can effectively kill HIV infected cells and limit HIV replication in vitro . Early clinical trials with CD4CAR modified T cells were shown to be safe but had limited antiviral efficacy[2 , 3] . The lack of in vivo functionality of the transferred T cells may have been due to suboptimal T-cell handling and expansion , or because CD4CAR-expressing CD8+ T-cells were susceptible to HIV infection and elimination[1 , 2] . Protection of the CD4CAR modified cells from viral entry is essential in order to ensure T-cell functionality and survival[4 , 5] . Hematopoietic Stem/progenitor Cell ( HSPC ) -based gene therapy has several advantages over T cell adoptive therapy . First , the regenerative nature of HSPC provides a lifelong supply of engineered T cells against antigen-expressing target cells , which is key to achieve long term immune surveillance and a functional cure of HIV infection . Secondly , modified cells undergo normal T cell differentiation and selection , eliminating potentially self-reactive T cells and increasing the potential for the development of immunological memory[6–8] . Our previous studies using humanized mice demonstrated that HSPCs modified with a protective CD4CAR resulted in successful differentiation of CD4CAR expressing T cells and significant suppression of HIV replication , suggesting a high degree of feasibility in redirecting immunity with an HSPC-based approach[4] . The nonhuman primate model is an ideal preclinical surrogate for the development of cure strategies in HIV+ patients . A series of well-characterized HIV-like viruses are available to recapitulate acute , chronic , and cART-suppressed infection [9 , 10] . Furthermore , use of a large animal model facilitates the detailed measurement of viral reservoirs in peripheral blood and in tissues . We and others have used macaque models extensively to evaluate autologous HSPC transplantation to combat a number of human diseases , including HIV infection . We have previously demonstrated that infection with the highly CCR5-tropic , HIV-enveloped simian/human immunodeficiency virus SHIV-1157ipd3N4 ( “SHIV-C” ) resembles suppressed infection in patients , including suppression by cART , rebound following cART withdrawal , and seeding of viral reservoirs in tissues [11 , 12] . Further , we have shown that autologous HSPCs can be modified to resist infection , for example via CCR5 gene editing [13] , or expression of a potent inhibitor of HIV/SHIV fusion , the enfuvirtide-related peptide C46[14] . We have optimized these experiments in pigtail macaques ( M . nemestrina ) , which carry a TRIM5 genotype that is permissive to lentivirus-mediated gene therapy approaches [15] . In short , our nonhuman primate model recapitulates virological and immune facets of HIV infection in patients , and facilitates evaluation of gene therapy-based HIV cure strategies . Here , we asked whether HIV/SHIV-specific immunity could be engendered in HSPCs and their progeny , via modification of autologous HSPCs with a C46CD4CAR-expressing lentivirus vector . Modified cells were evaluated in vitro , and in SHIV-infected nonhuman primates .
Our lentivirus constructs contain a CD4 based CAR ( CD4CAR ) that is composed of the human CD4 extracellular and transmembrane domain linked to the human CD3ζ signaling domain ( 4 ) . CD4 is the primary receptor for HIV , hence to protect CD4CAR expressing T cells from viral infection , we co-expressed the C46 fusion inhibitor in the CAR-containing vector ( C46CD4CAR ) ( Fig 1A ) [16] . We also generated a control vector that contains C46 and a truncated form of CD4CAR that lacks the signaling domain of CD3 ζ ( C46CD4CARΔzeta ) ( Fig 1A ) . Expression of CD4CAR ( without C46 ) resulted in increased HIV infection of Jurkat T cells ( 35 . 8% HIV+ , as compared to 12% for unmodified cells ) ( Fig 1B ) . However , expression of C46CD4CAR blocked HIV infection ( 0 . 21% HIV+ cells ) , indicating that C46 protected gene modified cells from HIV infection . We next tested if the CD4CAR molecule can functionally respond to antigen in pigtail macaque T cells . We transduced pigtail macaque T cells with control C46CD4CARΔzeta or C46CD4CAR vector , then stimulated the cells with either uninfected or HIV-infected cells that expressed HIV envelope . C46CD4CAR transduced pigtail macaque T cells produced IL-2 and IFNγ in response to stimulation , indicating that the CD4CAR molecule is functional in the NHP cells ( Fig 1C ) . In contrast , control cells that expressed C46CD4CARΔzeta did not respond to HIV infected cells . These data show that C46CD4CAR cells are protected against HIV infection and respond functionally and specifically to HIV antigen in NHP cells in vitro . To examine the effects of the CD4CAR in vivo , four male juvenile pigtail macaques were transplanted with autologous HSPCs that were transduced with lentivirus expressing C46CD4CAR ( “CAR1” and “CAR2” ) or C46CD4CARΔzeta ( “Control 1” and “Control 2” ) . As shown in S1 Fig , percent lentivirus marking from each animal’s HSPC infusion product ranged from 4 . 65% to 40% in colony forming assays . After HSPC transplant , recovery kinetics of total white blood cells , platelets , neutrophils , and lymphocytes in both control and CAR animals were normal [14 , 17] ( S2 Fig ) . We detected stable gene marking of PBMCs from all animals prior to SHIV challenge ( Fig 2A ) . In addition , we were able to detect C46CD4CAR or C46CD4CARΔzeta modified cells in peripheral blood by using an anti-human CD4 antibody clone ( 13B8 . 2 ) that detects human , but not pigtail macaque CD4 ( Fig 2B ) . Because this antibody will only label the human CD4CAR in our animals , we refer to CAR+ cells as “huCD4+” . We found that 0 . 1% to 1 . 25% of CD45+ peripheral blood leukocytes from control or CAR animals were huCD4+ ( Fig 2C ) . Importantly , huCD4+ CAR cells from both control and CD4CAR animals differentiated into multiple hematopoietic lineages , including T cells ( CD45+CD3+ ) , NK cells ( CD3−CD2+NKG2A+ ) [18] , B cells ( CD45+CD3−CD20+ ) and monocytes and macrophages ( CD3−CD20−CD14+ ) ( Fig 2D ) . These results show that autologous transplantation of C46CD4CAR-transduced HSPC is safe and well tolerated , and results in stable , multilineage engraftment with typical kinetics of hematopoietic recovery . To study the effect of C46CD4CAR transplantation on SHIV replication , animals were infected with SHIV-C for ~24 weeks followed by 28 weeks of combination antiretroviral therapy ( cART ) and subsequent cART withdrawal . At least 12 weeks after cART cessation , animals were then sacrificed for necropsy ( Fig 3A ) . Both CD4CAR and control animals had slightly higher plasma viral loads prior to cART , and did not achieve full virus suppression following cART ( Fig 3B ) . This was likely due to residual immune suppression from the transplant procedure [11 , 19] . The CAR1 animal had approximately 1 log higher viral load as compared to control animals during acute and chronic SHIV infection , while the CAR2 animal , in which more than 1% of PBMCs were modified with C46CD4CAR prior to SHIV infection , had lower peak viremia during acute infection and showed progressively decreasing viral loads prior to cART ( Fig 3B ) . While CAR 1 and control animals appear to have reached set points after 4 weeks of SHIV infection , CAR 2 animal demonstrated a trend of continuous reduction of viral load throughout primary infection ( Fig 3B ) . Interestingly , when we compared average viral load after cART withdrawal to average viral load during primary infection ( week 2 to week 22 ) , we found that both CAR containing animals had lower average rebound viremia ( 1 . 4–2 . 11 log lower than primary setpoint ) as compared to the control animals ( 0 . 4–0 . 8 log lower than primary setpoint ) ( Fig 3C ) . These findings are consistent with a model in which C46CD4CAR cells are capable of establishing virus-specific immune memory and responding to recrudescent SHIV viremia . We next measured antigen-dependent responses in C46CD4CAR and control animals by monitoring CAR gene marking as a function of SHIV plasma viremia . Lentivirus-marked cells were readily detectable by Taqman in CAR and control animals over the course of our nearly two-year study ( Fig 4A ) . Interestingly , CAR animals , but not controls , showed increased gene marking in the periphery at multiple time points . These were coincident with increases in SHIV viremia , notably during primary infection and viral rebound after cART withdrawal ( Fig 3B ) . To investigate further , we used flow cytometry to stain for huCD4+ PBMCs at multiple time points following SHIV infection . Consistent with Taqman-based gene marking data , we found that huCD4+ cells from CAR animals , but not control animals , expanded upon SHIV infection and post-cART withdrawal viral rebound ( Fig 4B–4E ) . This confirms that functional C46CD4CAR cells require intact CD3ζ signaling in order to expand in response to SHIV antigen . Furthermore , we observed an increase of C46CD4CAR+ cells during acute and chronic SHIV infection ( Fig 4D and 4E ) , reminiscent of a primary immune response to infection . After the cessation of cART treatment , the percentage of CD4CAR+ cells again increased rapidly , mimicking a memory response . The CAR2 animal , which had higher gene marking prior to SHIV infection ( Fig 2A ) , contained as many as 10% and 12 . 6% huCD4+ PBMCs during primary untreated infection and after cART withdrawal , respectively . We also observed expansion in percentage of huCD4+ cells among T cells ( S3A Fig ) and in huCD4+ T cell numbers ( S3B Fig ) . The expansion of CD4CAR+ cells is primarily driven by CAR expressing T cells as shown in Sup Fig 3A and 3B . During viral rebound , the 4–10 fold higher levels of CAR marking in C46CD4CAR animals relative to controls was consistent with the 1 . 5–2 log decrease in rebound viremia relative to primary infection in these animals ( Fig 3C ) . These data suggest that CAR-marked cells engraft long term and are capable of antigen-specific expansion months or years after transplantation . Our in vitro data suggest that expansion of C46CD4CAR cells is specific to HIV/SHIV antigen ( Fig 1 ) . To examine how CAR modified cells respond to SHIV replication in vivo , we monitored the naïve , effector , and memory phenotypes of CAR T cells longitudinally in our transplanted animals following SHIV infection . Prior to SHIV infection , huCD4+ ( CAR+ ) and unmarked T cells ( CAR- ) shared similar percentages of naïve ( CD28+CD95− ) , effector ( CD28−CD95+ ) and memory ( CD28+CD95+ ) subsets ( Fig 5A ) . Strikingly , huCD4+ cells became predominantly effector T cells after SHIV infection , consistent with a response to SHIV antigen . During cART-dependent viral suppression , when the percentage of huCD4+ T cells contracted , we found that most displayed a naïve or memory phenotype . After cART withdrawal , huCD4+ T cells again displayed a predominant effector phenotype . Antigen-dependent increases in the percentage of effector cells were observed in C46CD4CAR animals , but not in CD4CARΔzeta controls ( Fig 5B–5E ) . To investigate if CAR+ effector cells can mediate specific killing of HIV Env expressing cells , we performed ex vivo killing assays using PBMCs from CAR or control animals during primary SHIV infection prior to cART treatment and after cART withdrawal ( S4A Fig ) . We used U1 cells stimulated to express HIV envelope as targets . Notably , this human cell line lacks rhesus MHC molecules , and therefore should only be killed following CAR-dependent recognition of the HIV envelope . While PBMCs from control animals failed to mediate specific killing of Env+ U1 cells , PBMCs from CAR animals effectively mediated specific killing of Env+ U1 cells over control U1 cells ( S4B and S4C Fig ) . Interestingly , we observed a trend of an increase in IFNγ production of cryopreserved T cells from CAR animals in response to SHIV peptide pool stimulation , suggesting a potential synergy between CAR T cells and anti-SHIV natural T cell response ( S5A–S5C Fig ) . However , we did not observe a clear difference in production of natural anti-SHIV or anti-HIV Env antibody ( S6A and S6B Fig ) . These findings are consistent with our previous data ( Figs 1 and 3 ) , demonstrating that C46CD4CAR HSPC-derived cells generate a long-lived , functional response to SHIV antigen . Previous CD19 CAR T cell therapy has been associated with cytokine release syndrome , and the principle cytokines elevated in patients treated with CD19 CAR T cells were IFNγ and IL-6[20] . To monitor toxicity associated with the CD4CAR HSPCs treatment , we measured plasma cytokine levels from transplanted CAR and control animals blood collected prior to SHIV challenge , during untreated SHIV infection , during cART treatment and after cART withdrawal ( S7A Fig ) . As shown in S7B Fig , most proinflammatory cytokines , including IFNγ , IL-1β , IL-2 , IL-6 , MIP1β , MIP1α , MCP1 have mostly undetectable levels or no clear difference between control and CAR animals ( S7C Fig ) . We observed lower sCD40L levels for CAR control animals during untreated SHIV infection and after cART , which may be a result of reduced SHIV-mediated inflammation in CAR animals . Overall , we observed no toxicity associated with the CAR transplant as compared to control animals . We extended our analysis of C46CD4CAR cells by examining trafficking to multiple tissue sites , including those that have been characterized as viral reservoirs[21–23] . Both C46CD4CAR and C46CD4CARΔzeta cells were found in multiple lymphoid tissues , including various lymph nodes , gut , and bone marrow ( S8 Fig ) . Similar to PBMCs , CAR animals had higher percentage of huCD4+ cells among T cells in various tissues as compared to control animals ( S8A Fig ) . As with huCD4+ PBMC , tissue-associated CAR cells were multilineage , including CD4+ and CD8+ T cells , NK cells , and macrophages/monocytes . There were no obvious differences in cell composition between C46CD4CAR and C46CD4CARΔzeta modified cells ( S8B–S8E Fig ) . To examine the ability of C46CD4CAR cells to protect against SHIV-dependent depletion of CD4+ cells in the gut , biopsies were taken from the GI tract ( colon or duodenum/jejunum ) before SHIV infection and after cART withdrawal , and analyzed by flow cytometry . Control animals displayed a profound loss of CD4+ cells , both in terms of CD4+CD3+ T cell percentage ( Fig 6A and S9A Fig ) and CD4/8 ratio ( Fig 6B and S9B Fig ) . Strikingly , CD4+ T-cell percentage and CD4/8 ratio were substantially higher in C46CD4CAR animals following cART withdrawal as compared to the control animals , suggesting that functional CAR cells contributed to protection of immune homeostasis in this compartment . Furthermore CD4+ effector memory T-cells ( CD3+CD4+CCR7−CD45RA− ) , which are major target cells of HIV infection , were also protected in the gut of C46CD4CAR animals ( Fig 6C and S9C Fig ) . At necropsy , gene marking in lymphoid tissues ( including spleen , mesenteric lymph nodes , axillary lymph nodes , inguinal lymph nodes , and submandibular lymph nodes ) and gut ( including duodenum , jejunum , ileum , cecum , colon , and rectum ) was significantly higher in CAR animals relative to controls ( Fig 6D ) . Interestingly , SHIV RNA measurements in these tissues showed that CAR containing animals had substantially lower viral loads ( Fig 6E ) . Although we did not observe significant differences in gene marking in the brain ( including hippocampus , basal ganglia , thalamus , parietal cortex , and cerebellum ) between CAR and control animals , SHIV RNA measurements were also lower in this compartment in CAR animals , relative to controls ( Fig 6D and 6E ) . In particular , the CAR2 animal had dramatically lower SHIV mRNA ( 4–5 logs ) across all lymphoid tissues as compared to control animals ( S10 Fig ) . Collectively , these results demonstrate that C46CD4CAR cells in tissues are capable of long term , multilineage engraftment , and are protected against SHIV replication , consistent with our observations in peripheral blood .
The seminal case study for HIV cure/remission is the Berlin patient , who received an allogeneic , HLA-matched HSPC transplant from a donor homozygous for CCR5Δ32[24] , and has stimulated the search for HSPC-based cure approaches . Allogeneic HSPC transplantation without CCR5Δ32-protected donor cells in 2 HIV+ recipients initially resulted in undetectable HIV-1 after patients achieved full donor chimerism; this was likely due to a “graft versus reservoir” effect in which donor lymphocytes destroyed latently infected host cells . Ultimately , this intervention failed to eradicate latently infected cells , which rebounded after cART cessation[25] . These studies indicated that a combinatorial approach , rendering the blood and immune system resistant to infection and at the same time harnessing the immune system to attack infected cells , would be required . Numerous studies have used various gene therapy and gene editing approaches to genetically modify autologous stem cells , rendering them resistant to HIV infection[13 , 14 , 26–30] . Our nonhuman primate model of suppressed HIV infection is highly relevant to HIV cure studies in humans , utilizing a virus that is suppressed by a clinically relevant cART regimen , establishment of viral persistence in secondary lymphoid tissues , and rebound of viral replication following cART withdrawal[11 , 12] . We have previously shown that C46-expressing T-cells are protected against CCR5- and CXCR4-tropic viruses , and support a more robust immune response against infected cells in vivo [14 , 16] . Here , we demonstrate that HIV/SHIV-specific CAR cells possess strong antiviral activity even at low levels . These cells should act as sentinels , generating a robust immune response to reactivated infected cells months or years after they are introduced , without requiring expression in a high percentage of immune cells . The most novel aspect of our approach is the generation of CAR cells from autologous HSPCs . Stem cell-based expression of CARs contributes a long-lived source of these cells , capable of providing lifelong immune surveillance against recrudescent virus . In contrast , adoptive transfer of CAR-modified T-cell products must overcome barriers including immune exhaustion , limited trafficking to tissues , and lack of functionality at these sites [31 , 32] . Furthermore , although CD4CAR T-cells persist long term in patients [33] , it is unclear whether persisting cells are capable of responding to increased antigen loads , for example during cART treatment interruption . Here , we show for the first time in a clinically relevant large animal model , that autologous transplantation with a CAR-modified HSPC is safe and can be used to redirect long-term anti-viral immunity . We observed multilineage engraftment of autologous , gene-modified cells that persisted for almost 2 years . Moreover , we found that CAR-expressing cells expanded in response to SHIV infection in an antigen-driven fashion , and differentiated into effector cells in a CD3ζ domain-dependent manner . Intriguingly , we found that engineered CAR cells contracted during cART treatment during lower levels of antigen expression , followed by a rapid expansion after cART withdrawal , mimicking a memory response . As a result , CAR animals had decreased viremia during post-cART viral rebound , as compared to control animals . More importantly , we were able to detect CAR cells in multiple lymphoid tissues , including gut-associated lymphoid tissues . CD4+ T-cells at these sites facilitate viral replication , are significantly depleted during untreated infection , and are slow to regenerate during cART treatment[34] . We found significantly lowered SHIV mRNA in lymph nodes , gut and brain from C46CD4CAR animals as compared to control animals . Strikingly , both CAR animals showed substantially improved CD4/CD8 ratios and higher percentages of CD4+ and CD4+ effector memory cells in the gut after cART withdrawal , suggesting CD4+ T-cell protection was C46CD4CAR-dependent . A small percentage of CAR-modified cells appeared to be sufficient to redirect an effective immune response against SHIV-infected cells in our study . For both CAR animals , we observed robust expansion of C46CD4CAR cells after SHIV infection and cART withdrawal , which was dependent on CD3ζ signaling . The immediate , memory-like response for CAR cells after cART withdrawal from both CAR animals likely contributed to improved control of SHIV viremia and CD4 protection in the gut as compared to control animals . Many of the same optimization parameters used in T-cell-based CAR products are also applicable to HSPC-based CAR cells . For example , 2nd and 3rd generation chimeric antigen receptors with co-stimulatory molecules such as 41BB and/or CD28 [35] may further boost the primary and secondary responses of HSPC-derived CAR T cells , although the impact of these modifications on thymopoiesis has not yet been tested . Another novel aspect of our HSPC-based approach is the generation of CAR cells in lineages other than T-cells ( Figs 2 and S3 ) . The contribution of CAR-expressing cells other than T-cells in our model remains to be determined . While CAR expressing natural killer cells can contribute to clearance of infected cells[36 , 37] , C46CD4CAR is likely not functional in other cell types due to the lack of signaling pathway for CD3ζ[38] . If necessary , cell type-specific expression of the chimeric antigen receptor from transduced HSPCs may further improve efficacy and safety . Our studies clearly demonstrate the potential of using CAR gene therapy in HSPCs to redirect anti-HIV immunity against HIV-1 infection . These results set the stage for future attempts to eradicate viral infection and provide more effective immune surveillance for HIV , using optimized CAR vectors and combinatorial approaches , for example with latency reversing agents and/or additive immunotherapies . Importantly , these findings have broad implications beyond HIV: additional preclinical studies should be performed to explore HSPC-expressed CARs against other infectious diseases and cancer in greater detail .
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health ( ”The Guide” ) , and was approved by the Institutional Animal Care and Use Committees of the Fred Hutchinson Cancer Research Center and University of Washington , Protocol # 3235–01 . All animals were housed at and included in standard monitoring procedures prescribed by the Washington National Primate Research Center ( WaNPRC ) . This included at least twice-daily observation by animal technicians for basic husbandry parameters ( e . g . , food intake , activity , stool consistency , overall appearance ) as well as daily observation by a veterinary technician and/or veterinarian . Animals were housed in cages approved by “The Guide” and in accordance with Animal Welfare Act regulations . Animals were fed twice daily , and were fasted for up to 14 hours prior to sedation . Environmental enrichment included grouping in compound , large activity , or run-through connected cages , perches , toys , food treats , and foraging activities . If a clinical abnormality was noted , standard WaNPRC procedures were followed to notify the veterinary staff for evaluation and determination for admission as a clinical case . Animals were sedated by administration of Ketamine HCl and/or Telazol and supportive agents prior to all procedures . Following sedation , animals were monitored according to WaNPRC standard protocols . WaNPRC surgical support staff are trained and experienced in the administration of anesthetics and have monitoring equipment available to assist , including monitors of heart rate , respiration , blood pressure , temperature , and blood oxygenation . Monitors supplied readily and easily interpretable alerts , including audible alarms , LCD readouts , etc . For minor procedures , the presence or absence of deep pain was tested by the toe-pinch reflex . The absence of response ( leg flexion ) to this test indicates adequate anesthesia for this procedure . Similar parameters were used in cases of general anesthesia , including the loss of palpebral reflexes ( eye blink ) . Analgesics were provided as prescribed by the Clinical Veterinary staff for at least 48 hours after the procedures , and could be extended at the discretion of the clinical veterinarian , based on clinical signs . Decisions to euthanize animals were made in close consultation with veterinary staff , and were performed in accordance with guidelines as established by the American Veterinary Medical Association Panel on Euthanasia ( 2013 ) . Prior to euthanasia , animals were first rendered unconscious by administration of ketamine HCl . Four juvenile male pigtail macaques were transplanted with autologous , lentivirus modified HSPCs as previously described [14] . In short , animals were mobilized with granulocyte-colony stimulating factor ( G-CSF ) and stem cell factor ( SCF ) for 4 days prior to collection of large volume bone marrow aspirates and bead-based positive selection of CD34+ cells . Over a 48-hour ex vivo culture period , cells were transduced twice with the lentiviruses indicated in Fig 1 at a multiplicity of infection ( MOI ) of 5 ( CAR 1 ) or 10 ( CAR 2 , Control 1 , Control 2 ) . During HSPC transduction ex vivo , animals received a myeloablative conditioning regimen consisting of a fractionated dose of 1020 cGy total body irradiation . Following conditioning , the HSPC product was infused back into the autologous animal . A small aliquot of the infused cell product was plated in Colony Gel Medium ( Reach Bio , Seattle , WA ) and analyzed as previously described[14 , 39] . Individual colonies and total genomic DNA ( gDNA ) isolated at the indicated post-transplant time points were measured by gel-based and Taqman-based PCR methods , respectively , as previously described [14 , 17] . Animals were allowed to recover for approximately 200 days prior to infection with SHIV-C , which was administered to animals via the intravenous challenge route as previously described [11] . Combination antiretroviral therapy consisted of 20 mg/kg Tenofovir and 40 mg/kg FTC dosed 1X/day subcutaneous , and 150mg Raltegravir , dosed 2X/day oral with food . Plasma viral loads , peripheral T-cell counts , longitudinal tissue surgeries , and necropsy tissue collections were conducted as previously described [12 , 14] . Taqman-based peripheral blood measurements were performed from gDNA isolated from total leukocytes . Total RNA and gDNA from tissue samples were isolated using a Precellys 24 homogenizer and CK28-R hard tissue homogenizing beads ( Bertin Corp . ) as previously described [12] . Normalized SHIV RNA copy number in tissue was calculated by normalizing SHIV RNA copy number to the crossing threshold of macaque RNase P subunit p30 RNA . PCR-based assays for SHIV were designed not to detect HIV-based lentiviral vectors and average viral loads and log reduction in viral load were calculated as described previously [12] . Briefly , average viral loads following cART withdrawal were calculated by averaging each measurement from the first detection of recrudescent plasma viremia to the final measurement taken at necropsy ( 11–13 weekly data points ) . Average viral loads before cART were calculated by averaging plasma viremia over the same number of weekly data points from primary infection , beginning with the first detection of virus following intravenous SHIV challenge . To detect huCD4+ CAR modified cells , PBMCs and tissue necropsy samples were stained with the following antibodies: anti-human specific CD4 antibody ( for detection and analysis of CD4CAR modified cells; Beckman Coulter , clone 13B8 . 2 ) , anti-NHP CD45 ( BD Biosciences , clone D058-1283 ) , anti-CD4 ( eBiosciences , clone OKT4 ) , anti-CD8 ( eBiosciences , clone SK1 ) , anti-CD20 ( eBiosciences , clone 2H7 ) , anti-NK2Ga ( Beckman Coulter , clone A60797 ) , anti-CD14 ( Beckman Coulter , clone IM2707U ) , anti-CD95 ( BD Biosciences , clone DX2 ) , anti-CD28 ( BD Biosciences , clone D28 . 2 ) , and anti-CD3 ( BD Biosciences , clone SP34-2 ) . Fluorophore conjugates included FITC , PE , PE-Cy5 , PE-Cy7 , alexa700 , V500 , efluor450 , APC , and APC-efluor780 . To test C46CD4CAR cell functions in NHP cells , NHPPBMCs were purified from healthy pigtail macaque blood and stimulated with bead bound anti-CD3 ( BD Biosciences , clone SP34-2 ) and anti-CD28 ( BD Biosciences , clone D282 . 2 ) for 3 days . Afterwards , cells were transduced with either C46CD4CAR or C46CD4CARΔzeta lentivirus . 2 days after transduction , cells were co-incubated with either T1 cells or HIV-infected Sup-T1 cells ( AIDSreagent ) for 16 hours , followed by 6 hours of GolgiPlug treatment . Afterwards cells were first surface-stained with anti-CD3 , anti-human CD4 ( for detection of CD4CAR transduced cells ) and then intracellular-stained with anti-IFNγ ( eBiosciences , clone 4S B3 ) , anti-IL-2 ( BD Biosciences , clone MQ1-17H12 ) and analyzed by flow cytometry . To measure natural T cell response to SHIV infection , PBMCs were isolated from transplanted pigtail macaque peripheral blood collected during primary SHIV infection prior to cART treatment , and after cART withdrawal and cryopreserved . Total PBMCs were then thawed and stimulated with either no stimulation or SIVmac Gag , Pol , tat , nef , vif and HIV Env peptide pool overnight , followed with 6 hours of GolgiPlug ( BD biosciences ) incubation . Afterwards , cells were harvested and stained with anti-NHP CD3 , CD4 , CD8 , huCD4 and intracellular IFNγ . Jurkat cells ( clone E6-1 , AIDSreagent ) were either untransduced or transduced with 2MOI CD4CAR ( without C46 ) or C46CD4CAR for 2 days and infected with HIV-1NL4 . 3 ( 100 ng p24/106 cells ) for 3 days . Afterwards , cells were intracellularly stained with anti-Gag ( clone KC57 ) and analyzed by flow cytometry . PBMCs were isolated from transplanted pigtail macaque peripheral blood collected during primary SHIV infection prior to cART treatment , and after cART withdrawal . Total PBMC were co-incubated with unstimulated U1 ( HIV latent cell line , Env- ) or PMA activated U1 ( HIV Env+ ) at 3:1 , 10:1 and 20:1 ratios for 10 hours . In order to quantify target killing , target cells were pre-stained with celltrace Vioblue ( ThermoFisher Scientific ) prior to co-incubation , and percent killing was calculated as the percent loss of live celltrace Vioblue+ target cells after co-incubation with effector PBMC . Plasma was isolated from transplanted pigtail macaque peripheral blood collected during primary SHIV infection prior to cART treatment , and after cART withdrawal and frozen at -80 . After necropsy , plasma was thawed and 25ul ( max ) was used to carry out Milliplex none-human-primate multiplex assay ( EMD Millipore ) for detection of proinflammatory cytokines IFNγ , IL-1β , IL-2 , IL-6 , MCP-1 , MIP1β , MIP1α , sCD40L , TNFα . Detection range of cytokines are between 2 . 44 to 10 , 000 pg/ml . Data shown was average of 2 replicate wells .
|
Hematopoietic Stem/Progenitor Cell ( HSPC ) based gene therapy can be used to treat many infectious and genetic diseases . Here , we used an HSPC-based approach to redirect and enhance host immunity against HIV-1 . We engineered HSPCs to carry chimeric antigen receptor ( CAR ) genes that detect and destroy HIV-infected cells . CAR therapy has shown huge potential in the treatment of cancer , but has only been applied in peripheral blood T-cells . HSPC-based CAR therapy has several benefits over T cell gene therapy , as it allows for normal T cell development , selection , and persistence of the engineered cells for the lifetime of the patient . We used a CAR molecule that hijacks the essential interaction between the virus and the cell surface molecule CD4 to redirect HSPC-derived T-cells against infected cells . We observed >2 years of stable production of CAR-expressing cells without any adverse events , and wide distribution of these cells in lymphoid tissues and gastrointestinal tract , which are major anatomic sites for HIV replication and persistence in suppressed patients . Most importantly , HSPC-derived CAR T-cells functionally responded to infected cells . This study demonstrates for the first time the safety and feasibility of HSPC based therapy utilizing an HIV-specific CAR for suppressed HIV infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"blood",
"cells",
"hiv",
"infections",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"immune",
"physiology",
"viral",
"transmission",
"and",
"infection",
"cytokines",
"pathogens",
"immunology",
"microbiology",
"vertebrates",
"cloning",
"animals",
"mammals",
"retroviruses",
"primates",
"immunodeficiency",
"viruses",
"viruses",
"developmental",
"biology",
"rna",
"viruses",
"molecular",
"development",
"molecular",
"biology",
"techniques",
"viral",
"load",
"old",
"world",
"monkeys",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"white",
"blood",
"cells",
"monkeys",
"animal",
"cells",
"medical",
"microbiology",
"hiv",
"t",
"cells",
"microbial",
"pathogens",
"molecular",
"biology",
"macaque",
"immune",
"system",
"eukaryota",
"blood",
"cell",
"biology",
"anatomy",
"virology",
"viral",
"pathogens",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"viral",
"diseases",
"lentivirus",
"amniotes",
"organisms"
] |
2017
|
Long-term persistence and function of hematopoietic stem cell-derived chimeric antigen receptor T cells in a nonhuman primate model of HIV/AIDS
|
How organisms determine particular organ numbers is a fundamental key to the development of precise body structures; however , the developmental mechanisms underlying organ-number determination are unclear . In many eudicot plants , the primordia of sepals and petals ( the floral organs ) first arise sequentially at the edge of a circular , undifferentiated region called the floral meristem , and later transition into a concentric arrangement called a whorl , which includes four or five organs . The properties controlling the transition to whorls comprising particular numbers of organs is little explored . We propose a development-based model of floral organ-number determination , improving upon earlier models of plant phyllotaxis that assumed two developmental processes: the sequential initiation of primordia in the least crowded space around the meristem and the constant growth of the tip of the stem . By introducing mutual repulsion among primordia into the growth process , we numerically and analytically show that the whorled arrangement emerges spontaneously from the sequential initiation of primordia . Moreover , by allowing the strength of the inhibition exerted by each primordium to decrease as the primordium ages , we show that pentamerous whorls , in which the angular and radial positions of the primordia are consistent with those observed in sepal and petal primordia in Silene coeli-rosa , Caryophyllaceae , become the dominant arrangement . The organ number within the outmost whorl , corresponding to the sepals , takes a value of four or five in a much wider parameter space than that in which it takes a value of six or seven . These results suggest that mutual repulsion among primordia during growth and a temporal decrease in the strength of the inhibition during initiation are required for the development of the tetramerous and pentamerous whorls common in eudicots .
How to determine the numbers of body parts is a fundamental problem for the development of complete body structures in multicellular organisms . Digit numbers in vertebrates are evolutionarily optimized for the specific demands of the organism [1]; the body-segment number in insects is constant despite the evolutionarily diversified gene regulation in each segment [2–4]; and five petals are indispensable to forming the butterfly-like shape that is unique to legume flowers [5] . Studies of animal structures , such as vertebrate limbs and insect segments , strongly suggest that crosstalk between pre-patterns ( e . g . , morphogen gradients ) and self-organizing patterns underlies the developmental process of organ-number determination [6–13] . In plant development , a self-organization based on the polar transport of the phytohormone auxin [14–16] is conserved among seed plants [17] and seems to be the main regulator of the development of a hierarchal body plan , called a shoot , consisting of a stem and lateral organs such as leaves . The number of concentration peaks in most self-organizing patterns , such as Turing pattern and the mechanisms proposed for plant-pattern formation , is proportional to the field size [15 , 18 , 19] . Despite having a diversified field size for floral-organ patterning , the eudicots , the most diverged clade among plants , commonly have pentamerous or tetramerous flowers containing five or four sepals and petals ( the outer floral organs ) , respectively , and rarely have other numbers of organs [20 , 21] . Here , we focus on the developmental properties that so precisely and universally determine the floral organ numbers through self-organizing processes . Phyllotaxis , the arrangement of leaves around the stem , provides insight into floral development , because studies of floral organ-identity determination [22] have verified Goethe’s foliar theory , which insists that a flower is a short shoot with specialized leaves [23] . Phyllotaxis is mainly classified into two types: spiral phyllotaxis , which has a constant divergence angle and internode length , and whorled phyllotaxis , which has several leaves at the same level of a stem [24] . For spiral phyllotaxis , Hofmeister described a hypothesis of pattern formation in 1868 [24] , which we summarize in three basic rules: the time periodicity of primordia initiation , the initiation of a primordium at the largest available space at the edge of the meristem ( the undifferentiated stem-cell region ) , and the relative movement of primordia in a centrifugal direction from the apex due to the growth of the stem tip . Following that hypothesis , numerous mathematical models incorporating contact pressure [25 , 26] , inhibitor diffusion [27] , reaction-diffusion [18 , 28] , and mechanical buckling of the epidermis [29 , 30] were proposed to explain the observed phyllotactic patterns . Over the past ten years , these mathematical models were tested and interpreted in light of modern molecular biology . Several studies have suggested that the competitive polar transport of the auxin accounts for two of Hofmeister’s rules , the periodicity of initiation and the initiation at the largest space , which together are capable of reproducing both spiral phyllotaxis and whorled phyllotaxis [15 , 16 , 31] . Despite their simple rules and uncertain molecular basis , the phyllotaxis models can account for several of the quantitative properties observed in organ patterning . For example , one model showed that the divergence angle between successive leaves is 180 degrees for the first and second leaves , 90 degrees for the second and third leaves , and oscillating thereafter , converging to the golden angle , 137 . 5 degrees , which agrees with the phyllotaxis of true leaves in Arabidopsis thaliana after the two cotyledons [32 , 33] . Similar oscillatory convergence to a particular divergence angle occurs in the sepal primordia of the pentamerous flower of Silene coeli-rosa , Caryophyllaceae . In S . coeli-rosa , the divergence angle is 156 degrees at first , and then it oscillates , converging on 144 degrees [34] . The golden angle also appears in the floral organs of several Ranunculaceae species [35 , 36] . The agreements between the phyllotaxis models and actual floral development suggest that mathematical models can give useful clues to the underlying mechanisms of not only phyllotaxis but also floral organ patterning . There are at least three fundamental differences , however , between real floral development and the phyllotaxis models . The first difference is the assumption of constant primordium displacement during tip growth , which comes from Hofmeister’s hypothesis and has been incorporated into most phyllotaxis models . Although the helical initiation has been thought to always result in spiral phyllotaxis , many eudicots form the whorled-type sepal arrangements in their blooming flowers subsequent to helical initiation [37] ( Fig 1; e . g . , Caryophyllaceae [34] , Solanaceae [38] , Nitrariaceae [39] , and Rosaceae [40] ) . The remnants of helical initiation are more obvious in the pseudo-whorls ( e . g . , Ranunculaceae [41] ) , where the distance between each organ primordium and the floral center varies slightly even in the whorls of mature flowers , which usually have more varied floral organ numbers [20 , 35] , suggesting that post-meristematic modifications of primordia positions [42] play an essential role in generating the whorled arrangement and determining the floral organ number during floral development . In contrast , most phyllotaxis models have assumed constant growth of the primordia , so that the whorls appear only after the simultaneous initiation of several primordia [19] . The second difference comes from the fact that floral development is a transient process , whereas most phyllotaxis models have focused on the steady state of the divergence angle . Although the golden angle ( 137 . 5 degrees ) is quite close to the inner angle of regular pentagon ( 144 degrees ) , the developmental convergence from 180 degrees ( cotyledon ) to 137–144 degrees in phyllotaxis requires the initiation of more than five primordia , both in A . thaliana leaves and in the mathematical models [16 , 33] . In contrast , the divergence angle between the second and third sepal primordia in pentamerous eudicot flower development is already close to 144 degrees [34] . The third difference comes from the accuracy of the floral organ number in many eudicots . Although the polar auxin-transport model reproduced both wild-type and mutant A . thaliana floral organ positioning [43] , the organ number in the model was more variable , even with an identical parameter set ( Fig 3 in [43] ) , than that in experimental observations ( Table 1 in [44] ) . Moreover , among eudicot species , the appearance of pentamerous flowers is robust , despite the diversity of the meristem size and the outer structures , including the number and position of outside organs such as bracts [20] . Together , the differences between real floral development and previous phyllotaxis models indicate that floral development requires additional mechanisms to determine the particular organ number . To resolve the inconsistencies between the earlier models and actual floral development , we set out a simple modeling framework , integrating Hofmeister’s rules with two additional assumptions , namely , the repulsion between primordia that can repress primordium growth and the temporal decrease in initiation inhibition of new primordium , which were proposed independently in the contact pressure model [25 , 45 , 46] and the inhibitory field model [33 , 47 , 48] , respectively , for phyllotaxis . First , when we incorporated mutual repulsion among primordia into the growth process , a whorled-type pattern emerged spontaneously following the sequential initiation of primordia . The mutual repulsion obstructed the radial movement of a new primordium after a specific number of primordia arose , causing a new whorl to emerge . The number of primordia in the first whorl tended to be four or eight . Second , when we assumed that older primordia have less influence on the initiation of a new primordium , the pentamerous whorl arrangement , which is the most common arrangement in eudicot flowers , became dominant . We analytically show the conditions for the development of tetramerous and pentamerous whorls , and we predict possible molecular and physiological underpinnings .
Following the earlier models [49] , we represented the meristem as a circular disc with radius R0 and the primordia as points ( Fig 2A ) . A new primordium arises at the point along the edge of the meristem ( R0 , θ ) , in polar coordinate with the origin at the meristem center , where θ gives the minimum value of the inhibition potential Uini . As one of the simplest setups for sequential initiation [37] , we followed the assumption of earlier models for spiral phyllotaxis [49] , which state that new primordia arise sequentially with time intervals τ , as opposed to the simultaneous initiation studied previously for whorled phyllotaxis [19] ( Fig 1 ) . Although the structures outside of the flower , such as bracts and other flowers , as well as the position of the inflorescence axis , may affect the position of organ primordia , the pentamerous whorls appear despite their various arrangement [20] . Therefore , as the first step of modelling of floral organ arrangement , we assumed that whorl formation is independent of any positional information from structures outside of the flower . Thus , we calculated the inhibition potential only from floral organ primordia which are derived from a single floral meristem . The potential functions for the initiation inhibition by preexisting primordia have been extensively analyzed in phyllotaxis models [16 , 47 , 49] . The potential decreases with increasing distance between an initiating primordium and the preexisting primordia account for the diffusion of inhibitors secreted by the preexisting primordia [27 , 50] , and the polar auxin transport in the epidermal layer , as proposed in previous models of phyllotaxis [15 , 16 , 31] and the flowers [43] . We employed an exponential function exp ( −dij/λini ) as a function of θ , where dij denotes the distance between a new primordium i and a preexisting primordium j at ( rj , θj ) as d i j = R 0 2 + r j 2 - 2 R 0 r j cos ( θ - θ j ) ⋅ ( 1 ) The function decreases spatially through the decay length λini exponentially , induced by a mechanism proposed for the polar auxin transport , i . e . , the up-the-gradient model [15 , 16] . Up-the-gradient positive feedback amplifies local auxin concentration maxima and depletes auxin from the surrounding epidermis , causing spatially periodic concentration peaks to self-organize [15 , 16] and thus determine the initiation position of the primordia [51] . The amplification and depletion work as short-range activation and long-range inhibition , respectively [52] , which are common to Turing patterns of reaction-diffusion systems [18] . Since the interaction of local maxima in the reaction-diffusion systems follows the exponential potential [53 , 54] , the up-the-gradient model likely explains the exponential potential between the auxin maxima , while the rigorous derivation requires further research . The decay length λini depends not only on the ratio of the auxin diffusion constant and the polar auxin-transport rate [15] but also on other biochemical parameters for polar transport and the underlying intracellular PIN1 cycling [55] . Another mechanism , referred to as the with-the-flux model [56 , 57] , has been proposed for the polar auxin transport . Although with-the-flux positive feedback can also produce spatial periodicity , the primordia position corresponds to auxin minima [57] , which is inconsistent with observations [51] . On the other hand , the with-the-flux mechanism can explain auxin drain from the epidermal layer of the primordia to internal tissue [58] . Since the drain gets stronger as the primordia mature [58 , 59] , the auxin drain could cause decay of the potential depending on the primordia age . The auxin decrease in maturing organs can also be caused by controlling auxin biosynthesis [60 , 61] . Therefore , we integrated another assumption , namely that the inhibition potential decreases exponentially with the primordia age at the decay rate α ( Fig 2B ) . Temporally decaying inhibition was proposed previously to represent the degradation of some inhibitors [47 , 48] and account for various types of phyllotaxis by simple extension of the inhibitory field model [33] . Taken together , the potential at the initiation of the i-th primordium is given by U i n i ( θ ) = ∑ j = 1 i - 1 exp ( - α ( i - j - 1 ) ) exp ( - d i j λ i n i ) ⋅ ( 2 ) Most phyllotaxis models have assumed , based on Hofmeister’s hypothesis , that the primordia move outward at a constant radial drift depending only on the distance from the floral center without angular displacement , which makes helical initiation result in spiral phyllotaxis [49] . Here , we assumed instead that all primordia repel each other , even after the initiation , except for movement into the meristematic zone ( Fig 2C ) following observation of the absence of auxin ( DR5 expression ) maxima at the center of the floral bud ( e . g . , [62] ) . Even at the peripheral zone away from the meristem , the growth is not limited . Hence there is no upper limit for the distance between primordia and the center . The repulsion exerted on the k-th primordium is represented by another exponentially decaying potential when there are i primordia ( 1 ≦ k ≦ i ) : U g , k ( r , θ ) = ∑ j = 1 , j ≠ k i exp ( - d k j λ g ) , ( 3 ) where the decay length , introduced as λg , can differ from λini . The primordia descend along the gradient of potential Ug to find a location with weaker repulsion . The continuous repulsion can account for post-meristematic events such as the mechanical stress on epidermal cells caused by the enlargement of primordia [63 , 64] or the gene expression that regulates the primordial boundary [42] . The present formulation ( Eq 3 ) is similar to the contact pressure model , which has been proposed for re-correcting the divergence angle after initiation [25 , 45 , 46] . Another type of post-initiation angular rearrangement has been modeled as a function of the primordia age employed as i −j −1 in the present model ( Eq 2 ) and the distance between primordia with some stochasticity [65] . Eq 3 accounts for not only the angular rearrangement but also the radial rearrangement with stochasticity in both directions as will be described in the next subsection . We modeled the initiation process numerically by calculating the potential Uini ( Eq 2 ) for angular position θ incremented by 0 . 1 degree on the edge of the circular meristem . We introduced a new primordium at the position where the value of Uini took the minimum , provided that the first primordium is initiated at θ = 0 . We modeled the growth process by using a Monte Carlo method [66] to calculate the movement of primordia in the outside of the meristem depending on the potential Ug , k ( Eq 3 , Fig 2C ) . After the introduction of a new primordium , we randomly chose one primordium indexed by k from among the existing primordia and virtually moved its position ( rk , θk ) to a new position ( r k ′ , θ k ′ ) in the outer meristem ( r k , r k ′ ≥ R 0 ) . The new radius r k ′ and the angle θ k ′ were chosen randomly following a two-dimensional Gaussian distribution whose mean and standard deviation were given by the previous position ( rk , θk ) and by two independent parameters , ( σr , σθ/rk ) , respectively . Whether or not the k-th primordium moved to the new position was determined by the Metropolis algorithm [66]; the primordium moved if the growth potential ( Eq 3 ) of the new position was lower than that of the previous position ( i . e . , U g , k ( r k ′ , θ k ′ ) < U g , k ( r k , θ k ) ) . Otherwise , it moved with the probability given by P M P = exp ( - β Δ U g ) , ( 4 ) where Δ U g = U g , k ( r k ′ , θ k ′ ) − U g , k ( r k , θ k ) and β is a parameter for stochasticity . This stochasticity represents a random walk biased by the repulsion potential . A case PMP = 0 represents that primordia movement always follows the potential ( ΔUg < 0 ) . The first primordium stays at the meristem edge r = R0 until the second one arises when PMP = 0 because the growth potential is absent , while it can move randomly outside of the meristem when PMP ≠ 0 . To maintain the physical time interval of the initiation process at τ steps for each primordium , the number of iteration steps in the Monte Carlo simulation during each initiation interval was set to iτ , where i denotes the number of the primordia . We also studied the movement following Ug by numerical integration ( fourth-order Runge-Kutta method ) of ordinary differential equations to confirm the independence of the numerical methods ( S1 Fig ) . All our programs were written in the C programming language and used the Mersenne Twister pseudo-random number generator ( http://www . math . sci . hiroshima-u . ac . jp/m-mat/MT/emt . html ) [67] . Because the initiation time interval is constant , one possible scenario for forming a whorled pattern should involve decreasing or arresting the radial displacement of primordia ( Fig 1 , forth row ) . Therefore , we focused on the change in radial position and velocity to find the whorled arrangement , while angular positions were not taken into account in the present manuscript .
Numerical simulations showed that several whorls self-organized following the sequential initiation of primordia . Although several previous phyllotaxis models showed the transition between a spiral arrangement following sequential initiation and a whorled arrangement following simultaneous initiation [15 , 16 , 19] , they were not able to reproduce the emergence of a whorled arrangement following sequential initiation , which is the situation observed in many eudicot flowers ( Fig 1 ) [34 , 37 , 38 , 40 , 41] . In the present model , a tetramerous whorl appeared spontaneously that exhibited four primordia almost equidistant from the meristem center ( Fig 2D , left and middle ) , by arresting radial movement of the fifth primordium at the meristem edge until the seventh primordium arose ( arrowhead in Fig 2D , right ) . Likewise , subsequent primordia produced the same gap in radial distance for every four primordia ( Fig 2D , middle and right ) , leading to several whorls comprising an identical number of primordia ( Fig 2D ) . The radial positions of all primordia were highly reproducible despite stochasticity in the growth process ( error bars in Fig 2D–2F , middle and right ) . Therefore , we identified the whorled arrangement by radial displacement arrest ( arrowhead in Fig 2D , right ) . The initiation order and angle of the first tetramerous whorl in the model reproduced those observed in A . thaliana sepals [68] ( S2A Fig ) . The first primordium scarcely moved from the initiation point until the second primordium arose because growth repulsion was absent . The second primordium arose opposite the first , whereas the third and fourth primordia arose perpendicular to the preceding two . The angular position of the primordia did not change once the whorl was established because the primordia within a whorl blocked the angular displacement by the growth potential Ug ( S3 Fig ) . Introducing mutual repulsion among the primordia throughout the growth process caused the whorled arrangement to spontaneously emerge ( Fig 2D ) . This was in contrast to the model of constant growth in which all primordia move away depending only on the distance from the floral apex [49] . A study of post-meristematic regulation by the organ-boundary gene CUP-SHAPED COTYLEDON2 ( CUC2 ) showed that A . thaliana plants up-regulating CUC2 gene have an enlarged primordial margin and have whorled-like phyllotaxis following the normal helical initiation of primordia [42] , suggesting that repulsive interactions among primordia after initiation are responsible for the formation of the floral whorls . In the present model , the meristem size R0 controls the transition from non-whorled ( Fig 2E ) to whorled arrangement ( Fig 2D ) . Radial spacing of the primordia was regular when R0 was small ( Fig 2E , middle ) because the older primordia pushed any new primordium across the meristem ( Fig 2E , left ) , causing continuous movement at the same rate ( Fig 2E , right ) . Above a threshold meristem size R0 , a tetramerous whorl appeared spontaneously . The primordium number within each whorl increased up to eight with increasing R0 , but the number tended to be more variable ( S2B Fig ) . In the A . thaliana mutant wuschel , which has a decreased meristem size , the pattern of four sepals does not have square positions at the stage when the wild-type plant forms a tetramerous sepal whorl [69] . Conversely , the clavata mutant , which has an increased meristem size , has excessive floral organs with larger variation [69] . Our model consistently reproduced not only the transition from the non-whorled arrangement ( Fig 2E ) to the tetramerous whorled arrangement ( Fig 2D ) but also the variable increase in the primordia number within a whorl as the meristem size R0 increased . The pentamerous whorl stably appeared in the presence of temporal decay of initiation inhibition ( α > 0 in Eq 2 ) . The whorls comprising five primordia appeared in the same manner as the tetramerous whorls , namely , via the locking of the sixth primordium at the initiation site ( Fig 2F , right; S2C Fig ) . In order to study the organ number within each whorl extensively , known as the merosity [70] , we counted the number of primordia existing prior to the arrest of primordium displacement , which corresponds to the merosity of the first whorl ( arrowheads in Fig 2D and 2F , right ) . We defined arrest of primordium displacement as occurring when the ratio of the initial radial velocity of a new primordium immediately after initiation to that of the previous primordium was lower than 0 . 2 . The definition does not affect the following results as long as the ratio is between 0 . 1 and 0 . 6 . We found that the key parameter for merosity is the relative value of R0 normalized by the average radial velocity V = σ r / 2 π ( see S1 Text ) and the initiation time interval τ ( Fig 3 ) . The arrest of radial displacement did not occur below a threshold of R0/Vτ ( the left region colored red in Fig 3A ) , whereas the whorled arrangement appeared above the threshold value of R0/Vτ . As R0/Vτ increased further , tetramery , pentamery , hexamery , heptamery , and octamery appeared , successively ( Fig 3A ) . The present model showed dominance of special merosity , i . e . , tetramery and octamery in the absence of temporal decay of inhibition ( α = 0 in Eq 2; Fig 3A ) ; pentamery in the presence of temporal decay ( α > 0; Fig 3B and 3C ) , in contrast to previous phyllotaxis models for whorled arrangement in which the parameter region leading to each level of merosity decreased monotonically with increasing merosity [19] . The major difference between α = 0 and α > 0 was that θ3 , the angular position of the third primordium , took an average value of 90 degrees when α = 0 ( arrowhead in Fig 3A bottom magenta panel ) and decreased significantly as α increased ( arrowhead in Fig 3A bottom cyan panel ) . In a pentamerous flower Silene coeli-rosa , the third primordium is located closer to the first primordium than the second one [34] . This is consistent with the third primordium position at α > 0 , indicating the necessity of α , as we will discuss in the next section . The parameter region R0/Vτ for pentamery expanded with increasing α , whereas the border between the whorled and non-whorled arrangements was weakly dependent on α ( Fig 3C ) . The tetramery , pentamery , and octamery arrangements were more robust to R0/Vτ and α than the hexamery and heptamery arrangements . Dominance of the particular number also appears in the ray-florets within a head inflorescence of Asteraceae [71] , in which radial positions show the whorled-type arrangement [72] . Meanwhile , the leaf number in a single vegetative pseudo-whorl transits between two to six by hormonal control without any preference [73] . Moreover , the transition between the different merosities occurred directly , without the transient appearance of the non-whorled arrangement . This is in contrast to an earlier model [19] in which the transition between different merosity always involved transient spiral phyllotaxis . The fact that the merosity can change while keeping its whorled nature in flowers ( e . g . , the flowers of Trientalis europaea[74] ) supports our results . To our knowledge , ours is the first model showing direct transitions between whorled patterns with different merosities as well as preferences for tetramery and pentamery , the most common merosities in eudicot flowers . To further validate our model of the pentamerous whorl arrangement , we quantitatively compared its results with the radial distances and divergence angles in eudicot flowers . Here we focus on a Scanning Electron Microscope ( SEM ) image of the floral meristem of S . coeli-rosa , Caryophyllaceae ( Fig 4A–4C ) [34] , because S . coeli-rosa exhibits not only five sepals and five petals in alternate positions , which is the most common arrangement in eudicots , but also the helical initiation of these primordia , which we targeted in the present model . In addition , to our knowledge , this report by Lyndon is the only publication showing a developmental sequence for both the divergence angle Δθk , k+1 = θk+1 −θk ( 0 ≤ Δθk , k+1 < 360 ) and the ratio of the radial position , rk/rk+1 , referred to as the plastochron ratio [75] , in eudicot floral organs . Reconstructing such developmental sequences of both radial and angular positions is an unprecedented theoretical challenge , while those which describe the angular position alone for the ontogeny of spiral phyllotaxis ( 180 degree , 90 degree and finally convergence to 137 degree [16 , 33]; the ‘M-shaped’ motif , i . e . , 137 , 275 , 225 , 275 and 137 degrees [76 , 77] ) have been reproduced numerically . By substituting the initial divergence angle between the first and second sepals of S . coeli-rosa into Δθ1 , 2 = 156 but not any plastochron data into the simulation ( θ1 = 0 and θ2 = 156 degree ) , we numerically calculated the positions of the subsequent organs ( Fig 4D ) . The observed divergence angle Δθ2 , 3 = 132 degree indicates α > 0 , because Δθ2 , 3 = Δθ1 , 3 = ( 360−156 ) /2 = 102 degree at α = 0 , in the present model setting r1 ≅ r2 . Even when r1 > r2 , the divergence angle was calculated as Δθ2 , 3 = 113 degree ( r1 = R0+2Vτ , r2 = R0+Vτ , R0 = 1 , Vτ = 0 . 14 , and λini = 0 . 05 estimated from the S . coeli-rosa SEM image [34]; see S4 Fig for detail ) , which is still less than the observed value . As α became larger , the inhibition from the second primordium became stronger than that from the first one , making Δθ2 , 3 consistent with the observed value in S . coeli-rosa ( Fig 4E , top ) . For the subsequent sepals and petals , the model faithfully reproduced the period-five oscillation of the divergence angle and the plastochron ratio until the ninth primordium ( Fig 4E ) , notably in the deviation of the divergence angle from regular pentagon ( 144 degree ) and the increase of plastochron ratio at the boundary between the sepal and petal whorls . Moreover , a similar increase in the plastochron ratio occurred weakly between the second and third primordia in the first whorl ( closed arrowhead in Fig 4E ) , indicating a hierarchically whorled arrangement ( i . e . , whorls within a whorl ) . Such weak separation of the two outer primordia from the three inner ones within a whorl is consistent with the quincuncial pattern of sepal aestivation that reflects spiral initiation in many of eudicots with pentamerous flowers ( e . g . , Fig 2D–E in [21] ) . Even with an identical set of parameters , the order of initiation in the first pentamerous whorl can vary depending on the stochasticity in the growth process . The variations of the initiation order in simulations may be caused by the absence of the outer structure , because the axillary bud seems to act as a positional information for the first primordia in S . coeli-rosa floral development ( Fig 4B ) . The positioning of the five primordia in the first whorl was reproducible in 70% of the numerical replicates , within less than 20 degrees of that in S . coeli-rosa or that of the angles in a regular pentagon . Mismatches in the inner structure ( from the tenth primordium , i . e . , the last primordium in petal whorl ) might be due to an increase in the rate of successive primordia initiation later in development [35] , which we did not assume in our model . The agreements between our model and actual S . coeli-rosa development of sepals and petals in both the angular and the radial positions suggests that the S . coeli-rosa pentamerous whorls are caused by decreasing inhibition from older primordia . A possible mechanism to arrest the radial displacement of a new primordium , a key process for whorl formation ( arrowheads in Fig 2D and 2F ) , involves an inward-directed gradient of the growth potential Ug , k ( Eq 3 ) of a new primordium so that its radial movement is prevented . To confirm this for tetramerous whorl formation ( Fig 3A ) , we analytically derived the parameter region such that the radial gradient of the growth potential at the angle of the fifth primordium Ug , 5 ( Eq 3 ) , which is determined by the positions of the preceding four primordia , is inward-directed . For ease in the analytical calculation , we set α = 0 and PMP = 0 . The first four primordia positions were intuitively estimated ( see S2 Text ) as r 1 = R 0 + 3 τ V , θ 1 = 0 r 2 = R 0 + 3 τ V , θ 2 = 180 r 3 = R 0 + 2 τ V , θ 3 = 90 r 4 = R 0 + τ V , θ 4 = 270 , ( 5 ) which agreed with the numerical results with an error of less than several percent regardless of the parameter spaces . Hereafter we demonstrate a case Vτ = 6 . 0 . The position of the fifth primordium derived from the positions of four existing primordia ( Eq 5 ) becomes θ5 = 90 when R0 ≤ 2 , whereas θ5 ∼ 135 when R0 > 2 ( S5 Fig ) . Next , we calculated the potential for the fifth primordium in radial direction by substituting Eq 5 and the position of the fifth primordium θ5 into Eq 3 . The function becomes U g , 5 ( r , θ 5 ) = ∑ j = 1 4 exp ( - d 5 j λ g ) = ∑ j = 1 4 exp ( - r j 2 + r 2 - 2 r j r cos ( θ j - θ 5 ) λ g ) ⋅ ( 6 ) The potential exhibits a unimodal ( 2 < R0 < 10; Fig 5A ) or bi-modal ( R0 < 2 , R0 > 10; Fig 5B and 5C ) shape . At R0 < 10 , the potential gradient at the initiation position of the fifth primordium ∂Ug , 5 ( r , θ5 ) /∂r∣r = R0 is outward-directed ( Fig 5A ) , providing almost constant growth resulting a non-whorled arrangement in the simulations ( Fig 3A , red region ) . At R0 > 10 , we defined the radial position of the local maximum closest to the fifth primordium as rmax ( open arrowhead in Fig 5B and 5C; red squares in the upper half of Fig 5D ) and the local minimum as rmin ( blue circles in Fig 5D; 0 < rmin < rmax ) . The potential gradient ∂Ug , 5 ( r , θ5 ) /∂r∣r = R0 has a negative value when R0 < rmin or rmax < R0 ( Fig 5C ) , causing the fifth primordium to constantly move outward . On the other hand , the potential gradient is positive , i . e . , directed inward ( Fig 5B ) , when rmin < R0 < rmax ( between the two solid arrowheads in Fig 5D ) , causing the arrest of radial movement of the fifth primordium . The values of rmin and rmax , analytically calculated as function of R0 and τ ( solid black line in Fig 5E ) , were faithfully consistent with the parameter boundaries between the non-whorled pattern and the tetramerous-whorled pattern and between the tetramerous-whorled and pentamerous-whorled patterns , respectively , in the numerical simulations ( Fig 5E ) . The assumption r1 = r2 ( Eq 5 ) according to our numerical results ( Fig 2D ) , which is a similar setup to co-initiation of two primordia , is not a necessary condition for consistency ( S6 Fig ) . Thus the inward-directed gradient of the growth potential ( Eq 3 ) , which works as a barrier to arrest the outward displacement of the fifth primordium , causes the formation of tetramerous whorl . The inward radial gradient of the potential Ug , k ( Eq 3 ) also accounted for the emergence of pentamerous whorls at α > 0 . Unlike the case of α = 0 , the angular position of the third primordium θ3 at the global minimum of Uini decreases from 90 degrees as α increases ( Fig 6A ) . For example , the recursive calculations for the minimum of Uini gave the angular positions of the two subsequent primordia , θ3 ≅ 62 and θ4 ≅ 267 , respectively , at α = 2 . 0 ( Vτ = 6 . 0 , R0 = 20 . 0 , and PMP = 0 ) . Those angular positions were consistent with the numerical results ( e . g . , Fig 2F and S2B Fig ) . The gradient of the growth potential ∂Ug , 5 ( r , θ5 ) /∂r at the edge of the meristem for the fifth primordium that arises at θ5 ≅ 129 is negative ( Fig 6B ) . Therefore , the fifth primordium moves outward at constant velocity so that the tetramerous whorl is unlikely to emerge . The inward-directed potential at the position of the new primordium first appears when the sixth primordium arises around 343 degrees , which was derived by the recursive calculation ( Fig 6C ) . The first primordium ( the rightmost potential peak in Fig 6C ) prevents the outward movement of the sixth primordium ( red circle in Fig 6C ) . Arrest of radial displacement of the sixth primordium is maintained until the seventh primordium arises to allow the radial gap between these primordia to appear ( i . e . , a pentamerous whorl emerged ) . After the appearance , the growth potential gradients of the sixth and the seventh primordia become outward-directed , providing their constant growth with keeping the radial gap to the first whorl . Likewise , the other merosities can be explained by similar recursive calculations of the angular position from the initiation potential ( Eq 2 ) and the radial gradient of the growth potential ( Eq 3 ) . Based on these analytical results ( Figs 5 and 6 ) and the dimensionless parameter G = τV/R0 , which represents the natural logarithm of the average plastochron ratio [49 , 75] , we quantitatively compared the present model against previous phyllotaxis models assuming simultaneous initiation based on the initiation potential [19] . The tetramerous and pentamerous whorls appeared in , at most , 1 . 3-fold and 1 . 2-fold ranges of G , respectively , in the earlier study ( Fig 4D in [19] ) ; however , they appeared in much wider ranges in our model ( i . e . , 3-fold to 5-fold and 1 . 2-fold to 5-fold ranges of G , respectively; Fig 3C ) . Here , another key parameter is the temporal decay rate of the initiation inhibition α that shorten the transient process approaching to the golden angle ( Fig 6A ) than those of spiral phyllotaxis [32 , 33] . λini , representing the gradient of the initiation potential ( Eq 2 ) , little affected the border between the whorled and non-whorled arrangements at α = 0 ( Fig 7A and 7B ) ; λini affected the border only when α ≠ 0 ( Fig 7C and 7D ) . The independency of λini at α = 0 is consistent with the result shown by the previous model , which did not incorporate temporal decay of the potential and indicated that the phyllotactic pattern depends little on the functional type of initiation potential [49] . On the other hand , the gradient of the growth potential ( Eq 3 ) regulated by λg caused a drastic transition between the whorled and non-whorled arrangements ( Fig 7E and 7F ) . Unlike G , λini , and α ( Fig 6A ) , λg hardly affects the angular position , as demonstrated in the previous sections , but it controls how far the growth potential works as a barrier to determine the merosities of the whorls ( Fig 7E and 7F ) . Thus , λg , α , and G differentially regulate phyllotaxis of the floral organs , suggesting the involvement of distinct molecular or physiological underpinnings . We have seen that both the mutual repulsion of growth regulated by λg and the temporal decay of initiation inhibition controlled by α are responsible for the formation of tetramerous and pentamerous whorls following sequential initiation . These mechanisms can be experimentally verified by tuning λg and α . Here , we discuss several candidates for the molecular and physiological underpinnings . Future studies should also clarify the limits and applicability of the common developmental principle elucidated here by exploring more complex development in a wide variety of flowers . Because our model assumes sequential initiation of the primordia , it does not cover the floral development of all eudicots; sepal primordia arise simultaneously in some eudicot clades ( Fig 1; e . g . , mimosoid legume [94] ) . Likewise , in later development , several primordia arise at once in the stamen and carpel whorls ( e . g . , Ranunculaceae [35] ) . The transitions between simultaneous and sequential development have two additional intriguing implications for evolutionary developmental biology . First , the initiation types may affect the stochastic variation of floral organ numbers , possibly caused by the absence or presence of pseudo-whorls ( Fig 1 ) and the noisy expression domain of homeotic genes [95] . Second , such transitions occur even in animal body segmentation [3 , 4] , possibly caused by evolution of both gene regulatory network topologies and embryonic growth [7 , 9–11] . The limitations of the model can be reduced by introducing initiation whenever and wherever the potential ( Eq 2 ) is below a threshold , allowing simultaneous as well as sequential initiation [19] . The threshold model exhibiting both types of initiation does not by itself result in the dominance of particular merosities [19] . Incorporating two mechanisms , mutual growth repulsion and temporally decreasing inhibition at the point of initiation , into the threshold model could explain the dominance of particular merosities following both the sequential and the simultaneous initiation of floral organ primordia ( Fig 1 ) . Another problem is the absence of trimerous whorls in the present model ( Fig 3 ) . The transition between the trimery and tetramery or pentamery , and vice versa , occurred multiple times during the evolution of angiosperms . Therefore , trimerous flowers are scattered across the basal angiosperms , monocots , and a few families of eudicots [96 , 97] . Elucidating the developmental mechanisms underlying the transitions between the different merosities , as well as those between sequential and simultaneous initiation , will be an important avenue for future studies . One problem in determining floral organ number is how to generate whorls comprised of a specific number of organs . By introducing a growth assumption ( i . e . , continuous repulsion among primordia throughout development , which was originally proposed as the contact pressure model [25 , 45 , 46] and is supported by experimental observations [42] ) into a dynamical model of phyllotaxis [49] , we showed that the whorled arrangement arises spontaneously from sequential initiation . Moreover , when we allowed the inhibition to decay over time [33 , 47 , 48] , pentamerous whorls became the dominant pattern . The merosity tended to be four or five in much larger parameter spaces than those in which it tended to be six or seven . The emergence of tetramerous and pentamerous whorls could be verified experimentally by tuning the two parameters α and λg .
|
Why do most eudicot flowers have either four or five petals ? This fundamental and attractive problem in botany has been little investigated . Here , we identify the properties responsible for organ-number determination in floral development using mathematical modeling . Earlier experimental and theoretical studies showed that the arrangements of preexisting organs determine where a new organ will arise . Expanding upon those studies , we integrated two interactions between floral organs: ( 1 ) spatially and temporally decreased inhibition of new organ initiation by preexisting organs , and ( 2 ) mutual repulsion among organs such that they are “pushed around” during floral development . In computer simulations incorporating such initiation inhibition and mutual repulsion , the floral organs spontaneously formed several circles , consistent with the concentric circular arrangement of sepals and petals in eudicot flowers . Each circle tended to contain four or five organs arranged in positions that agreed quantitatively with the organ positions in the pentamerous flower , Silene coeli-rosa , Caryophyllaceae . These results suggest that the temporal decay of initiation inhibition and the mutual repulsion among growing organs determine the particular organ number during eudicot floral development .
|
[
"Abstract",
"Introduction",
"Model",
"Results/Discussion"
] |
[] |
2015
|
A Dynamical Phyllotaxis Model to Determine Floral Organ Number
|
Systematic identification of protein-drug interaction networks is crucial to correlate complex modes of drug action to clinical indications . We introduce a novel computational strategy to identify protein-ligand binding profiles on a genome-wide scale and apply it to elucidating the molecular mechanisms associated with the adverse drug effects of Cholesteryl Ester Transfer Protein ( CETP ) inhibitors . CETP inhibitors are a new class of preventive therapies for the treatment of cardiovascular disease . However , clinical studies indicated that one CETP inhibitor , Torcetrapib , has deadly off-target effects as a result of hypertension , and hence it has been withdrawn from phase III clinical trials . We have identified a panel of off-targets for Torcetrapib and other CETP inhibitors from the human structural genome and map those targets to biological pathways via the literature . The predicted protein-ligand network is consistent with experimental results from multiple sources and reveals that the side-effect of CETP inhibitors is modulated through the combinatorial control of multiple interconnected pathways . Given that combinatorial control is a common phenomenon observed in many biological processes , our findings suggest that adverse drug effects might be minimized by fine-tuning multiple off-target interactions using single or multiple therapies . This work extends the scope of chemogenomics approaches and exemplifies the role that systems biology has in the future of drug discovery .
Identification of protein-ligand interaction networks on a proteome-wide scale is crucial to address a wide range of biological problems such as correlating molecular functions to physiological processes and designing safe and efficient therapeutics [1] . Recent protein-ligand interaction studies have revealed that protein targets involved in entirely different pharmacology can bind similar small molecule drugs [2]–[4] . Large scale mapping of polypharmacology interactions indicates that drug promiscuity is a common phenomenon across the proteome [5] . It has been found that approximately 35% of known drugs or leads were active against more than one target . Moreover , a significant number of promiscuous compounds ( approximately 25% ) have observed activity in completely different gene families . Such drug promiscuity presents both opportunities and challenges for modern drug discovery . On one hand , it is possible to develop high-efficacy drugs by inhibiting multiple targets [6] or to reposition existing drugs to treat different diseases [7] , [8]; on the other hand , the off-target effect may result in adverse drug reactions that account for around one-third of drug failures during development [9] . As a result , there is increasing interest in the identification of multiple targets associated with a phenotype [6] and in developing combinatorial therapies to boost clinical efficacy [10] . Chemogenomics has emerged as a new discipline to systematically establish target relationships based on the structural and biological similarity of their ligands [3] , [11]–[18] . However , the success of chemogenomics depends on the availability of bioactivity data for the receptors and their associated ligands . For new drug targets , such data are either insufficient or unavailable . Further , the adverse drug reaction may involve receptors that are not well characterized . Complementary to chemogenomics methods , we have developed a chemical systems biology approach to identifying off-target binding networks through their ligand binding sites . The method requires 3D-structure information for the protein but not the ligand , thereby extending the scope of existing chemogenomics approaches . Moreover , the identified off-target binding network is integrated with the reconstructed biological pathways so that the effect of the drug on the biological system can be understood at the system level . In brief ( see Methods for further details ) , our chemical systems biology approach proceeds as follows: 1 ) The ligand binding site of the primary target is extracted or predicted from a 3D experimental structure or homology model and characterized by a geometric potential [19] . 2 ) Off-target proteins with a similar ligand binding site to the primary target are identified across the human structural genome using a Sequence Order Independent Profile-Profile Alignment ( SOIPPA ) [20] . The atomic details of the interactions between the drug and the putative off-targets from step 2 are characterized using protein-ligand docking methods . Based on a normalized docking score the high-ranking off-targets are further investigated . 4 ) The identified panel of off-targets is subject to structural and functional cluster analysis and incorporated into a network that includes multiple metabolic , signal transduction , and gene regulation pathways . The first and second steps have been implemented in the software package SMAP , available from http://funsite . sdsc . edu . In this paper , we apply this strategy to identify and analyze a panel of unknown off-targets for Cholesteryl Ester Transfer Protein ( CETP ) inhibitors . CETP inhibitors represent a new preventive therapy for cardiovascular disease through raising HDL cholesterol . However , clinical studies have revealed that one of the CETP inhibitors , Torcetrapib , has deadly off-target effects as a result of hypertension [21]–[25] and consequently was withdrawn from phase III clinical trial . In contrast to Torcetrapib , another CETP inhibitor JTT-705 does not have unwanted side-effects that increases blood pressure [25] . In addition , JTT-705 is able to block cell proliferation and angiogenesis through Ras and P38 kinase pathways [26] . As will be shown , the multiple off-targets of these CETP inhibitors identified here are involved in both positive and negative control of stress regulation and immune response through an interconnected metabolic , signal transduction and gene regulation network . Our predictions are strongly correlated to the observed clinical and in vitro observations , providing a molecular explanation for the difference in side-effect profiles of these two CETP inhibitors . These findings suggest that adverse drug reactions might be modulated by the fine-tuning of the off-target binding network and exemplify the role of systems biology in the future of drug discovery .
The ligand binding site of CETP ( PDB id: 2OBD ) is assumed to be a long tunnel interacting with two cholesteryl oleates ( 2OB ) and two 1 , 2-dioleoyl-Sn-glycero-3-phosphocholines ( PCW ) molecules in the native state ( Fig . S1 ) , however , the exact location of inhibitor binding is unknown . Docking studies using the software Surflex [27] , eHits [28] and AutoDock [29] indicate that the CETP inhibitors are able to bind to all four sites , with a slight preference for the pocket occupied by PCW . Thus , all four sites were used to search for the off-target binding sites of CETP inhibitors . Although only approximately 15% of human proteins have known 3D structures deposited in the Protein Data Bank ( PDB ) [30] , the structural coverage of the human proteome increases to 57% if homologous proteins are included ( e-value less than 1 . 0e-3 and aligned sequence lengths greater than 30 residues using a Blast [31] search ) . The structural coverage is reduced to around 40% if the aligned length is greater than 120 residues ( Fig . S2 ) . After removing structures with redundant sequences ( sequence identity = 100% ) , 5 , 985 structures and models from the PDB were selected for off-target search by SMAP . Besides bactericidal/permeability increasing protein ( PDB Id: 1ewf ) that is classified in the same fold and Pfam [32] family as CETP ( FATCAT [33] p-value = 1 . 26e-11 , RMSD = 4 . 53 ) , 273 off-fold structures are found with similar binding sites to CETP ( SMAP p-value less than 1 . 0e-3 ) . Reverse virtual screening of the 273 structures against JTT-705 , the smallest CETP inhibitor , was carried out with Surflex [27] and eHits [28] ( see Methods ) to detect the binding capability of these proteins . To reduce the impact of protein flexibility , the complex structure , whenever available in PDB , is used for docking . Proteins that have steric crashes with JTT-705 were removed from the list and a panel of CETP off-targets consisting of 204 structures was constructed for further study as shown in Table S1 . The majority of these off-targets have binding sites that match to one of the two sites that are adjacent to PCW in CETP . Excluding cytochrome P450s that bind drugs promiscuously , most of the putative off-targets are involved in lipid/fatty acid transport or binding , signal transduction pathways and immune response . Based on both SMAP p-values and docking scores ( p-value<1 . 0e-3 , Surflex score>3 . 50 and eHiTs score<−4 . 50 ) , six classes of structure were consistently found at the top of the list: CD1B like antigen recognition domains ( CD1B ) ; nuclear hormone receptor ligand binding domains ( NR ) ; lipid transport proteins ( LPTP ) ; fatty acid binding proteins ( FABP ) ; EF hand-like calcium binding proteins ( EF ) ; and heme binding proteins ( HEME ) . The first four classes of proteins are able to bind cognate ligands similar to those that bind to CETP , such as fatty acids , lipoproteins , and lipids [34] . Although these putative off-targets do not have detectable global structural similarities to CETP according to their CE Z-scores ( Fig . S3 ) , they have local structural similarity and are related to each other , forming an interconnected off-target network . As shown in Fig . 1 , 76% of the putative off-targets ( 154/204 ) form the three largest clusters . The largest helix bundle cluster includes NR , EF , HEME and other proteins ( Fig . S4 ) . In this paper , we focus on the six selected classes of proteins and demonstrate how they correlate to the clinical findings . Other putative off-targets are subject to on-going computational and experimental studies . Most of the predicted ligand binding sites of CD1B , LPTP , and FABP have a similar topology to that of CETP . The drug molecule binds to a cavity formed by anti-parallel beta-sheets and capped by other structural components such as a helix . The others , NR , EF , and HEME all have alpha-helical architectures that are completely different from the secondary structure surrounding the binding site of CETP . These differences illustrate the necessity of tools like SMAP that can find local structural similarities even when global similarity is non-existent . From a functional perspective , it is not surprising that lipid binding proteins act as off-targets for CETP inhibitors since they are required to bind similar cognate hydrophobic ligands such as PCW . It is noteworthy that glycolipid transfer protein , one of the lipid binding proteins , has significant structural similarity to nuclear hormone receptors . For example , the FATCAT [33] p-value is 1 . 77e-3 when comparing one glycolipid transfer protein ( PDB id: 1TFJ ) with that of retinoid X receptor ( PDB id: 1YOW ) , but the RMSD is 9 . 82 Å for a rigid superimposition . However , if the components of these structures are allowed to twist , the RMSD drops to 2 . 57 Å when the helices surrounding the binding site are well aligned ( Fig . S5 ) . The structural similarity between glycolipid transfer protein and other all-helical proteins increases confidence in our result that the lipid-activated nuclear receptor ( NR ) is one of the major off-targets of CETP inhibitors . We searched for possible functional correlations between CETP and the putative off-targets using the iHOP [35] literature network ( http://www . ihop-net . org/UniPub/iHOP/in ? dbrefs_1=NCBI_LOCUSLINK__ID|1071 ) . Several top-ranked off-targets appear in the same sentences with each other more than 3 times in the literature . They include phospholipid transfer proteins , nuclear receptors , including PPAR , major histocompatibility complex class II that is similar to CD1B , apolipoprotein A-1 , and angiotension I converting enzyme . The functional similarity between CETP and the off-targets is further quantitatively measured using gene ontology ( GO ) relationships found with the FunSimMat web server [36] ( http://funsimmat . bioinf . mpi-inf . mpg . de/index . php ) . From 204 off-targets , 148 structures had annotated GO terms and 94 structures had detectable similarities with a Resnik score [37] larger than 0 . 0 . Among these 94 structures , lipid transport/binding proteins , CD1B , and nuclear hormone receptors were ranked top , followed by globin-like , EF hand-like and other proteins ( Table S2 ) . To further support our off-target predictions we conducted docking studies on CETP and the identified off-targets , which also provides insights into the molecular mechanisms of off-target binding . It has been established that the binding affinity calculated from docking programs is not necessarily reliable [38]–[40] . When using an energy-based scoring function , the errors come predominantly from the inaccurate parameterization of the individual energy terms . We find that the docking scores for CETP and its putative off-targets are linearly dependent on the number of carbon atoms on the docked molecules because the hydrophobic term dominates the scoring ( Fig . S6 ) . Based on this observation we developed a procedure to minimize the systematic error in the scoring function . Rather than considering the raw docking score we used the z-score to represent the relative binding affinity . The z-score is derived from a large number of random drug-like molecules and is dependent on both the number of carbon atoms in the ligand and the nature of the protein binding site . A large negative z-score indicates a high probability of true binding . Based on this procedure , the normalized docking scores ( NDS ) of the six classes of off-targets are listed in Table 1 . These data indicate that binding of CETP inhibitors to putative off-targets is indeed statistically significant . Furthermore , the vector distance of the carbon atom size dependent average docking score for CETP and the majority of off-targets is less than 1 . 0 ( Table S3 ) . This implies that the ligands are able to bind to CETP and to the off-targets with similar binding affinities , since their predicted binding affinity differences are less than 1 . 0 , which is the standard deviation of docking scores ( see Methods ) . Finally , the correlation of ligand binding profiles between CETP and its off-targets [4] are relatively high ( Table S3 and Fig . S7 ) . Importantly , the binding profiles for the three CETP inhibitors ( Torcetrapib , Anacetrapib , and JTT-705 ) are different from each other across the panel of off-targets . JTT-705 is the most promiscuous inhibitor . In contrast , Torcetrapib failed to dock into some of the off-targets , and Anacetrapib is suitable to be docked into the least number of off-targets . The difference between their off-target binding profiles can be partly explained by their different complexity [41] and sizes . The molecular volumes of JTT-705 , Torcetrapib and Anacetrapib are 407 . 31 , 498 . 42 , and 527 . 28 Å3 , respectively . As shown in Table 1 the estimated volume of the off-target binding pockets varies greatly . Thus , the smallest ligand , JTT-705 , can be accommodated in all of these pockets , but the larger-sized Torcetrapib and Anacetrapib are difficult to fit into the smaller sized pockets . It could be argued that the failure in docking Torcetrapib and Anacetrapib into the smaller sized pockets is because the induce fit of the receptor is not explicitly modeled . However , for most of the NRs , both antagonist and agonist conformations are tested . Thus it is less likely that the unfitness of Torcetrapib and Anacetrapib for some of the off-targets is a result of not specifically considering induced fit in the docking calculation . The different off-target binding profiles of these CETP inhibitors have significant implications for the observed side-effects , as discussed subsequently . By incorporating the predicted off-targets into biological pathways it is possible for us to correlate the predicted off-target interactions with the observed pleotropic effects of Torcetrapib , Anacetrapib and JTT-705 . Among them , the negative effect of Torcetrapib on blood pressure in phase III clinical trials could be deduced . Also deducible was an explanation for the increased death from infection and cancer [21] . Conversely , JTT-705 has gotten encouraging safety results from phase II clinical trials and no side-effects of hypertension have been observed thus far . Similar positive results are observed for Anacetrapib during phase I clinical trials . It should be noted that at this time that JTT-705 and Anacetrapib are in clinical trials involving only a small number of patients during short term studies . Results from long term studies are needed to confirm the absence of negative effects for these two drugs . In addition , JTT-705 is found to be able to block cell proliferation and angiogenesis through Ras and P38 kinase pathways [26] . To illustrate these findings , using a survey of the literature , we constructed a hierarchical biological network that connects drugs , off-targets , pathways and clinical observations . Using this network we could explore the implications of administering CETP inhibitors on different pathways through their interactions with corresponding off-targets ( Fig . S8 ) . The network consists of several interconnected metabolic , signal transduction , and gene regulation pathways . Each component of the network is separately shown in Fig . 2 , Fig . 3 , Fig . 4 , and Fig . S9 , and is discussed in detail in the following sections . It is notable that several predicted off-targets , especially the nuclear hormone receptors , are essential components in the network , involved in both positive and negative controls of several cellular systems . Nuclear hormone receptors are known as lipid-activated transcription factors that play key roles in lipid metabolism , inflammatory processes and the hormone system . The regulatory controls of our predicted nuclear hormone receptors are on pathways involved in hypertension , inflammation and cancer development . Torcetrapib , Anacetrapib and JTT705 showed different binding affinities to these receptors and thus different clinical outcomes resulting from the combinational responses of these receptors in related pathways .
In vitro , in vivo and clinical studies indicate that CETP inhibitors exhibit pleotropic effects in humans through the interaction with unknown off-targets . We have identified a panel of proteins that likely bind to CETP inhibitors leading to the observed clinical indications . The putative off-target interactions are consistent with existing experimental data and provide insights into the molecular mechanisms of the side-effect profile of CETP inhibitors . Drug promiscuity depends not only on the similarity of ligand binding pockets in the related proteins but also the complexity of the drug itself [41] . In general , smaller molecules are able to bind more targets . The same trend has been predicted for CETP inhibitors; the smallest JTT-705 is the most promiscuous and the largest , Anacetrapib , is the least promiscuous . However , in contrast to conventional wisdom that implies the more specific the binding the lesser the side-effects , the most promiscuous inhibitor , JTT-705 , does not cause the side-effect of hypertension that is observed in the more specific Torcetrapib . Considering the regulation of blood pressure by NRs , it is possible that JTT-705 acts as an antagonist of NRs to down-regulate aldosterone . However , our results suggest that CETP inhibitors prefer binding to the agonist rather than the antagonist conformation of the NR . Experimental evidence also implies that JTT-705 actually activates NR to mediate Ras and p38 kinase pathways [26] . Thus , it is more likely that the side-effect of CETP inhibitors is modulated by a combination of biological controls involved in many physiological processes such as cell proliferation [81] , inflammation and hypertension . In other words , JTT-705 is involved in activation of NRs that contribute to both positive and negative controls of aldosterone . Although Torcetrapib is more specific and binds less off-targets than JTT-705 , it only activates those NRs that up-regulate RAAS resulting in hypertension . To fully understand how small molecules can modulate physiological or pathological processes through such combinatorial control , it is necessary to simulate the dynamic properties of the biological system . To this end , it is a critical first step to identify all of the putative molecular receptors involved in the biological process and to connect them into a logical integrated protein-ligand interaction network . The chemical systems biology approach developed here is limited by available protein structures that currently only cover approximately 50% of the human proteome , although the structural coverage of the human proteome will steadily increase with progress in structural genomics [82] and conventional structure determination . As a result , some potential off-targets may be missed because they are not included in the screening . In addition to establishing functional relationships between proteins using their sequences , structures and functional sites , there are significant efforts to relate drug targets to their ligands through chemical genomics analysis [12] . However , the chemical genomics approach is restricted by the availability of bioactivity data . When exploring off-targets that cover the whole human proteome , this limitation becomes obvious since only a small number of target families explored by pharmaceutical companies are in the bioactivity database [5] . Thus our method is complementary to existing chemical genomics approaches . Drug-target networks will be greatly expanded by combining chemical genomics data and a structural genome-wide off-target analysis . Several studies have attempted to extend the target-based method to the domain-based model through similar sequence motifs or global structures [83] . In this study we further expand the scope of the chemical genomics approach beyond sequence and fold similarity by searching for similar ligand binding sites . Hence a ligand binding site-based approach will provide an ever improving way to generate a candidate list of proteins participating in interconnected biochemical pathways and to establish their relationships to biological processes . It is hoped that these approaches will eventually provide the foundation for the in silico simulation of the influence of small molecules on biological systems . In the interim it is noted that the analysis of incomplete networks is still invaluable in making new discoveries in biomedicine as exemplified by several recent studies [3] , [11] . Besides SMAP used in this study , a number of web servers for ligand binding site search are available , for example , SiteEngine [84] , SitesBase [85] , [86] , CavBase [87]–[89] , SuMo [90] , PdbSiteScan [91] , eF-Site [92] , [93] , pvSOAR [94] , and pevoSOAR [95] . Compared with these servers , SMAP has several distinguishing features making it particularly suitable for identifying off-targets on a structural genome-wide scale . First , SMAP does not require prior knowledge of both the location and the boundary of the ligand binding site . Instead , whole proteins are scanned to find the most similar local patch in the spirit of local sequence alignment such as the Smith-Waterman algorithm [96] . This feature makes SMAP appropriate for practical problems since typically the boundary of the ligand binding site is not clearly defined or depends on the ligand in the complex structure . Second , SMAP integrates geometric , evolutionary and physical information into a unified similarity score akin to a sequence alignment score . However , unlike conventional sequence alignment , the SMAP alignment is sequence order independent; a necessary requirement when comparing local binding sites . Third , because SMAP uses the reduced structure representation , it is not sensitive to structural uncertainty and flexibility . Thus SMAP can be applied to homology models and handle flexible ligand binding sites . Finally , we have developed a probability model to efficiently estimate the statistical significance of the binding site similarity . The model allows us to reliably identify similar ligand binding sites in a high throughput fashion . Despite these advantages of SMAP , it is expected that the best results will come from the combination of different tools as demonstrated by many studies in bioinformatics and molecular modeling . Despite the success of ligand binding search algorithms in protein function prediction and drug design [2] , [4] , [20] , [87] , [88] , [95] , [97]–[99] currently no algorithm can retrieve all of the binding sites that bind a cognate ligand such as ATP . However , in the context of searching for off-targets of drug molecules , the actual number of false negatives may be limited based on the nature of the drug . False negatives in the ligand binding site search are due mainly to large conformational changes of the ligand and corresponding physical and geometric changes in the binding site . Most existing drugs are designed to selectively inhibit an exquisite target . They are more rigid and less adaptable to the changing environment of the binding site than the cognate ligand . For example , a protein kinase ATP competitive inhibitor is designed to inhibit only the ATP binding site of the protein kinase , not that of other superfamilies such as P-loop hydrolases . On the other hand , although rational drug design may take the same cognate ligand binding site into account , it rarely explores the cross-reactivity between binding sites that are not naturally designed for the same cognate ligand but are able to bind the same drug . Studies by others have shown that the drug binding site can be considered as a negative image of the drug to screen compound database [100] or vice versa to model the drug binding site [101] . Hence ligand binding site similarity search is a valuable tool to identify off-targets that accommodates only the drug molecule but not necessarily all proteins that bind to the same cognate ligand across gene families . In general , the chemical systems biology approach developed in this paper is specific in identifying potential off-targets for drug-like molecules and could be used in concert with experimental design employing in vitro screening , in vivo screening and clinical trials . Even with the current limited structural coverage of the human proteome , our predications are able to provide a testable hypothesis as to the suitability of a lead compound prior to conducting a clinical trial . Thus our findings have implications for drug discovery and development . In contrast to the conventional drug discovery process in which drug leads are optimized to reduce promiscuous binding , the possible combinatorial control of aldosterone regulation by CETP inhibitors suggests that adverse drug effects can be minimized through fine tuning of multiple off-target interactions . Although it is desirable for a drug to bind the primary target in a highly specific way , this is difficult to achieve considering the inherent similarity among protein binding pockets within and across gene families . Moreover , many biological process involve combinatorial control to provide redundancy and homeostasis [102] . In such cases it becomes very difficult to modulate the systems behavior by inhibiting or activating only one single target protein . Thus , a multiple-target approach [6] and combination therapy [10] have been actively pursued to boost clinical efficacy in the treatment of diseases such as cancer and diabetes . However , these combined approaches are rarely systematic with the purposeful intent of developing therapeutics that bind to a primary target to treat the disease , but at the same time are considered to bind to desirable off-targets that modulate side-effects . In some cases this combined goal is achieved serendipitously as would seem to be the case for JTT-705 . Instead of using a single molecule , it may be more feasible to use multiple components to treat a disease state and at the same time to reduce drug side-effects . Different from conventional combination therapy where all of components target disease related proteins , here only a subset of the molecules are directly therapeutic , other molecules serve the purpose of reducing side-effects by targeting non-disease related proteins . We speculate that many drugs which failed due to off-target effects can be rescued by this target-off-target combination therapy . For example , it is expected that the side-effect of Torcetrapib can be reduced by introducing molecules that binds to molecular components involved in the negative control of aldosterone regulation . Such therapies can be only rationally designed by exploring the system properties of the biological network .
5 , 985 structures or models that cover approximately 57% of the human proteome were searched against CETP ( PDB id: 2obd ) ligand binding sites using the sequence order independent profile-profile alignment ( SOIPPA ) algorithm [20] . A new statistical model was introduced to the original approach to estimate the significance of the alignment score [103] . In brief , the alignment score for a given alignment length is fitted to an extreme value distribution ( EVD ) : ( 1 ) Where: ( 2 ) where S is the raw SOIPPA similarity score . μ and σ are fitted to the logarithm of N , which is the alignment length between two proteins: ( 3 ) ( 4 ) Six parameters a , b , c , d , e , and f are 5 . 963 , −15 . 523 , 21 . 690 , 3 . 122 , −9 . 449 , and 18 . 252 for the McLachlan similarity matrix used in this study , respectively . Using this statistical model , 276 off-targets are identified with p-values less than 1 . 0e-3 . The putative 276 off-targets are subject to further investigation using more computationally intensive protein-ligand docking . After removing three structures with the same fold as CETP , JTT-705 , the smallest CETP inhibitor , is docked to the remaining 273 structures using two commonly used fast docking programs , Surflex 2 . 1 [27] ( default setting ) and eHits 6 . 2 [28] ( fastest setting ) . 69 structures with a Surflex docking score smaller than 0 . 0 or an eHits score larger than 0 . 0 are considered to be difficult to fit JTT-705 due to significant steric crashes ( and hence the other two inhibitors based on size ) and are removed from the putative off-target list . The remaining 204 structures are subject to further investigation using the docking software AutoDock4 . 0 [29] and other more computationally intense methods as described below . An all-against-all global structural similarity analysis between the 204 putative off-targets was computed using CE [104] . A graph is constructed with each of the structures as a node . An edge is formed between two nodes if their CE z-score is larger than 4 . 0 ( a superfamily level similarity ) [104] . The volume of the binding pocket is computed using the CASTp server [105] ( http://sts-fw . bioengr . uic . edu/castp ) with default settings . Drug-like molecules are downloaded from ZINC ( http://zinc . docking . org ) [106] . From this database , six sets of molecules are randomly selected with a fixed number , 5 , 10 , 15 , 20 , 25 and 29 carbon atoms , respectively; each set includes 100 molecules . These molecules are docked to CETP and its putative off-targets using eHiTs [28] and AutoDock4 . 0 [29] . The correlation of the docking score to the number of carbon atoms is derived from linear regression for each of the protein receptors . From the linear fitting curve , the average docking score for molecules with a certain number of carbon atoms can be estimated . Based on the fitted average docking score , a normalized docking score DS is calculated as a z-score: ( 5 ) Where Si is the raw docking score for the molecule with i carbon atoms , μi is the fitted average docking score for the number of carbon atoms i , σ is the standard deviation , which is not dependent on the size of molecules and is approximately 1 . 0 in all cases . The vector distance of the average docking score D between CETP and its off-targets is calculated from the average values of the docking scores for randomly selected molecules with fixed numbers of 5 , 10 , 15 , 20 , 25 and 29 carbon atoms as follows: ( 6 ) where SCETP and Soff are the average values of carbon atom size dependent docking scores to CETP and its off-targets , respectively . In this case study , we identify a panel of off-targets of CETP inhibitors using a chemical systems biology approach . All of the identified off-targets belong to different protein superfamilies from the primary target , but are structurally and functionally related , being mainly involved in lipid metabolism , immune response and signaling networks . Among them , CD1 , nuclear hormone receptors and lipid transport proteins are the most likely off-targets with highly consistent results from multiple resources including functional correlation , ligand binding site similarity , hydrophobic scales , and predicted binding affinities . Moreover , the elucidated off-target effects from these proteins are strongly correlated to clinical and in vitro observations . Their combinatorial control of biological process plays a key role in the modulation of the adverse drug effect of CETP inhibitors . This study demonstrates that a chemical systems biology approach , which systematically explores protein-ligand interactions on a genome-wide scale and incorporates them into biological pathways , will provide us with valuable clues as to the molecular basis of cellular function . At the same time , it will help to transform the conventional single-target-single-drug drug discovery process to a new multi-target-multi-molecule paradigm .
|
Both the cost to launch a new drug and the attrition rate during the late stage of the drug discovery and development process are increasing . Torcetrapib is a case in point , having been withdrawn from phase III clinical trials after 15 years of development and an estimated cost of US $800 M . Torcetrapib represents a new class of therapies for the treatment of cardiovascular disease; however , clinical studies indicated that Torcetrapib has deadly side-effects as a result of hypertension . To understand the origins of these adverse drug reactions from Torcetrapib and other related drugs undergoing clinical trials , we introduce a systematic strategy to identify off-targets in the human structural proteome and investigate the roles of these off-targets in impacting human physiology and pathology using biochemical pathway analysis . Our findings suggest that potential side-effects of a new drug can be identified at an early stage of the development cycle and be minimized by fine-tuning multiple off-target interactions . The hope is that this can reduce both the cost of drug development and the mortality rates during clinical trials .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"computational",
"biology/literature",
"analysis",
"computational",
"biology/macromolecular",
"structure",
"analysis",
"biotechnology/protein",
"chemistry",
"and",
"proteomics",
"biotechnology/small",
"molecule",
"chemistry",
"cardiovascular",
"disorders/hypertension",
"pharmacology/adverse",
"reactions",
"computational",
"biology/systems",
"biology"
] |
2009
|
Drug Discovery Using Chemical Systems Biology: Identification of the Protein-Ligand Binding Network To Explain the Side Effects of CETP Inhibitors
|
Eukaryotic cells form stress granules under a variety of stresses , however the signaling pathways regulating their formation remain largely unknown . We have determined that the Saccharomyces cerevisiae lysine acetyltransferase complex NuA4 is required for stress granule formation upon glucose deprivation but not heat stress . Further , the Tip60 complex , the human homolog of the NuA4 complex , is required for stress granule formation in cancer cell lines . Surprisingly , the impact of NuA4 on glucose-deprived stress granule formation is partially mediated through regulation of acetyl-CoA levels , which are elevated in NuA4 mutants . While elevated acetyl-CoA levels suppress the formation of glucose-deprived stress granules , decreased acetyl-CoA levels enhance stress granule formation upon glucose deprivation . Further our work suggests that NuA4 regulates acetyl-CoA levels through the Acetyl-CoA carboxylase Acc1 . Altogether this work establishes both NuA4 and the metabolite acetyl-CoA as critical signaling pathways regulating the formation of glucose-deprived stress granules .
When eukaryotic cells are exposed to environmental stress such as extreme temperatures , nutrient deprivation , or toxic chemicals , they quickly employ defense mechanisms to promote survival . An essential part of the stress response is the formation of stress granules ( SGs ) , cytoplasmic ribonucleoprotein ( RNP ) granules that contain translationally repressed messenger RNA ( mRNAs ) and numerous other proteins . Not only is accumulation of SGs a hallmark of many neurodegenerative diseases , but mutations in known SG proteins that increase their tendency to aggregate are causative [reviewed in 1] . There is also growing evidence that SGs are contributing to both progression and chemotherapy resistance of cancer cells by promoting survival and inhibiting apoptosis [2–6] . The implication of stress granules as contributors to disease and chemotherapy resistance illustrates the importance of defining how this stress response is regulated within the cell . Stress granule formation is a conserved cellular stress response and SGs are hypothesized to be triage centres where mRNAs are protected in a non-translating state prior to either being sorted to Processing Bodies ( P-bodies ) for degradation or back into active translation . Further mRNAs can also move from P-bodies to SGs , highlighting the dynamic nature of these RNP structures . Though P-bodies and SGs share some of the same protein components [7] , SGs and P-bodies form independently in humans [8 , 9] and under most stress conditions in yeast [10–12] . However , there is evidence that P-bodies promote formation of glucose deprivation SGs ( GD-SGs ) in yeast [13] . Although composition of SGs varies with stress , the core constituents of SGs are largely conserved and include poly ( A ) +mRNA , 40S ribosomal subunit , translation initiation factors , PolyA-binding protein Pab1 , along with a number of proteins implicated in SG assembly such as Pub1/TIA-1 , and Pbp1/ATAXIN-2 ( yeast/MAMMALIAN names ) [reviewed in 14] . These dynamic mRNPs also sequester proteins from key cell signaling pathways including pro-apoptosis factors and other signaling proteins which may protect the cell from initiating apoptosis during modest stress [15] . As SG-inducing stresses all cause inhibition of translation initiation and as drugs that trap mRNA in polysomes inhibit SG formation [13 , 16 , 17] , it is generally believed that blocking translation initiation is necessary for SG assembly . However , not all stresses that inhibit translation initiation cause SG formation [13 , 18] , indicating that inhibition of translation is not the sole “trigger” of SG assembly but requires other signaling pathways , which like the composition of SGs are proposed to be stress-dependent . As SGs are transient structures that can assembly and dissolve within minutes even when translation initiation is inhibited , it is strongly argued that SG dynamics is regulated largely through post-translational modification ( PTMs ) [reviewed in 1] . In yeast , SGs have been documented to form under a variety of stress conditions , however GD-SG formation is by far the most characterized [13 , 19 , 20] . SGs form rapidly within 5 and 10 minutes after glucose deprivation and their formation are partially dependent on assembly factors Pbp1 and Pub1 [13] . Surprisingly the signaling pathway ( s ) triggering SG assembly upon glucose deprivation is poorly understood . Of the three primary glucose sensing and signaling pathways in S . cerevisiae two , TOR and PKA , do not appear to play a role in regulating GD-SG formation [21 , 22] . It has been suggested that the third pathway , AMPK1/SNF1 may play a signaling role in glucose-deprived stress granule formation potentially through the activation of PAS kinases Psk1 [23] . However it remains to be tested if Snf1-activation of Psk1 is actually driving GD-SG formation . Though the role of lysine acetylation in GD-SG dynamics in yeast has not been assessed , there is growing evidence that lysine acetylation is playing a critical role in various aspects of SG dynamics . In mammalian cells lysine acetyltransferase ( KAT ) CBP and lysine deacetylases ( KDACs ) HDAC6 and SIRT6 have been shown to both co-immunoprecipitate and co-localize with SG proteins [24–26] . In mouse embryonic fibroblasts ( MEF ) cells SIRT6 has been implicated in both the assembly and disassembly of heat-shock SGs [25] , but the role of HDAC6 is not as straight forward . The downregulation of HDAC6 activity in MEF cells inhibited SG formation [26] , while in QBI-293 cells it increased SG formation [24] . In the later study it was determined that acetylation of TDP-43 by CBP drove its aggregation into stress granules , while HDAC6-dependent deacetylation of TDP-43 reduced SG formation . In S . cerevisiae there are also intriguing hints that lysine acetylation contributes to SG dynamics . Deletion mutants of KAT GCN5 and KDACs HST1 and SIR2 were identified to exhibit increased SG formation under non-stress conditions in a high content microscopy screen [10] . Furthermore we recently determined that Esa1 , the catalytic domain of the yeast KAT NuA4 , co-purified SG components Pab1 , Pbp1 , Lsm12 and Pbp4 and that both Pab1 and Pbp1 are in vitro targets of NuA4 [27] . Lysine acetyltransferases maybe exceptional signaling effectors for monitoring glucose and the metabolic state of the cell as their activity is intimately linked to acetyl-CoA , their acetyl donor [28–31] . Acetyl-coA is a central metabolite whose cellular levels are tightly linked to nutrient availability [reviewed in 32 , 33 , 34] . In S . cerevisiae there are two pools of acetyl-CoA , one mitochondrial and one nucleocytosolic , the latter of which affects nuclear and cytosol lysine acetylation events . Under conditions where glucose is in abundance , nucleocytosolic levels of acetyl-CoA and histone acetylation are high . However once cells exhaust their sources of glucose , acetyl-CoA and histone acetylation decrease [28 , 35 , 36] . Even within a cell cycle , oscillations of acetyl-CoA levels have been shown to regulate KAT activity against both histones and other substrates [28] . Further studies in both yeast and human cells show that nucelocytosolic depletion of acetyl-CoA induces autophagy , largely through KAT-dependent transcriptional regulation of autophagy genes [37 , 38] . Given that KATs are naturally designed to monitor nutrient availability and the clear links between lysine acetylation and SG dynamics , we sought to determine if KATs contribute to the regulation of GD-SG formation in yeast . We demonstrate that NuA4 is required for the assembly of glucose-deprived stress granules and identify a conserved role for Tip60 , the mammalian homolog of Esa1 [39] , in the regulation of SGs in human breast cancer cells . Surprisingly , we find that NuA4 is partially regulating the formation of GD-SGs through acetyl-CoA levels . Together our work defines novel signaling pathways that regulates SG dynamics upon glucose deprivation and suggests that NuA4 regulation of acetyl-CoA levels is likely mediated through Acc1 .
Given the interaction between Esa1 and stress granule proteins [27] and the roles of NuA4 in glucose metabolism [40 , 41] , we sought to determine whether NuA4 has a role in SG formation upon glucose deprivation . NuA4 is a multi-subunit complex , composed of the essential catalytic subunit Esa1 [42] , five other essential subunits Act1 , Arp4 , Epl1 , Swc4 , Tra1 , and seven non-essential subunits Eaf1 , Eaf3 , Eaf5 , Eaf6 , Eaf7 , Yaf9 , and Yng2 [39] . Molecular and structural dissection has revealed NuA4 to be modular in nature [43] , and that assembly of its multiple sub-complexes depends on the Eaf1 subunit [44 , 45] . Wild type , mutants of core SG subunits pbp1Δ and pub1Δ , along with NuA4 non-essential mutants eaf1Δ and eaf7Δ were transformed with a Pab1-GFP expressing plasmid [46] . pbp1Δ and pub1Δ cells were used as controls as both mutants display reductions in GD-SGs [13] . The percentage of cells with Pab1-GFP foci and the number of foci/cell ( S1 Table contains all quantification for strains studied ) were quantified in both 2% glucose media and after 10 minutes of glucose deprivation . None of the mutants examined display a significant increase or constitutive formation of SGs under glucose conditions compared to wild type . As expected , upon glucose deprivation Pab1-GFP SGs are induced in wild type cells but the induction was significantly reduced in pub1Δ and pbp1Δ cells [13] ( Fig 1A and 1B ) . Upon glucose deprivation eaf1Δ and eaf7Δ cells show a decrease in Pab1-GFP SGs to a similar extent as pbp1Δ and pub1Δ cells ( Fig 1A and 1B ) . Decreases in GD-SG formation were seen for NuA4 mutants’ eaf3Δ and eaf5Δ ( S2 Table ) . Furthermore , while the temperature-sensitive allele of ESA1 , esa1-L254P ( esa1-ts ) [42] , form GD-SGs at the permissive temperature ( 25°C ) , when pre-incubated at the non-permissive temperature ( 37°C ) prior to 10 minutes of glucose deprivation , there is a significant decrease in Pab1-GFP foci formation ( Fig 1C and 1D ) , indicating that the catalytic activity of NuA4 is required for Pab1-GFP SG formation . To determine if NuA4 is only promoting Pab1 assembly into GD-SGs or whether it had a greater role in SG or P-body formation we examine the localization of endogenously tagged core stress granule proteins Pbp1-GFP , Pub1-GFP , and Pab1-GFP [13 , 20] along with the P-body marker Lsm1-GFP [47] in wild type , eaf1Δ and eaf7Δ strains after 10 minutes of glucose deprivation . Similar to plasmid based Pab1-GFP experiment ( Fig 1 ) , NuA4 mutants display a decrease in endogenously tagged Pab1-GFP localization to GD-SGs ( Fig 2A & 2B ) . In addition similar defects were observed for Pub1-GFP and Pbp1-GFP , indicating that NuA4 is impacting GD-SG assembly , and not just Pab1-GFP localization to GD-SGs . Though decreased levels of Pab1-GFP and Pub1-GFP were detected in eaf1Δ cells , deletion of EAF7 had no effect on protein levels ( S1 Fig ) , suggesting that defects in glucose-deprived SG formation is not due to changes of the SG marker levels . In contrast , NuA4 mutants did not impact P-body marker Lsm1-GFP foci formation . While NuA4 is required for glucose-deprived SG formation it is not required for the formation of heat-shock or ethanol SGs ( S2 Fig ) . Taken together , this work has identified NuA4 as a novel signaling pathway for formation of SGs upon glucose deprivation . To determine if other KATs or KDACs in S . cerevisiae play a role in GD-SG formation we systematically screened a library of single non-essential KAT and KDAC mutants ( S2 Table ) . The KAT/KDAC mutant library was transformed with a Pab1-GFP expressing plasmid and screened for SGs under both glucose replete and depleted conditions . In addition to NuA4 mutants ( eaf1Δ , eaf3Δ , eaf5Δ and eaf7Δ ) , deletion mutants of the KAT GCN5 also displayed a modest decrease in GD-SG formation . Therefore , we sought to determine if Gcn5 does impact GD-SGs and/or if it has a functionally redundant role with NuA4 in GD-SG dynamics . Glucose-deprived stress granule formation was measured in wild type , eaf7Δ , gcn5Δ , and eaf7Δgcn5Δ cells expressing Pab1-GFP from its endogenous loci in both glucose and glucose deprivation conditions ( 10 minutes , 30 minutes and 60 minutes ) . eaf7Δ was selected as a proxy for NuA4 as it displays defects in GD-SG , but unlike eaf1Δ , eaf7Δ cells displays minimal fitness defects [48] . Furthermore , while eaf1Δgcn5Δ cells are inviable [45 , 49 , 50] , eaf7Δgcn5Δ cells display only minor growth defects ( S2C Fig ) . While gcn5Δ cells did not display a significant reduction in GD-SG assembly compared to wild type cells using the endogenously integrated Pab1-GFP , eaf7Δgcn5Δ cells display a significant reduction in GD-SG formation at 10 minutes compared to both wild type and single KAT mutants ( Fig 3A and 3B ) . To better understand the contribution of Eaf7 and Gcn5 to kinetics of SG formation/disassembly , SGs were measured over 10 , 30 and 60 minutes glucose deprivation ( Fig 3C ) . As previously shown in wild type cells GD-SGs quickly form and then start resolving by 30 minutes as the cells begin to adapt to glucose deprivation [11] . While gcn5Δ cells had no impact on GD-SGs over the time course , eaf7Δ cells display a delay in the formation of GD-SGs . In contrast , the percentage of cells with SGs do not increase over the glucose deprivation time course in eaf7Δgcn5Δ cells . As eaf7Δgcn5Δ cells display minimal growth defects compared to the single mutants , it is unlikely the defects in GD-SG formation reflect cell sickness . Rather these results suggest that EAF7 and GCN5 are functionally redundant in GD-SG regulation and in the absence of Eaf7 , and presumably NuA4 , Gcn5 is able to compensate albeit with slower kinetics for GD-SG formation . Given the conservation of KAT complexes and stress granule biology , we predicted that the Tip60 complex , the mammalian homolog of the NuA4 complex , might also contribute to SG dynamics . To test this possibility we examined SGs in HeLa and MCF7 cell lines in the presence and absence of the inhibitor NU9056 which inhibits the activity of the Tip60 protein , the catalytic subunit of the complex and the mammalian homolog of S . cereviaise Esa1 [51] . While cells treated with 20 μM NU9056 alone for 24 hours resulted in a decrease in acetylated histone H4 ( Fig 4A ) it did not promote SG formation in unstressed mammalian cells as assessed using the SG marker TIA-1 [17] ( Fig 4B and 4C ) . We next asked whether inhibition of Tip60 decreases SG formation promoted by sodium arsenite or bortezomib , a chemotherapeutic drug that has been shown to induce SG formation in bortezomib-resistant cancer cell lines [4] . HeLa and MCF7 cells were treated with NU9056 for 24 hours and then either 10 μM bortezomib for 4 hours or 100 μM sodium arsenite for 1 hour . Stress granule formation was analyzed by immunofluorescence using the SG markers TIA-1 , FMRP [52] , and DDX3 [53] . Treatment with NU9056 resulted in a significant decrease in bortezomib-induced ( Fig 4D & 4E ) and sodium arsenite-induced ( Fig 4F & 4G ) SGs in both cell types . These results show that NuA4/Tip60 have a conserved function in regulating SG dynamics . We next sought to determine if the signaling pathway through which NuA4 regulates glucose-deprived SG formation was through either regulation of translation initiation or SNF1/AMPK pathway . Though NuA4 has not been implicated in translation initiation , it does have a role in the transcriptional regulation of ribosome genes [54] and both genetic and physical interactions with ribosome subunits [27 , 45 , 49] . If NuA4 mutant defects in GD-SG formation were due defects in inhibition of translation upon glucose deprivation we would anticipate that in NuA4 mutants polysomes would remain associated with RNA upon glucose deprivation . However , this is not the case as ribosomal profiling determined that like wild type cells , eaf1Δ and eaf7Δ cells are capable of stalling translation initiation by polysome disassembly in response to glucose deprivation ( Fig 5A ) . Another candidate pathway through which NuA4 might regulate GD-SG formation is the SNF1/AMPK pathway , a major sensor of glucose state that is regulated by NuA4 acetylation of Sip2 [41] . Sip2 is one of three β-regulatory subunits of the Snf1 complex and a repressor of Snf1/AMPK activity . NuA4 acetylation of Sip2 enhances its interaction and inhibition of Snf1 . In contrast , in NuA4 mutants Sip2 acetylation decreases , destabilizing its interaction with Snf1 , leading to increased Snf1 activity . If NuA4 mutant defects in GD-SG formation were due to hyperactive Snf1 , we would anticipate that other mutants with hyperactive Snf1 would display defects in SG-formation . However , hxk2Δ cells which have hyperactive Snf1 [55] do not have defects in GD-SG formation ( Fig 5B and 5C ) . Surprisingly , we also determined that in cells in which SNF1 is deleted , GD-SGs still occurs ( Fig 5B and 5C ) indicating that Snf1 kinase activity is not required for the assembly of Pab1-GFP into glucose-deprived SGs . Together this shows that NuA4 is contributing to GD-SG dynamics through a novel mechanism . As cellular levels of acetyl-CoA reflect glucose availability , we next sought to determine if changes in acetyl-CoA impact glucose-deprived SG formation . As exogenous acetate treatment of S . cerevisiae is converted to acetyl-CoA ( Fig 6A ) , we asked if glucose-deprived SG formation in wild type cells was impacted by exogenous acetate . As previously shown [28] , treatment of exponential-phase wild type cells with 100 mM acetate results in increased acetyl-CoA levels ( Fig 6B ) . We found that acetate treatment significantly reduced the number of cells with GD-SGs ( Fig 6C and S3A Fig ) . As the repression by acetate might reflect conversion of acetate to glucose by the glyoxylate cycle we asked if acetate could still suppress GD-SGs in icl1Δ cells , a mutant that disrupts the glyoxylate cycle pathway [56] ( Fig 6C and S3A Fig ) . Acetate could suppress GD-SGs in icl1Δ cells , indicating the suppression is not due to conversion of acetate to glucose but rather through another metabolite or mechanism . Further , galactose and metabolic intermediates citrate and pyruvate cannot suppress GD-SG formation suggesting the effect of acetate is not simply due to reintroduction of an energy source ( S4A and S4B Fig ) . Interestingly 2% ethanol can partially suppress GD-SGs ( S4A and S4B Fig ) which might reflect conversion of ethanol to acetyl-CoA [57] , though higher levels of ethanol induce SGs [58] . Similar to NuA4 mutants , acetate treatment did not suppress heat-shock ( Fig 6D and S3B Fig ) or ethanol SGs ( S2D and S2E Fig ) and acetate treatment did not impact P-body ( Lsm1-GFP foci ) formation ( Fig 6E and S3C Fig ) . As increased acetyl-CoA results in increased protein acetylation [29 , 30] , we asked whether acetate suppression of GD-SGs requires Eaf7/NuA4 or Gcn5 . We found that acetate treatment could further reduce the number of cells with GD-SGs in both eaf7Δ and eaf7Δgcn5Δ ( Fig 6F and S3D Fig ) . Together this work indicates that exogenous acetate treatment can suppress GD-SG formation and this can occur through a mechanism independent of Eaf7/NuA4 or Gcn5 activity . To confirm that the impact of exogenous acetate on SG formation is through the metabolite acetyl-CoA , we next sought to determine if genetic manipulation of cellular acetyl-CoA levels could impact GD-SG formation . Reducing the activity of the essential acetyl-CoA carboxylase Acc1 results in an increase in cellular acetyl-CoA [29] . A doxycycline regulated tet07 repressible promoter was used to control the expression of the essential protein ACC1 ( tet07-ACC1 ) [59] . Exponentially grown tet07-ACC1 cells transformed with a Pab1-GFP expressing plasmid were treated with either vehicle control or 10uM doxycycline for a 2 . 5 hour incubation period prior to glucose deprivation . As previously shown , doxycycline treatment results in reduced Acc1 protein levels , increased cellular acetyl-CoA levels , and increased histone acetylation ( Fig 7A–7C ) . While , doxycycline treatment did not induce SGs in glucose replete conditions , reducing Acc1 activity suppressed glucose-deprived SG formation ( Fig 7D and S5A Fig ) . Thus , reducing the expression of ACC1 , correlates with increasing acetyl-CoA levels and a reduction in glucose-deprived SG formation . If acetyl-CoA is regulating glucose-deprived SG formation , we predicted that decreasing cellular acetyl-CoA would increase SG formation . Saccharomyces cerevisiae have two acetyl-CoA synthetases , Acs1 and Acs2 , which produce cellular acetyl-CoA from acetate . While ASC2 is essential under glucose culture , the non-essential ACS1 is induced upon glucose depletion and is required for survival on non-fermentable carbon sources [30 , 60–62] . While we could not detect changes in GD-SG formation in either a strain carrying a temperature sensitive allele of ACS2 [30] at 33°C or tet07-ACS2 strain treated with 10uM doxycycline ( S6 Fig ) , this was not the case with ACS1 . Exponentially grown acs1Δ cells have a significant decrease in acetyl-CoA levels and a decrease in histone H4 acetylation as compared to wild type cells ( Fig 8A and 8B ) . Deletion of ACS1 results in a significant increase in glucose-deprived SG formation ( Fig 8C and S5B Fig ) . Interestingly , the increase in GD-SGs in acs1Δ cells was reduced to wild type GD-SG levels by deletion of EAF7 . As the acs1Δeaf7Δ cells did not display phenotypes similar to eaf7Δ cells , it suggests that Acs1 and Eaf7 may contribute to SG dynamics through distinct pathways . Alternatively , deletion of EAF7 may result in an increase in acetyl-CoA levels that could compensate for the reduction of acetyl-CoA seen in the absence of Acs1 . To explore this hypothesis we measured acetyl-CoA levels in exponentially grown eaf7Δ and gcn5Δ cells and found that both have a significant increase in acetyl-CoA levels compared to wild type cells ( Fig 8D ) . In addition , acetyl-CoA level are further increased in the eaf7Δgcn5Δ mutant , which parallels the increased suppression of GD-SGs seen in this mutant ( Fig 3B ) . We next asked if the impact of Eaf7 on acetyl-CoA levels is in parallel or epistatic to Acc1 ( Fig 8E ) . As previously shown tet07-ACC1 cells treated with doxycycline display an increase in acetyl-CoA , however acetyl-CoA levels do not significantly increase upon the deletion of EAF7 ( compare eaf7Δtet07-ACC1 vs tet07-ACC1 treated with doxycycline ) . This suggests that Eaf7 and Acc1 are epistatic , not additive , and work within the same pathway . If the mechanism through which eaf7Δ cells suppress GD-SGs was only through Acc1 ( or downstream pathways ) and increased acetyl-CoA levels , one would predict that doxycycline treated tet07-ACC1 cells should display similar reductions in GD-SGs as eaf7Δ cells ( untreated eaf7Δtet07-ACC1 cells ) , as both have similar levels of acetyl-CoA . Instead we found that GD-SGs in doxycycline treated tet07-ACC1 cells is significantly higher than eaf7Δ cells ( untreated eaf7Δ tet07-ACC1 cells ) ( Fig 8F ) . Nor did we detect a significant change in the percentage of cells with GD-SGs between eaf7Δ ( untreated eaf7Δtet07-ACC1 cells ) and eaf7Δacc1 cells ( doxycycline treated eaf7Δ tet07-ACC1 cells ) . Taken together the non-additive impacts on acetyl-CoA levels suggest that Eaf7 and Acc1 are epistatic and that eaf7Δ mutants are suppressing GD-SGs through increased cellular acetyl-CoA levels and a yet to be described secondary pathway . To test the possibility that NuA4 modulates acetyl-CoA level through regulation of Acc1 , we performed an Acc1 activity assay on whole cell extracts from wild type and the NuA4 deletion mutant eaf1Δ cells . As expected Acc1 activity detected in the whole cell extracts is increased upon the addition of citrate ( Fig 9 ) [63 , 64] . Though protein levels of Acc1 are not impacted by presence or absence of Eaf1 ( S7 Fig ) , we found that Acc1 activity is significantly decreased in eaf1Δ cells ( Fig 9 ) . The reduction in Acc1 activity in eaf1Δ suggests that NuA4 is required for full Acc1 activity and provides insight into the mechanism by which NuA4 mutants display increased acetyl-CoA levels and suppression of GD-SG formation .
Here we demonstrate that NuA4 KAT function is required for SG formation upon glucose deprivation ( Figs 1 and 2 , S2 Fig and S2 Table ) , and this is partially mediated through the regulation of the metabolite acetyl-CoA ( Fig 8 ) . While it could be argued that the impacts of Eaf7/5/3 on GD-SGs could be through their non-NuA4 role as part of the TINTIN complex [reviewed in 65] , this is unlikely as defects in GD-SG formation were also found in the eaf1Δ cells and temperature sensitive esa1-ts allele . Furthermore , as seen for many phenotypes associated with NuA4 mutants [45 , 49 , 66] , GD-SG defects displayed by the scaffold mutant eaf1Δ are more pronounced than those of TINTIN mutants ( Fig 2 ) . We determined that NuA4 mutant eaf7Δ possess a delay in SG assembly rather than a complete inhibition of GD-SG formation ( Fig 3 ) . As eaf7Δ cells display minimal fitness defects , this result suggests that the delay in SG formation is not a reflection of sickness or cell cycle defects . Rather as eaf7Δgcn5Δ double mutant shows a significant reduction in GD-SGs throughout the 60 minute glucose deprivation time course ( Fig 3B ) , it suggests that while NuA4 is the primary KAT regulating GD-SG formation that in its absence Gcn5 can partially compensate . How is NuA4 regulating formation of SGs upon glucose deprivation ? Given the role of NuA4 in histone acetylation and transcription , SG defects of NuA4 mutants could reflect defects in mRNA expression and in turn protein levels of structural SG proteins like Pab1 or Pub1 . Despite the fact that deletion of EAF1 reduces Pab1-GFP and Pub1-GFP protein levels while deletion of EAF7 has no effect ( S1 Fig ) , both eaf1Δ and eaf7Δ cells have significant reduction in GD-SG formation . This suggests that NuA4-dependent regulation of at least Pab1 or Pub1 protein levels have minor contribution to GD-SG dynamics . Indeed , most mutants identified in genome-wide screens with defects in SGs do not impact Pab1 protein levels , suggesting regulation of protein abundance is likely not a common mechanism for SG regulation [10 , 67] . Further , we eliminate two other likely candidate mechanisms through which NuA4 could be regulating GD-SG , inhibition of translation and AMPK1/Snf1 ( Fig 5 ) . The later finding is particularly unexpected given that NuA4 inhibits Snf1 activity [41] . Neither a mutant that had elevated Snf1 activity , hxk2Δ [55] , or deletion of SNF1 itself , impacted GD-SG formation as assessed by the Pab1-GFP SG marker . Though it has been proposed , but not tested , that Snf1-dependent activation of PAS kinase Psk1 is required for the incorporation of Ppb1 into GD-SGs [23] , our results suggest this might not be the case . One possibility is signaling pathways other than Snf1 are activating and regulating Psk1 . Alternatively Snf1 and Psk1 may contribute to other aspects of SG dynamics at a later time point . Indeed , our study assessed SGs at 10 minutes glucose deprivation , while the Psk1 study assessed SGs after 60 minutes , a time point where we see resolution of SGs ( Fig 3 ) . Similarly , the role of AMPK in mammalian SG dynamics is also conflicted . While AMPK1 is required for cold shock induced SG formation in COS7 cells [68] , in HeLa cells increased AMPK1 activity does not induce SGs , but pre-activation of AMPK1 prior to diethyl maleate treatment causes smaller and more SGs [69] . Clearly further studies will be needed to fully assess the role of Snf1/AMPK in SG dynamics . Our work suggests that NuA4 is regulating GD-SG formation through both an acetyl-CoA dependent and acetyl-CoA independent pathways . Surprisingly we found that NuA4 and Gcn5 are not just mediating the downstream effects of acetyl-CoA oscillations , but rather that these KATs also influence acetyl-CoA levels . Genetic and chemical manipulations that increase or decrease acetyl-CoA levels led to the suppression or induction of GD-SGs ( Figs 6–8 ) . Similar to the ability of NuA4 mutants to only suppress GD-SG formation , acetyl-CoA suppression of SG formation is also likely stress specific as exogenous acetate treatment could not suppress heat-shock ( Fig 6D ) or ethanol ( S2D Fig ) induced SG formation . Furthermore our work suggests that acetyl-CoA is not the only driver of GD-SG formation . Reduction of acetyl-CoA levels in acs1Δ cells increase the percentage of cells with SGs upon glucose deprivation , but not under glucose replete conditions ( Fig 8C ) . This suggests that decreases in acetyl-CoA alone is not enough to drive SG formation , but can escalate formation upon glucose deprivation . We found that while deletion of ACS1 elevated the number of cells that formed SGs upon glucose deprivation ( Fig 8C ) , deletion of ACS2 had no influence on GD-SG formation ( S6 Fig ) which was surprising given that Acs2 is the nucleoctyosolic acetyl-CoA synthases essential for production of acetyl-coA on glucose [30] . In contrast , the cellular function of Acs1 is not as clear . While Acs1 has been suggested to be a mitochondrial acetyl-CoA synthase [70] , Acs1 has been localized to the cytoplasm and nucleus [71 , 72] . Further Acs1 is required for acetyl-CoA production on non-fermentable carbon sources , even in the absence of ACS2 [62] , indicating that Acs1 function is not limited to the mitochondria but also contributes to nucleocytosolic acetyl-CoA regulation . As ACS1 is transcriptionally induced upon glucose depletion [60 , 61] , it is possible that Acs1-dependent generated acetyl-CoA upon glucose deprivation is required to attenuate or for the resolution of GD-SGs . While we cannot exclude the possibility that other metabolites derived from acetyl-CoA are responsible for suppression of GD-SGs , the genetics suggest that acetyl-CoA is a regulator of the GD-SG response . How is acetyl-CoA supressing glucose-deprived SGs ? A tempting scenario is that upon glucose deprivation nucleocytosolic acetyl-CoA depletion contributes to SG formation by a yet-to-be-characterized mechanism , and that mutations or treatments that increase nucleocytosolic acetyl-CoA , such as exogenous acetate treatment ( Fig 6 ) , reduction in Acc1 ( Fig 7 ) or deletion of EAF7 ( Fig 8 ) , mask this pathway . Though nucleocytosolic acetyl-CoA levels are a sensitive gauge for the metabolic state of the cell , high during glucose rich growth and low during fasted or carbon-poor states [reviewed in 34] , the extent and timing of acetyl-CoA oscillations upon acute glucose deprivation has yet to be studied . Given that exogenous acetate can be converted to acetyl-CoA and then metabolised within minutes even in yeast cells in which the growth rate is significantly reduced [28] , detailed flux analysis of nucleocytosolic acetyl-CoA upon stresses such as glucose deprivation remains a challenge to measure . However the idea that increased acetyl-CoA can suppress a stress response is not unique . Increased nucleocytosolic acetyl-CoA levels suppress autophagy through histone acetylation and repression of autophagy genes in both human and yeast cells [37 , 38] . Could similar events be occurring to suppress GD-SG formation ? While transcriptional events maybe contributing to SG dynamics , the rapid assembly of GD-SGs and the fact that SG assembly can occur in the absence of translation suggest that the acetylation state of non-histone proteins maybe a more likely mechanism regulating GD-SG assembly . Hence if the speculated model is correct and acetyl-CoA pools rapidly decrease upon glucose deprivation , one potential model is that under glucose replete “normal” acetyl-CoA levels KAT ( s ) are acetylating a target ( s ) that suppress SG formation . Indeed lysine acetylation has been detected on SG subunits Pbp1 , Pub1 and Pab1 under glucose replete conditions [73–75] , however it has yet to be determine if changes in glucose impact the acetylation status of these sites and if these lysine acetylations impact SG formation . Direct acetylation of SG proteins by NuA4 could also be the secondary molecular mechanism through which NuA4 is regulating GD-SGs distinct from the regulation of acetyl-CoA levels ( Fig 8F ) . We also cannot exclude the model that acetylation of SG subunits are promoting GD-SG formation as seen for acetylation of TDP-43 that promotes its aggregation into stress granules [24] . Given the depth and dynamic nature of lysine acetylation that is occurring in the cell , a likely scenario is that lysine acetylation controls multiple aspects of SG dynamics . If acetyl-CoA is suppressing GD-SG formation through hyper-activation of a KAT , our work suggests it is not through either NuA4 or Gcn5 . Exogenous acetate does not require Eaf7 or Gcn5 to suppress GD-SG formation ( Fig 6F ) and elevating acetyl-CoA levels by decreasing Acc1 protein levels , still suppresses GD-SGs even in eaf7Δ cells ( Fig 8F ) . Furthermore , if suppression was through a KAT , one would predict its deletion would result in increased GD-SG formation . However , our screen of known and putative non-essential KATs did not identify such a mutant ( S2 Table ) , albeit the essential KAT ECO1 was not screened and functional redundancies of the KATs may also be hindering identification . As proposed above , if KATs and KDACs regulate multiple aspects of SG dynamics , including promotion , maintenance and disassembly , our simple KAT/KDAC screen would not suffice and will require detailed kinetics or identification of specific acetylation sites . Alternatively , the impact of acetyl-CoA on GD-SGs may be occurring independent of KATs . Non-enzymatic acetylation of proteins occurs widely [75] and increased acetyl-CoA could led to acetylation of exposed lysine charge patches influencing aggregation of SG proteins directly or acetylation of yet-to-be-identified pathways contributing to GD-SG formation . Acetyl-CoA is also required for a wide variety of enzymes and pathways including the stimulation of KDAC activity [76] . As KDAC mutants display increased number of SGs even under non-stress conditions ( S2 Table and [10] ) , acetyl-CoA activation of KDACs may contribute to the mechanism ( s ) by which acetyl-CoA delicately balances GD-SGs . Given the plethora of acetylation sites on SG proteins and the breath of enzymes/pathways regulated by acetyl-CoA , likely multiple mechanisms are mediating acetyl-CoA’s role in GD-SG dynamics . Further this work challenges the paradigm that KATs are only mediating the effects of acetyl-CoA through acetylation of histones and other substrates . Rather our work implicates NuA4 as regulating acetyl-CoA level through Acc1 the rate-limiting enzyme for de novo fatty acid synthesis ( Figs 8 and 9 ) . The additive effects of eaf7Δ and gcn5Δ potentially suggest that each KAT is regulates acetyl-CoA levels through distinct mechanism or they have a shared substrate that regulates acetyl-CoA levels , and in the absence of one KAT , the other is able to partially compensate . The first scenario is the most likely as ACC1 transcription is decreased in gcn5 cells [77] , but transcriptome studies have not detected a role NuA4 in regulation of the gene expression of ACC1 [78 , 79] nor do we detect an impact on Acc1 protein levels in eaf1Δ or eaf7Δ cells ( S7 Fig ) . How is NuA4 regulating Acc1 ? Acetylome studies have determined that Acc1 is heavily acetylated in both human and yeast cells and that acetylation changes upon stress [73–75 , 80 , 81] . Nor is acetylation limited to Acc1 , but sites were detected on all proteins of the Fatty Acid Synthases ( FAS ) pathway , suggesting that lipogenesis could be directly regulated by acetylation of enzymes . Future studies will need to be conducted to determine whether proteins of the FAS pathway , including Acc1 , are regulated by NuA4-dependent acetylation . In conclusion , our study not only identifies NuA4 as a regulator of GD-SG assembly in yeast , but identifies the central metabolite acetyl-CoA as a rheostat to fine-tune GD-SG dynamics . Though the exact mechanism by which NuA4 and acetyl-CoA regulate GD-SGs remains unclear , our work suggests that NuA4 is not acting solely downstream of acetyl-CoA , but contributes to regulating acetyl-CoA levels through Acc1 . The complex cross-talk between NuA4 and acetyl-CoA maybe critical to link glucose availability to the cellular stress response .
S . cerevisiae strains and plasmid used in this study are listed in S2 Table . Genetic deletion or epitope tag integrations were generated using standard PCR-mediated techniques as previously described [82] , and validated by PCR with sequence-specific primers and/or Western Blot . Yeast cultures were grown at 30°C unless otherwise indicated with constant shaking at 200 rpm in standard YPD medium ( 1% yeast extract , 2% peptone , 2% dextrose ) or synthetic complete SCD medium ( 0 . 67% yeast nitrogen base without amino acids , 0 . 2% amino acid drop out mix , 2% dextrose ) . Glucose deprivation , acetate and heat-shock stresses were conducted as described previously [13 , 28 , 83] . Briefly , yeast strains were grown in the appropriate media overnight before dilution to an OD600 of 0 . 1 and grown to mid-logarithmic growth ( OD600 of 0 . 5–0 . 8 ) prior to treatment . For glucose depletion and control experiments , cells were collected by centrifugation ( 3 min at 3 , 000 rpm at 30°C ) , washed in 30°C pre-warmed medium with or without glucose , re-suspended in either glucose or glucose-deprived medium , and then incubated with constant shaking for 10 mins unless otherwise indicated . For acetate treatment , acetate ( 2 . 5M Acetate Solution , Sigma-Aldrich; cat . # 3863 ) was added to re-suspension media to a final concentration of 100 mM for 10 min . For heat-shock stress , yeast cultures were collected by centrifugation ( 3 min at 3 , 000 rpm at 30°C ) , re-suspended in pre-warmed medium at 46°C , and then incubated at 46°C for 10 min . For ethanol stress , ethanol was added to exponentially growing cells to a final concentration of 15% for 10 min . Similar to acetate suppression of glucose deprivation , 2% ethanol , 2% citrate ( sodium citrate dihydrate: Fisher Scientific , S279-3 ) or 2% pyruvate ( sodium pyruvate , Sigma-Aldrich , cat# P2256 ) was added to glucose-deprived medium for 10 min . For the yeast strains carrying the tetracycline-repressible tet07 promoter , doxycycline ( Sigma-Aldrich; cat . # D3447 ) was added to exponentially growing cells to a final concentration of 10 uM for 2 . 5 h [59] prior to stress treatment and doxycycline exposure was maintained throughout the experiment . After stress treatment , 1 ml of yeast cells was collected by centrifugation ( 3 min at 3 , 000 rpm at room temperature ) , re-suspended in 20 μl of pre-warmed SC medium to maintain stress ( example SC +/- glucose ) , prior to immediate microscopic examination in absence of fixative agents . A Leica fluorescence microscope ( DMI6000; Leica Microsystems ) equipped with a high-performance camera ( Hamamatsu ) , DG4 light source ( Sutter Instruments ) and Volocity 4 . 3 . 2 software ( PerkinElmer ) was used for all of the imaging acquisition , processing , including blind quantification scoring of the SGs [84] . All experimental images were captured as Z-stacks ( 0 . 2 μm steps across 6 μm ) , which were collapsed and used for manual image quantification in a blind manner . A least three independent biological replicates for each yeast strain and condition were performed with 100 cells/replicate scored for SGs ( % of cells with SGs ) and 50 cells/replicate scored for the number of SGs/cell . From these values , the percentage of cells that form SGs was determined , and statistical significance below the p-value of 0 . 05 was measured using two-way ANOVA or unpaired t-tests . As indicated in text , whole cell extracts ( WCE ) were prepared using either the modified immunoprecipitation ( mChIP ) [85] or Trichloroacetic acid ( TCA ) [86] protocols . Protein concentration of the mChIP WCE was measured by using Bradford reagent ( Bio-Rad; cat . # 500–0006 ) according to the manufacturer’s instructions . Equal numbers of cells were used for TCA WCE and equal volumes of WCE were utilized for Western Blot analysis . Proteins were separated by SDS-PAGE gels and electrophoretically transferred to nitrocellulose membranes ( PALL; cat . # 66485 ) . Membranes were incubated for 60 min at room temperature ( RT ) in blocking buffer ( PBST-5% milk ) , prior to incubation overnight at 4°C with the primary antibody diluted in blocking buffer ( PBST-5% milk ) . The immunoblots were washed three times in PBST and incubated with the secondary antibody for 2 h at RT in blocking buffer ( PBST-5% milk ) . The immunoblots were washed again three times in PBST , and the proteins were visualized by chemiluminescence ( Clarity ECL substrate , Bio-Rad ) . The following dilutions of primary antibodies were used: 1:5 , 000 mouse monoclonal anti-GFP ( Sigma-Aldrich; cat . # 11814460001 ) ; 1:1 , 000 rabbit polyclonal anti-hyperacetylated histone H4 ( EMD Millipore; cat . # 06–946 ) ; 1:1 , 000 rabbit monoclonal anti-histone H3 ( EMD Millipore; cat . # 05–928 ) ; 1:1 , 000 rabbit polyclonal anti-Acc1 ( Cell Signaling; cat . # 3662 ) ; 1:10 , 000 rabbit polyclonal anti-G6PDH ( Sigma-Aldrich; cat . # A9521 ) . The following dilutions of secondary antibodies were used: 1:5 , 000 goat-anti-mouse horseradish peroxidase ( HRP ) -conjugated ( BioRad; cat . # 170–6516 ) and 1:5 , 000 goat-anti-rabbit HRP-conjugated ( Chemicon; cat . # AP307P ) . As indicated , protein level quantification was performed using either standard densitometry with an internal protein control or the entire lane signal using the TGX Stain-Free FastCast Acrylamide Kit ( Bio-Rad; cat . # 161–0181 ) as previously described [87 , 88] . Statistical significance below the p-value of 0 . 05 was measured using ANOVA or unpaired t-tests . Briefly , 50 ml of yeast cells cultured in glucose or subjected to glucose deprivation stress for 10 min were harvested by centrifugation ( 3 min at 3 , 000 rpm at 4°C ) in the presence of cycloheximide ( Sigma-Aldrich; cat . # C4859 ) at the final concentration of 100 μg/ml . The cells were washed once in cold distilled water containing 100 μg/ml cycloheximide , and cell pellets were stored at -80°C . Ribosomal distribution was then performed as previously described [89] . 10 A260 units diluted in 400 μl of lysis buffer were loaded onto 15–45% linear sucrose gradients . The gradients were centrifuged ( 90 min at 39 , 000 rpm ) in a SW41 rotor ( Beckman Instruments ) , after which the gradients were collected from the top with a gradient fractionation system ( Brandel ) . The A254 was measured continuously to generate traces . HeLa and MCF7 cell lines were purchased from the American Type Cell Collection ( ATCC ) . HeLa cells were grown and maintained in Dulbecco’s modified Eagle’s Medium ( DMEM ) supplemented with 2 mM L-glutamine and 10% fetal bovine serum . MCF7 cells were grown in DMEM medium , but supplemented with 2 . 75 μg/ml insulin . HeLa and MCF7 cells were treated with the following drugs: NU9056 ( Tocris; cat . # 4903 ) was diluted in DMSO and used at a final concentration of 20 μM for 24 h; Bortezomib ( LC Laboratories; cat . # B-1408 ) was diluted in DMSO and used at a concentration of 10 μM for 4 h; Sodium arsenite ( Sigma-Aldrich; cat . # S7400 ) was diluted in Phosphate Buffer Saline ( PBS ) and used at a final concentration of 100 μM for 60 min . Immunofluorescence was then performed on the cells at the end of the specific drug treatment . Cells were washed three times with 1X PBS , fixed in 4% paraformaldehyde , washed again three times with 1X PBS , and permeabilized with 0 . 5% Triton X-100 in 1X PBS . Cells were washed three times , incubated with the primary antibody for 60 min , then washed once with 0 . 1% Triton X-100 in 1X PBS , and washed twice with 1X PBS . Cells were incubated with Alexa Fluor secondary antibody for 60 min , and washes were done similarly as after the incubation with the primary antibody . Cells were mounted on slides using Vectashield Hard Set Mounting Media ( Vector Labs ) . Fluorescence microscopy was performed using the Zeiss Axio Imager Z1 microscope , and images were acquired with an AxioCam HR camera utilizing Zeiss Axiovision 4 . 5 software . The following primary antibodies were used: anti-FMRP ( 1:100; EMD Millipore; cat . # MAB2160 ) ; anti-DDX3 ( 1:100; Bethyl Laboratories; cat . # A300-474A ) ; anti-TIA-1 ( 1:50; SantaCruz Biotechnology; cat . # SC-1751 ) . The following secondary antibodies ( Life Technologies ) were used: rabbit ( 1:200; cat . # A110012 ) , mouse ( 1:200; cat . # A11005 ) , and goat ( 1:100; cat . # A11058 ) Alexa Fluor 594 nm antibodies . Cells were washed twice with cold 1X PBS and lysed for 10 min in TEB buffer ( 0 . 5% Triton X-100; 2 mM phenylmethylsulfonyl fluoride , 0 . 02% sodium azide in 1X PBS ) . Nuclei were pelleted by centrifugation ( 10 min at 6 , 500 g at 4°C ) , washed once in TEB buffer , and centrifuged again . The pellet was resuspended in 0 . 2 N hydrochloric acid in 1X PBS , and histones were extracted overnight at 4°C , then centrifuged ( 10 min at 6 , 500 g at 4°C ) . Protein concentration was measured by Bradford reagent , and histones were resolved by SDS-PAGE . Proteins were transferred to Immunobilon-P polyvinylidene difluoride membranes ( PVDF; EMD Millipore ) . Membranes were incubated for 60 min at RT in blocking buffer ( PBST-5% milk ) . Membranes were then incubated overnight at 4°C with the primary antibody diluted in blocking buffer ( PBST-2% milk ) . The immunoblots were washed three times in PBST and incubated with the secondary antibody for 60 min at RT in blocking buffer ( PBST-2% milk ) . The immunoblots were washed again three times in PBST , and the proteins were visualized by using the chemiluminescence detection kit ( Luminata ECL substrate , EMD Millipore ) . The following primary antibodies and dilutions were used: 1:1000 rabbit polyclonal anti-hyperacetylated histone H4 ( EMD Millipore; cat . # 06–946 ) ; 1:10 , 000 mouse monoclonal anti-Tubulin ( Sigma-Aldrich; cat . # T6199 ) . The following secondary antibodies were used: 1:20 , 000 goat anti-mouse HRP-conjugated ( Cappel; cat . # 55479 ) ; 1:20 , 000 goat anti-rabbit HRP-conjugated ( Cappel; cat . # 55690 ) . 50 ml of yeast cell culture were harvested and immediately processed for acetyl-CoA extraction as previously described [90] . Acetyl-CoA level was measured using an Acetyl-Coenzyme A assay kit ( Sigma-Aldrich; cat . # MAK039 ) according to the manufacturer’s instructions . A SynergyH1 Multi-Mode Plate Reader ( BioTek ) was used to measure the fluorometric product ( λex5 = 535/λem = 587 ) of each assay reaction . Acetyl-CoA carboxylase assay was performed similar to previously described [64] with adapted for yeast extraction . Exponentially growing cells were lysed in lysis buffer ( 100 mM Hepes pH 8 . 0 , 20 mM MgOAc , 300 mM NaOAc , 10% Glycerol , 10 mM EGTA , 0 . 1mM EDTA and 1% NP-40 ) supplemented with mini protease inhibitor cocktail tablet ( Roche; cat . # 4693159001 ) , KDAC inhibitors ( 50 mM nicotinamide , 50mM sodium butyrate , 0 . 4μM apicidin ( Sigma; cat . # EPI008-1KT ) , 2μM M344 ( Sigma; cat . # EPI008-1KT ) , 250μM splitomycin ( Sigma; cat . # EPI008-1KT ) , 5μM TSA ( Sigma; cat . # EPI008-1KT ) , 10mM Sodium 4-phenylbutyrate ( Sigma; cat . # EPI008-1KT ) and 1mM valproic sodium salt ( Sigma; cat . # EPI008-1KT ) ) and phosphatase inhibitors ( 5mM NaF , 1mM Na3VO4 , phosphatase inhibitor cocktail 2 and 3 ( Sigma; cat . # P5726 and P0044 , respectively ) ) by bead beating . Cell lysates were obtained by centrifugation at 16 , 000 x g for 15 min at 4°C . Acc1 assay was performed by incubating equal amounts of cell lysates with reaction buffer ( 50 mM Hepes pH 7 . 4 , 10mM MgCl2 , 1 mM MnCl2 , 2 mM DTT , 0 . 4 mM ATP , 0 . 075% fatty acid free BSA ( Sigma; cat . # A7030 ) , 12 . 5 mM NaHCO3 and 1 . 5 μCi 14C-NaHCO3 ( Perkin Elmer; cat . # NEC086H005MC ) with or without 10 mM citrate . The reactions were incubated at room temperature for 90 min , stopped by the addition of 0 . 6 N HCl and air-dried overnight at 37°C . The radioactivity of the products was determined by scintillation counting of the dried materials . Specific activity was calculated as the amount of substrate conversion divided by 90 min and the amount of total protein in the lysate as determined by microBCA kit ( Thermo Scientific; cat . # 23231 ) . ANOVA and Turkey’s post hoc test were performed with data from 3 biological replicates , each of which has 3 technical replicates .
|
In response to environmental stress , such as nutrient limitations or toxic chemicals , cells must quickly counteract these threats in order to survive . One way cells fight environmental challenges is through the formation of stress granules , which are aggregates of proteins and mRNA within the cytoplasm . Though their formation is essential for survival under multiple conditions , when stress granules are inappropriately formed they become causative for diseases such as amyotrophic lateral sclerosis and fragile X syndrome . Further stress granules contribute to chemotherapy resistance of cancer cells by promoting survival . Therefore it is critical to understand how stress granules are formed and disassembled . Here using the budding yeast Saccharomyces cerevisiae we determine that an enzyme called NuA4 is contributing to stress granule formation upon glucose deprivation . Why is this important ? We determine that Tip60 , the equivalent of NuA4 in mammalian cells , is also regulating stress granule formation in cancer cells . As there is a growing number of drugs that target this class of enzyme , there is the possibility that these drugs may reduce stress granule formation offering a novel therapeutic approach to treat numerous diseases .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"carbohydrate",
"metabolism",
"chemical",
"compounds",
"dna-binding",
"proteins",
"vertebrates",
"carbohydrates",
"glucose",
"metabolism",
"organic",
"compounds",
"glucose",
"animals",
"mammals",
"cell",
"signaling",
"fungi",
"model",
"organisms",
"glucose",
"signaling",
"experimental",
"organism",
"systems",
"amniotes",
"research",
"and",
"analysis",
"methods",
"saccharomyces",
"proteins",
"chemistry",
"histones",
"yeast",
"biochemistry",
"signal",
"transduction",
"organic",
"chemistry",
"post-translational",
"modification",
"cell",
"biology",
"acetylation",
"cats",
"monosaccharides",
"biology",
"and",
"life",
"sciences",
"chemical",
"reactions",
"physical",
"sciences",
"saccharomyces",
"cerevisiae",
"yeast",
"and",
"fungal",
"models",
"metabolism",
"organisms"
] |
2017
|
Lysine acetyltransferase NuA4 and acetyl-CoA regulate glucose-deprived stress granule formation in Saccharomyces cerevisiae
|
Dorsoventral patterning of the embryonic axis relies upon the mutual antagonism of competing signaling pathways to establish a balance between ventralizing BMP signaling and dorsal cell fate specification mediated by the organizer . In zebrafish , the initial embryo-wide domain of BMP signaling is refined into a morphogenetic gradient following activation dorsally of a maternal Wnt pathway . The accumulation of β-catenin in nuclei on the dorsal side of the embryo then leads to repression of BMP signaling dorsally and the induction of dorsal cell fates mediated by Nodal and FGF signaling . A separate Wnt pathway operates zygotically via Wnt8a to limit dorsal cell fate specification and maintain the expression of ventralizing genes in ventrolateral domains . We have isolated a recessive dorsalizing maternal-effect mutation disrupting the gene encoding Integrator Complex Subunit 6 ( Ints6 ) . Due to widespread de-repression of dorsal organizer genes , embryos from mutant mothers fail to maintain expression of BMP ligands , fail to fully express vox and ved , two mediators of Wnt8a , display delayed cell movements during gastrulation , and severe dorsalization . Consistent with radial dorsalization , affected embryos display multiple independent axial domains along with ectopic dorsal forerunner cells . Limiting Nodal signaling or restoring BMP signaling restores wild-type patterning to affected embryos . Our results are consistent with a novel role for Ints6 in restricting the vertebrate organizer to a dorsal domain in embryonic patterning .
The vertebrate embryonic dorsal organizer , historically referred to as the Spemann organizer , breaks the symmetry of the blastula by defining its dorsal side and ultimately gives rise to axial mesoderm , which forms the notochord , the defining anatomical feature of the chordate lineage . In fish and frogs , induction of the organizer relies on a maternal Wnt signaling pathway that leads to the accumulation of β-catenin in nuclei on the prospective dorsal side of the embryo [1] , [2] . A primary function of the organizer is to induce a region in the embryo that is competent to adopt dorsal fates , such as prechordal plate mesoderm and neural ectoderm , in the presence of widespread ventralizing BMP signaling . Proper partitioning of axial versus non-axial cell fates during gastrulation is essential to ensure proper embryonic patterning . BMP signaling patterns tissues along the dorsoventral axis ( DV ) , but does not act to partition axial versus non-axial fates . For example , in zebrafish bmp2b ( swirl ) ligand mutant embryos , loss of BMP signaling causes the expansion of dorsal neurectodermal and non-axial dorsal mesodermal cell fates at the expense of ventral cell fates without expanding the organizer itself [3] , [4] , [5] , [6] . Thus , in the absence of ventral cell fate specification , other mechanisms ensure that the organizer is confined dorsally . In zebrafish and Xenopus , several maternal and zygotic genes function to restrict the organizer to dorsal regions . Three related homeodomain-containing transcriptional repressors , Vox , Vent , and Ved play a key role in repressing dorsal organizer gene expression ventrolaterally in zebrafish [7] , [8] , [9] , [10] , [11] . These repressors are expressed ventrally and dorsolaterally , and their deficiency causes dorsal organizer gene expression to expand around the ventrolateral margin during late blastula stages at the expense of ventrolateral tissues . Xenopus Vox and Vent have been shown to directly repress the expression of the organizer genes chordin ( chd ) and goosecoid ( gsc ) [12] , [13] and depletion of Gsc has been reported to lead to a 25-fold increase in vent expression [14] . Similarly in zebrafish , Vox and Vent have been shown to bind to the gsc promoter and to physically associate with Gsc protein [8] , [9] . These and other data [15] , [16] illustrate the cross-regulatory interactions between opposing ventralizing and dorsalizing transcriptional repressors that are essential for proper embryonic patterning . Several additional genes are known to restrict organizer gene expression to dorsal regions and modulate the expression of vox , vent , and ved . Knockdown of Runx2bt2 , a maternal isoform of Runx2b , delays induction of vox and vent , and eliminates ved expression [17] . Embryos deficient in maternal Runx2bt2 exhibit an expansion of dorsal organizer gene expression at late blastula stages with a reciprocal loss of ventrolateral tissues [17] . Expression of vox , vent , and ved is maintained during late blastula and early gastrula stages by zygotic Wnt8a signaling [10] , [18] . By mid gastrulation , expression of these ventralizing transcriptional repressors is maintained by BMP signaling [10] , [19] . Thus , a gene regulatory network involving Runx2bt2 , Wnt8a and BMP signaling converges on vox , vent , and ved to maintain the specification of non-axial mesoderm . The maternally supplied transcription factor pou5f3 ( previously called pou5f1 ) also functions in restricting the organizer to the dorsal midline . Maternal-zygotic deficiency of pou5f3 ( MZpou5f3 ) leads to severe dorsalization resulting from derepression of organizer genes ventrolaterally in the embryonic margin , and incomplete induction of the BMP pathway [20] . MZpou5f3 mutants also exhibit aberrant morphogenesis and fail to form endoderm [20] , [21] , [22] . Pou5f3 likely functions as a transcriptional activator of genes , including vox , that are required to repress dorsal organizer gene expression ventrolaterally [23] , [24] . Thus , Pou5f3 is another mediator of organizer gene repression operating in parallel to the Wnt8a pathway and partially through the BMP pathway . One of the earliest organizer genes induced downstream of the maternal Wnt pathway in zebrafish is bozozok ( boz ) , a direct transcriptional repressor of bmp2b , vox , vent , and ved expression [8] , [25] , [26] , [27] , [28] , [29] , [30] . boz mutant embryos fail to form prechordal plate , notochord , forebrain , and ventral neural structures and display an increase of ventroposterior mesoderm [25] . boz mutant embryos can be rescued by suppressing Wnt8a signaling , indicating that Boz antagonizes zygotic wnt8a expression in the organizer to block non-axial fate development in the dorsal embryonic midline and allow axial development [26] . Boz stability is modulated by Lnx2b , a maternally supplied E3 ubiquitin ligase that can directly bind and ubiquitinate Boz [31] , [32] . Loss of Lnx2b causes expression of boz and other organizer genes to expand into lateral regions of the late blastula , illustrating the importance of proper turnover of Boz . The transcriptional repressors Vox , Vent , and Ved are essential for partitioning the mesoderm into axial versus non-axial domains in response to positive regulation from Runx2bt2 , Pou5f3 , the Wnt8 , and BMP pathways , and negative inputs from dorsalizing transcriptional repressors such as Boz and Gsc . It is less clear how these pathways are molecularly integrated to regulate vox , vent , and ved expression and it is likely that additional maternally-provided factors function in this process . Accordingly , we performed a genetic screen for maternal-effect mutations to identify novel mediators of vertebrate embryonic patterning . We isolated a novel recessive maternal-effect mutation p18ahub that causes a profound reduction in ventrolateral mesoderm with a reciprocal expansion in axial mesoderm , and frequently multiple independent axial-like domains . Consistent with radial expansion of the organizer , p18ahub mutant females produce embryos exhibiting ectopic dorsal forerunner cells , a unique population of non-involuting mesendodermal cells at the dorsal margin [33] , [34] , [35] . We can rescue p18ahub dorsalized mutant embryos either by limiting Nodal signaling or restoring BMP signaling . We determined through positional cloning that p18ahub is a mutation disrupting the integrator complex subunit 6 ( ints6 ) gene , which encodes a highly conserved component of the Integrator Complex , a large multisubunit complex implicated in 3′ end processing of spliceosomal snRNAs [36] . Previously , ints6 was named deleted in cancer 1 ( dice1 ) and investigated as a putative tumor suppressor gene in humans [37] , [38] , [39] , [40] . Using a forward genetic approach , we have revealed a novel role for Ints6 in limiting the extent of dorsal organizer tissues during vertebrate embryogenesis .
We isolated p18ahub , a recessive maternal-effect dorsalizing mutation , in an ENU-induced mutagenesis screen designed to identify novel maternal factors required for early embryonic development and patterning in zebrafish ( similar to that described in [41] , [42] ) . Mutant females yielded embryos with similar phenotypes whether they were crossed to mutant or wild-type ( WT ) males , indicating a strictly maternal-effect defect with no zygotic contribution . The first defect evident in embryos from p18ahub mutant mothers ( henceforth referred to as p18ahub mutant embryos ) was a delay in the initiation of epiboly , the morphogenetic process by which the blastoderm cells move over and encompass the yolk cell [43] . As WT embryos reached the late blastula/50% epiboly stage ( Figure 1A ) , mutant embryos typically had not initiated epiboly movements ( Figure 1B ) . In early gastrulation stages , WT embryos displayed a single dorsal thickening corresponding to the embryonic axis ( Figure 1C ) , whereas mutant embryos often developed a radial thickening possibly due to hyper convergence of cells around the entire embryonic margin ( Figure 1D ) . Approximately 50% of mutant embryos ( Figure 1G ) lysed prior to 24 hpf . Thirty-five percent of mutant embryos surviving to 24 hours post fertilization ( hpf ) exhibited radial symmetry around the animal-vegetal axis and lacked any recognizable structures ( Figure 1E , F ) . We categorized such embryos as class 6 dorsalized embryos . Class 5 embryos lacked recognizable structures but were not radially symmetric around the animal-vegetal axis and typically did not survive to 24 hpf . Class 4 through class 1 embryos ( Figure 1G ) exhibited progressively less severe dorsalization , as described [5] . To simplify the presentation of the data , we have combined phenotypic classes , C1–C3 , and , C4–C6 , in all figures , except that C5 embryos that did not survive to 24 hpf are included in a lysed category . To better examine the epiboly and lysis defects of p18ahub embryos , we conducted time-lapse imaging ( supplemental Movie S1 and Movie S2 ) of embryos from mid-blastula to mid-somitogenesis stages . p18ahub embryos were developmentally delayed and in some mutant embryos displayed prolonged epiboly ( Movie S2 ) . Some p18ahub embryos failed to undergo epiboly whatsoever and lysed by the equivalent of early gastrula stage in WT , with cells dispersing rapidly and the blastoderm disintegrating ( Movie S1 ) . In other p18ahub embryos , epiboly progressed to just prior to yolk plug closure ( 100% epiboly ) when the hypoblast rapidly retracted and either the embryo lysed or the tissue dived down into the animal pole of the yolk cell ( Movie S2 ) , accounting for the morphology of embryos like the one shown in Figure 1F . Based on these studies it is likely that the lysed p18ahub embryos for a given clutch displayed either the early lysis phenotypes or were class 5 or 6 dorsalized . To investigate if p18ahub embryos exhibit altered DV patterning , we examined the expression of genes of the dorsally-derived neurectoderm , dorsolaterally-derived somitic and paraxial mesoderm , and ventrally-derived pronephros , by in situ hybridization on 3- to 5-somite stage embryos . The expression of both six3 , which marks forebrain neurectoderm [44] , and pax2 . 1 , a marker of the midbrain-hindbrain boundary [45] , was circumferentially expanded in p18ahub embryos ( Figure 1H , I ) . krox20 , which is expressed in rhombomeres 3 and 5 in the hindbrain [46] , was often undetectable in severely dorsalized p18ahub embryos due to their severe delay ( 8/23 embryos displayed krox20 expression for the clutch represented in Figure 1I ) . In p18ahub embryos able to develop longer , krox20 expression was also expanded circumferentially ( Figure 1J , K ) . The expression of myod , a marker of paraxial and somitic mesoderm [47] , was scattered in clusters of cells distributed circumferentially ( Figure 1I , verified by examination of myod probe alone , not shown ) . Pronephric pax2 . 1 expression was often undetectable in mutant embryos ( compare Figure 1I p* with H; verified by examination of pax2 . 1 probe alone , not shown ) , indicative of a severe reduction of ventroposterior mesoderm . To investigate if patterning is also affected during gastrulation in p18ahub embryos , we examined the expression of the fore- and mid-brain marker otx2 [48] at mid gastrulation . We found that otx2 expression was expanded around the DV axis rather than restricted to a dorsoanterior region as in WT ( Figure 1L and M ) . Importantly , however , otx2 was not expanded posteriorly as in wnt8a mutants [18] , suggesting that the Wnt8a pathway is intact in p18ahub embryos . We also examined the expression of cyp26a and hoxb1b , markers of anterior neurectoderm and caudal hindbrain , respectively [49] , [50] . Consistent with dorsalization , p18ahub embryos displayed expanded cyp26a expression around the DV axis ( Figure 1N , O ) . Note also that the p18ahub embryos were delayed developmentally; the margin of the mutant embryo ( the equivalent time of bud stage for WT ) has not advanced as far as in WT . Compared to WT ( Figure 1N ) , the p18ahub embryo displays reduced marginal cyp26a expression , which is consistent with the WT cyp26a expression pattern at earlier gastrula stages [50] . Expression of hoxb1b in the posterior neurectoderm is expanded in p18ahub embryos around the margin , although with reduced intensity compared to WT ( Figure 1P , Q ) , likely also reflecting developmental delay [49] . It was necessary to age-match embryos for these experiments because we could not obtain sufficient numbers of p18ahub embryos at mid-gastrula stage due to their lysis . These data indicate that p18ahub embryos display an expansion of dorsal neurectodermal cell fates during gastrulation , but unlike zygotic wnt8a mutants , no significant expansion of anterior at the expense of posterior neurectoderm is observed in p18ahub embryos . To determine if the dorsalization of p18ahub embryos results from impaired BMP signaling , we examined BMP ligand gene expression , as well as expression of the BMP antagonist chordin ( chd ) in mutant embryos . bmp2b expression appeared normal in late blastula stage mutant embryos ( data not shown ) . However , bmp2b and bmp4 expression were significantly reduced in mutant embryos by the early gastrula stage ( Figure 2A , B , E , F ) and severely reduced by mid gastrulation ( Figure 2C , D , and data not shown ) , consistent with dorsalization . Since BMP gene expression is controlled by an autoregulatory feedback loop [3] , , the loss of bmp2b and bmp4 expression in mutant embryos indicates severely reduced BMP signaling . chd expression , which is restricted to the dorsal side of the early zebrafish gastrula [53] , is circumferentially expanded in mutant embryos at an early gastrula stage ( Figure 2G , H ) , consistent with reduced BMP signaling and excessive dorsal fate specification in p18ahub embryos . To determine if p18ahub embryos are defective mechanistically in BMP signal transduction , we injected mutant embryos with bmp2b mRNA , which moderately to severely ventralizes WT embryos [6] ( Figure 2J ) . We found that forced expression of bmp2b restored ventral fates to WT in two-thirds of mutant embryos or moderately ventralized them ( Figure 2I , K , O , P ) . Thus , the BMP signal transduction machinery in mutant embryos can function when BMP ligand is provided exogenously . To further test whether the endogenous BMP signaling machinery is functional in p18ahub embryos , including the endogenous BMP ligands , mutant embryos were injected with translation-blocking morpholinos ( MOs ) targeting the secreted BMP antagonists Chordin , Noggin1 , and Follistatin-like 2b [54] . Depletion of these BMP antagonists in mutant embryos caused mild to moderate ventralization , similar to that of WT embryos ( Figure 2M , N ) . Thus , endogenous BMP ligands can signal in p18ahub embryos at a WT or greater level if BMP ligand function is permitted . Dorsalization can also be caused by a ventral expansion of dorsal organizer gene expression . To investigate the organizer in p18ahub mutant embryos , we examined expression of the organizer gene goosecoid ( gsc ) [55] . We found that , although gsc expression was induced normally at the mid blastula stage ( data not shown ) , it was expanded in p18ahub embryos by early gastrulation ( Figure 3A , B ) and remained ectopically expressed through mid gastrula stages ( Figure 3C , D ) compared to WT embryos . Therefore , the dorsalization of p18ahub embryos involves a prominent expansion of gsc expression by early gastrulation , contrasting dorsalization resulting solely from defective BMP signaling [5] . Since gsc expression was induced normally in p18ahub mutant embryos , it suggested that the organizer is induced normally by the maternal Wnt signaling pathway [1] , [2] . To directly test this , we examined β-catenin nuclear localization as a readout of the maternal Wnt signaling pathway . We immunostained mid-blastula stage embryos ( 3 . 5 hpf ) to visualize β-catenin in nuclei on the presumptive dorsal side of the embryo [1] . No differences were evident between WT and mutant embryos in β-catenin intensity or its localization in nuclei at the dorsal margin ( Figure 3E , F , arrows ) . No nuclear localized β-catenin was observed ventrally in mutant embryos . Consistent with normal induction of the organizer , boz , a direct transcriptional target of the maternal Wnt pathway [56] , was expressed normally in mutant embryos through 6 hpf , the equivalent of early gastrula stage ( Figure 3G–J ) , unlike gsc and chd , which were expanded ( Figure 2H and 3B ) . Thus , the organizer is induced normally in mutant embryos but the expression of some organizer genes becomes derepressed around the margin between late blastula and early gastrula stages . Furthermore , the dorsalization of p18ahub embryos does not rely upon ventrolateral expansion of boz . Zygotic Wnt8a signaling in ventrolateral regions represses the expression of the organizer genes , gsc and chd , thus restricting the size of the organizer [7] , [8] , [9] , [10] , [11] , [18] , [19] , [26] . Loss of Wnt8a or its mediators Vox , Vent , and Ved causes the expression of some organizer genes to expand and dorsalizes the embryo [7] , [8] , [9] , [10] , [11] , [19] . Accordingly , we investigated the status of the Wnt8a pathway in mutant embryos . In WT late blastula and early gastrula stage embryos , wnt8a is expressed at the margin in a large domain extending from ventral to dorsolateral regions ( Figure 4A ) [18] , [57] . However , in p18ahub late blastula stage embryos , wnt8a expression was restricted ventrally to a smaller domain ( Figure 4B ) . Furthermore , wnt8a expression was undetectable in mutant embryos age-matched to their WT counterparts at early gastrula stage ( not shown ) , perhaps reflecting a loss in the competence of marginal cells to express wnt8a . Since wnt8a is required for anteroposterior ( AP ) patterning of neural tissue [18] and a defect is not evident in AP patterning in p18ahub mutants , it suggests that the early , transient expression of wnt8a may be sufficient for AP patterning in p18ahub mutant embryos . We also examined the expression of vox and ved , two genes encoding transcriptional repressors acting downstream of Wnt8a to restrict organizer gene expression to the dorsal midline [10] , [11] , [19] . We found that vox expression was reduced in mutant embryos compared to age-matched WT embryos at a late blastula stage ( Figure 4C , D ) . By early gastrulation , vox expression was greatly reduced or absent in mutant embryos ( Figure 4E , F ) . We also conducted a time course experiment where embryos were collected from WT and p18ahub females over half hour intervals beginning at a mid-blastula stage ( 3 hpf ) and subsequently examined for wnt8a and vox expression . In this experiment induction of neither gene was observed in p18ahub embryos by the equivalent of the early to mid gastrula stage ( 7 hpf ) in WT ( not shown ) . Maternal ved expression was evident in p18ahub embryos ( data not shown ) ; however , by an early gastrula stage ved expression was prominently reduced ( Figure 4G , H ) . Early gastrula stage embryos also displayed reduced or nearly absent expression of eve1 , a marker of ventroposterior mesoderm that requires BMP and Wnt8a signaling for its expression ( Figure 4I , J ) [58] , [59] , [60] . Thus , key mediators of the Wnt8a pathway are repressed in p18ahub embryos beginning as early as the late blastula stage , which can account for the loss of ventroposterior mesoderm . We could not restore Wnt8a signaling in mutant embryos via injection of wnt8a mRNA , because early overexpression of wnt8a mimics the maternal Wnt signal for organizer induction and severely dorsalizes embryos [57] . To determine if the zygotic Wnt8a signaling pathway is functional in p18ahub embryos , we tested for Wnt8a function in p18ahub embryos rescued by depletion of BMP antagonists . If Wnt8a signaling is required , then it would indicate that the pathway is functional but likely fails to function in p18ahub embryos due to lack of full induction or maintenance of expression . To moderately increase BMP signaling in p18ahub embryos , we depleted one BMP antagonist , Chd , via MO injection . Loss of Chd alone rescued the majority of p18ahub embryos to a mild V1 ventralized phenotype , similar to loss of Chd in WT embryos ( Figure 4K , L , N , O ) . Loss of wnt8a dorsalized WT embryos [18] ( Figure 4K ) and also dorsalized chd deficient embryos or caused posterior truncations ( Figure 4M , P ) . Importantly , p18ahub embryos that were enabled to specify ventral tissues by Chd depletion became dorsalized again when Wnt8a was also depleted ( Figure 4K–M , O , Q ) . These results indicate that the Wnt8a pathway is mechanistically intact in p18ahub embryos and that their ventralization or rescue to a WT phenotype by enhancement of BMP signaling depends on endogenous Wnt8a signaling . Along with the maternal Wnt pathway , the dorsal organizer also depends on Nodal signaling for its induction ( reviewed in [61] ) . Thus , we investigated the status of the Nodal pathway in p18ahub embryos . We examined expression of the Nodal ligand nodal-related 1 ( ndr1 , squint ) in mutant and WT embryos [62] . ndr1 induction is initiated on the dorsal side of the embryo ( Figure 5A ) and requires Wnt signaling similarly to boz [2] . At a mid blastula stage , we observed no significant differences in the expression of ndr1 between WT and p18ahub embryos ( Figure 5B ) and ndr1 expression was never observed more animally outside of its normal marginal domain through an early gastrula stage equivalent ( not shown ) . To further test if Nodal signaling is induced normally in p18ahub embryos , we examined expression of the Nodal feedback antagonist lefty1 ( lft1 , antivin1 ) . lft1 is initially expressed dorsally but is subsequently expressed around the margin of the late blastula [63] , [64] , [65] . At an early gastrula stage ( 6 hpf ) , we observed no significant differences in the expression level of lft1 in mutant versus WT embryos ( Figure 5C and D ) . From these data we conclude that Nodal signaling in p18ahub embryos is induced normally and likely operates normally at an early gastrula stage . By mid gastrula stages the pan-mesodermal gene no tail ( ntl ) is expressed in two distinct domains , one corresponding to axial mesoderm , the developing notochord , and another corresponding to ventrolateral non-axial mesoderm [66] . At the equivalent of mid gastrula and early somite stages , ntl expression in presumptive non-axial mesoderm was reduced or absent , while axial mesoderm appeared to be expanded leading to multiple independent axes in some mutant embryos ( 9/9 , 6/10 , and 4/10 in three independent clutches ) ( Figure 5E–H″ ) . Some mutant embryos displayed a significantly broadened single axial ntl domain with a prominent reduction in marginal ntl expression ( Figure 5F inset ) . By the equivalent of early somite stages of development , mutant embryos clearly possessed multiple independent presumptive notochords marked by ntl expression ( Figure 5H ) . floating head ( flh ) , a homeobox gene required for notochord specification , is induced at a late blastula stage independently of ntl but is a direct transcriptional target of ntl by mid gastrula stages [67] , [68] , [69] . Consistent with excessive axial ntl expression , flh was also circumferentially expanded in p18ahub embryos by mid gastrulation ( Figure 5I , J ) . Therefore , genes marking axial mesoderm ( gsc and flh ) are ectopically expressed in p18ahub embryos by early gastrula stage , despite normal Wnt-mediated organizer induction and normal induction of the Nodal pathway . Nodal signaling is required for the expression of the HMG- type transcription factor sox32 ( casanova ( cas ) ) in endodermal precursors by late blastula stages [35] , [70] . cas is required along with pou5f3 ( formerly spiel-ohne-grenzen , pou5f1 ) to induce sox17 and maintain endodermal precursor cells [21] . Both cas and pou5f3 are also required for the maintenance of dorsal forerunner cells [21] , [33] , a distinct population of non-involuting cells that also express sox17 at the leading edge of the dorsal margin of the blastoderm as it migrates over the yolk [34] , [71] . We observed endodermal sox17 expression in p18ahub embryos at mid gastrulation ( Figure 5L ) . However , endodermal precursor cells were located more animally than in WT ( Figure 5K ) , perhaps due to altered cell movements resulting from loss of BMP signaling and dorsalization [72] , [73] . Ectopic marginal expression of sox17 was observed in p18ahub mid gastrula stage embryos indicating that ectopic dorsal forerunner cells form in mutant embryos ( Figure 5K , L , arrows ) . In p18ahub embryos axial mesoderm and organizer gene expression are expanded ventrolaterally , but remain confined within a normal mesodermal domain . Importantly , we did not observe the animal-ward expansion of Nodal-dependent genes such as ntl or lft1 or excessive specification of mesendodermal precursor cells , which has been reported upon excessive Nodal signaling due to loss of Lefty [74] , [75] , [76] or ectopic expression experiments [77] , [78] . Thus , Nodal signaling is likely not excessive in p18ahub embryos . The secreted feedback inhibitor Lefty/Antivin ( Lft1 ) regulates Nodal signaling [63] , [64] , [65] , [79] . Misexpression of Lft1 in WT embryos severely limits mesendoderm induction with embryos closely resembling ndr1;ndr2 double mutants [62] . We found that injection of as little as 0 . 7 picograms ( pg ) of lft1 mRNA ( Figure 6A , +0 . 7× middle row ) , a dose that only weakly perturbs WT embryos ( Figure 6A , Minor head defects ) , could restore WT or nearly WT patterning in 33% of mutant embryos . Injection of 1 pg of lft1 mRNA ( Figure 6A , +1× middle row ) restored WT patterning in a larger fraction of mutant embryos and suppressed mesendoderm formation ( oep-like ) in 20% of mutant embryos . The same dose of lft1 mRNA injected into WT embryos blocked mesendoderm development in more than 50% of the embryos ( Figure 6A , +1× bottom ) . Injection of 3 pg of lft1 mRNA into mutant and WT embryos inhibited mesendoderm induction in a similar fraction of embryos ( Figure 6A , +3× ) . Thus , suppression of Nodal signaling can restore the balance between axial and non-axial fate specification in mutant embryos , similarly to restoring BMP signaling . We examined lft1 expression to determine if its reduction contributed to excess dorsal mesodermal gene expression at mid gastrulation in p18ahub embryos . In WT embryos lft1 was expressed around the embryonic margin as well as in the developing axis and dorsal forerunner cells ( not shown ) at mid gastrula stages ( Figure 6B ) . In p18ahub embryos lft1 was expressed within the presumptive prechordal plate as well as in clusters of cells scattered around the margin at a mid gastrula stage ( Figure 6C ) . These latter cells may represent remaining marginal cells having non-axial fates . Hence , an absence of lft1 expression cannot account for the severe dorsalization of p18ahub embryos . To identify the molecular nature of p18ahub , we mapped the mutation to a chromosomal position by examining linkage to simple sequence length polymorphic ( SSLP ) markers . We first found linkage of p18ahub to z1660 on chromosome 9 . Further fine mapping examining over 1100 meioses placed p18ahub within a 1 . 35 Mb interval between the SSLP marker z7120 and an SSLP marker that we generated for BAC CR545476 . 14 ( Zv9 ) . ( Figure 7A ) . The interval displays synteny with human chromosome 13 . It contains just over one dozen genes , none of which were known to function in DV patterning . We proceeded to sequence the open reading frames of cDNAs of genes within the interval . We found a T to A transition in an exon of the integrator complex subunit 6 ( ints6 ) gene ( GenBank Accession number KF700696 , OMIM 604331 ) , converting a nearly invariant valine to an aspartate at position 375 of the 854 amino acid predicted protein ( Figure 7B ) . Human and zebrafish Ints6 orthologs are 66% identical indicating an overall high degree of evolutionary conservation . The only recognizable domain in Ints6 is an N-terminal von Willebrand factor type A motif ( InterPro IPR002035 ) , a broadly employed motif mediating interactions between diverse proteins . There is a second gene , ddx26b , on the X chromosome in humans and on chromosome 14 in zebrafish that is highly related to ints6 . The zebrafish Ints6 and Ddx26b proteins are 61% identical also implying significant conservation of function between these homologs . We examined the expression of ints6 by in situ hybridization . ints6 transcripts were present in eggs and embryos through the late blastula stage ( Figure 7C , D , data not shown ) . ints6 transcripts declined during late blastula stages and were undetectable by early gastrulation ( Figure 7 , data not shown ) . We observed no alteration in the expression of ints6 in mutant embryos by in situ hybridization ( not shown ) . These expression data are consistent with the recessive , maternal-effect inheritance of p18ahub and the maternal requirement for ints6 . To determine if dorsalization of p18ahub embryos was caused by the mutation in ints6 , we injected mutant embryos with mRNA encoding WT Ints6 . We found that as little as 5 pg of WT ints6 mRNA rescued 50% of mutant embryos completely ( Figure 7E ) and 50 pg rescued nearly 80% of mutant embryos to WT ( Figure 7E , H , I ) . Thus , ints6 is the defective gene in p18ahub mutant embryos . We also injected mutant embryos with up to 300 pg of different preparations of p18ahub mutant ints6 mRNA and never detected rescue ( Figure 7F ) . Thus , p18ahub is likely a strong loss-of-function or null allele . Overexpression of inst6 in WT embryos produced no phenotype ( data not shown ) . The related zebrafish ddx26b mRNA also rescued p18ahub embryos ( Figure 7G ) , although it is not provided maternally to the embryo ( data not shown ) . We did not detect ddx26b expression via in situ hybridization on embryos through 24 hpf , although we were able to amplify cDNA corresponding to ddx26b from ovary RNA ( not shown ) .
We have identified a recessive maternal-effect mutation in integrator complex subunit 6 ( ints6 ) , a highly conserved member of an RNA processing machine for which no specific role in vertebrate development was known . Importantly , to our knowledge the p18ahub mutation represents the first mutation of the ints6 gene in any organism to be reported . Interestingly , the loss of ints6 causes a recessive , maternal-effect dorsalization whereby dorsal midline axial fates are expanded , generating multiple dorsal axes at the expense of ventrolateral fates , suggesting an expansion of the dorsal organizer itself or of organizer gene expression mediating dorsal fate specification . The dramatic radial dorsalization of affected embryos is not caused by an expansion of the maternal Wnt-mediated induction of the organizer or the induction of Nodal signaling . In ints6 mutant embryos axial mesoderm is expanded along the DV axis at late blastula stages but remains confined near the margin where mesendoderm is normally induced , indicating a DV patterning defect rather than a general expansion of mesoderm . The Integrator Complex was identified as a complex of 12 subunits that co-purifies with Deleted in split hand/split foot 1 ( DSS1 ) in human HEK293 cells [36] . The Integrator Complex associates with the C-terminal domain ( CTD ) of the large subunit of RNA polymerase II ( RNAP ) and has been implicated in 3′ end processing of the U1 and U2 snRNAs of the splicesosome [36] . Phosphorylation of serine 7 in the heptapeptide repeats present in the CTD directs the Integrator Complex-bound RNAP to snRNA genes rather than to the promoters of protein coding genes [36] . Prior to those studies , ints6 was referred to as deleted in cancer 1 and investigated as a putative human tumor suppressor gene , given the loss of its expression in several tumor-derived cell lines and tumor specimens from patients [38] , [39] . It was shown that the promoter of ints6 is subject to hypermethylation in transformed cells and overexpression of ints6 can suppress the anchorage-independent growth of prostate tumor-derived and non-small cell lung carcinoma-derived cell lines [37] , [40] . Although Ints6 is a member of a complex required to produce functional spliceosomal snRNAs [36] , [80] , the dorsalized phenotype of the ints6 mutant in the zebrafish is not suggestive of a broad maternal or zygotic function in snRNA 3′ end formation . We can rescue the dorsalized mutant phenotype by altering a number of pathways required for embryonic patterning . Such success in our rescue experiments is difficult to reconcile with a widespread splicing defect . Since we can rescue dorsalization by injecting WT ints6 mRNA , Ints6 functions in the embryo rather than during oogenesis , when maternal mRNAs are transcribed and spliced . Tao et al . have reported that MO-mediated knockdown of Integrator Complex subunits 5 , 9 , and 11 in the zebrafish results in defective hematopoiesis , possibly due to alteration of zygotic snRNA and mRNA processing [81] . The lysis defect that we observe in some early gastrula stage p18ahub embryos may reflect a more general function for Ints6 , possibly in snRNA 3′ end formation or another unidentified role . In this regard p18ahub may be a hypomorphic allele that perturbs Ints6 function as part of the Integrator Complex . The specific role of Ints6 within the Integrator Complex is unknown . RNAi-mediated knockdown of Ints6 in Drosophila S2 cultured cells causes low levels of splicing defects of snRNAs and of a U7-GFP reporter . These effects are orders of magnitude less than the alteration of snRNA processing reported in Drosophila S2 cells for RNAi-mediated knockdown of Ints9 or Ints11 , putative catalytic subunits of the Integrator Complex [80] . It is also possible that the p18ahub missense mutation in ints6 affects some but not all roles of Ints6 equivalently . Prior to our work C . elegans deleted in cancer 1 ( dic-1 ) was the only ints6 homolog to be investigated in any developmental context [82] . However , C . elegans DIC-1 is highly divergent from zebrafish Ints6 displaying only 26% identity and may not represent a true ortholog . C . elegans DIC-1 is required to maintain the viability of oocytes through oogenesis and embryonic viability in general , since RNAi depletion in embryos leads to arrested development and widespread cell death . No defects in embryonic patterning were reported [82] . Nuclear β-catenin was normally localized dorsally in mid-blastula stage ints6 mutant embryos indicating that maternal Wnt-mediated organizer induction is normal . Furthermore , boz expression was also normal in mutant embryos through the equivalent of an early gastrula stage [7] , [11] . The expression of ndr1 , lft1 , and ntl in both WT and ints6 mutant embryos was initially confined to the dorsal side of the embryo and expanded normally ventrolaterally around the blastoderm margin by late blastula stages [62] , [63] , [64] , [83] . The earliest alteration of organizer gene expression that we observed in p18ahub embryos was the derepression of marginal gsc expression at an early gastrula stage . By mid gastrulation , ntl and flh expression in the axial mesoderm was also extensively expanded in mutant embryos . Several lines of evidence distinguish the induction of the Nodal pathway in the axial mesoderm from its induction throughout the margin during late blastula and early gastrula stages , but the molecular basis for this distinction is not fully understood [2] , [62] , [84] , [85] , [86] . Genes downstream of ndr1 , like ntl , are also induced differentially in the axis versus the ventrolateral margin [62] , [85] . Axial ntl is required for notochord formation , whereas non-axial ntl expression is required to establish ventroposterior mesoderm [87] , [88] . Analyses of reporter gene constructs driven by different ntl enhancers indicate that , aside from Nodal signaling , both Wnt8a and BMP signaling contribute significantly to the ventrolateral expression of ntl [89] . In ints6 mutant embryos , loss of the non-axial expression domain of ntl is likely due to reduced ventrolateral BMP and Wnt8a signaling , thus leading to loss of non-axial mesoderm with a concomitant expansion of axial mesoderm . Vox , Vent , and Ved are three critical repressors of organizer gene expression that are required to maintain the integrity of non-axial mesoderm . Initial induction of vox and zygotic ved expression is mediated through runx2bt2 , a maternally provided splice isoform of runx2 . Depletion of Runx2bt2 leads to a notable absence of ved expression at a late blastula stage , while vox and vent are expressed [17] . These defects are distinct from that of ints6 mutants , which display nearly absent vox and ved expression by early gastrulation . The Wnt8a pathway is a critical negative regulator of organizer gene expression in ventrolateral non-axial mesoderm operating through Vox , Vent , and Ved [7] , [10] , [18] . wnt8a is induced in a reduced domain in ints6 mutant embryos but is not maintained by an early gastrula stage ( Figure 4B ) . ints6 mutants exhibit reduced expression of vox and ved by the early gastrula stage , a time when these transcriptional repressors rely on Wnt8a for their expression . By the early gastrula stage , organizer gene derepression is evident in ints6 mutants ( Figure 3B ) . Loss of Wnt8a signaling may cause organizer gene expression to expand , or alternatively the expanded expression of organizer genes in ints6 mutants may cause the loss of Wnt8a signaling . Simultaneous loss of both zebrafish β-catenin proteins causes gsc and chd expression to expand around the embryonic margin , similarly to ints6 mutants [90] . β-catenin 1 and 2 function redundantly to transduce zygotic Wnt8a signaling , which has been proposed to suppress the ndr1-mediated induction of chd and gsc in the ventrolateral margin [91] . Loss of wnt8a expression in ints6 mutant embryos could similarly cause chd and gsc to expand around the ventrolateral margin . Our results suggest that Wnt8a signal transduction mechanisms are intact in ints6 mutants , given that chd MO-mediated rescue of p18ahub depends on Wnt8a signaling . Our results indicate that ints6 mutants are not mechanistically defective in BMP signaling . DV patterning of ints6 mutants can be restored to nearly WT by depleting the BMP antagonist Chd alone , or along with Noggin1 and Follistatin-like 2b . Induction of bmp2b expression is normal in ints6 mutants and loss of bmp2b and bmp4 expression by early gastrulation likely is the result of expanded organizer gene expression , including chd expression , or loss of Wnt8 signaling . Maternal-zygotic pou5f3 ( MZpou5f3 ) mutants , like ints6 mutants , display expanded organizer gene expression at late blastula stages and are severely dorsalized . In contrast to ints6 deficient embryos , MZpou5f3 mutants fail to initiate bmp2b and bmp4 expression , and show greatly reduced bmp7 expression [20] . In MZpou5f3 mutants , expansion of organizer gene expression is likely due to loss of repressors that rely on Pou5f3 for their expression [24] . Similarly to ints6 mutants , pou5f3 mutants exhibit aberrant morphogenesis and also display a delay in the completion of epiboly , [22] , [92] . However , Pou5f3 has additional functions during gastrulation compared to Ints6 . MZpou5f3 mutants fail to form endoderm and exhibit reduced sox17 expression in dorsal forerunner cells [20] , [24] , in contrast to the expanded populations of sox17-expressing forerunner cells observed in p18ahub embryos . Endodermal sox17 expression was observed in ints6 mutant embryos , although it was restricted more animally than in WT ( Figure 5L ) . We examined the expression of pou5f3 in p18ahub embryos at a late blastula stage and observed no differences from WT ( not shown ) . Thus , pou5f3 expression does not rely on ints6 . It is possible that ints6 and pou5f3 cooperate in DV patterning and in morphogenesis of the early embryo . The definitive placement of ints6 within a genetic pathway as well as the determination of its precise molecular functions will require biochemical characterization of the Ints6 protein . Ints6 may mediate DV patterning by supporting the expression or function of the Wnt8a pathway or specifically modulating the axial versus non-axial response to Nodal . Importantly , our molecular genetic approach has revealed a novel function for Ints6 in vertebrate embryonic patterning . Future studies will reveal its function within the complex gene regulatory network that restricts the organizer to dorsal regions and modulates axial versus non-axial tissues of the vertebrate body plan .
The animal work performed in this study was approved by the Institutional Review Board of the University of Pennsylvania . p18ahub was isolated in an ENU-induced mutagenesis genetic screen for maternal-effect mutations utilizing a hybrid AB/TU genetic background for mapping purposes , similar to that described previously [41] , [42] . The p18ahub mutation was found to be on the TU chromosome . Thus the line was outcrossed to AB to maintain it . Mutant females from both the original hybrid line as well as outcrossed lines were used . Embryos for in situ hybridization and injection experiments were derived from crosses of p18ahub mutant females to TL males . Due to delay in morphogenesis , the stage of p18ahub embryos is often based on the stage of corresponding age-matched WT reference embryos . Other criteria for embryo staging are described in the text . Antisense RNA probes were synthesized from linearized plasmid templates and transcribed using either SP6 or T7 polymerases ( Promega ) and Roche digoxygenenin-labeled NTP mix ( 11277073910 ) . To examine the expression of ints6 and ddx26b , full-length open reading frames were amplified using the following primers and cloned into pDONR221 and ultimately pCSDest using the Gateway system ( Invitrogen ) . Note: coding sequences or complements are shown; attB1 and attB2 sites are omitted . ints6 gateway For GTCCATGTAGAAGGGGCGAATATCAAC ints6 gateway Rev GTGCTCGAGTCCTTCAAGTAGGGCAG ddx26b gateway For GAGATAGGTCATATGCCGATTGTAGC ddx26b gateway Rev TGAGATGTGACTTGCCACTATCTGC The hybridization procedure was essentially as described at the Zebrafish Information Network ( see http://zfin . org/zf_info/zfbook/chapt9/9 . 82 . html ) , except Roche BM Purple alkaline phosphatase substrate ( 11442074001 ) was used and terminated by washing embryos briefly in PBS followed by overnight fixation in 4% paraformaldehyde/PBS at 4°C . Stained embryos were stored in methanol at −20°C and cleared in a 2∶1 mixture of benzyl benzoate∶benzyl alcohol for photography . Embryos were mounted in Canada Balsam and oriented under glass coverslips for imaging on either a Zeiss Axioscope microscope fitted with a ProgRes CF CCD camera ( Jenoptik ) or Leica MZ12-5 dissecting microscope fitted with a Photometrics CoolSnap CF CCD camera ( Roper Scientific ) . In situ hybridization was performed 2 to 4 times for each probe independently on groups of mutant and age- or stage-matched WT control embryos , as indicated in the text , typically obtained from crosses on the same day . Embryos were fixed overnight at 4°C in 4% paraformaldehye/PBS and washed several times in PBS . Embryos were then dechorionated and placed in 100% methanol overnight or for long-term storage at −20°C . Embryos were rehydrated first in 50% MeOH for 10 minutes and then 3 times for 10 minutes in100% PBST ( PBS with 0 . 5% Triton X-100 ) . Embryos were blocked in Blocking Solution ( 10% fetal bovine serum/PBST ) for one hour before receiving a 1∶500 dilution of rabbit anti-β-catenin antibody ( Sigma C2206 ) in Blocking Solution and incubating at 4°C overnight . The next day embryos were washed 3 times for 10 minutes with 1 ml of Blocking Solution at room temperature before incubation in a 1∶500 dilution of HRP-conjugated goat-anti-rabbit antibody ( Sigma A9169 ) in Blocking Solution for at least two hours at room temperature . Embryos were then washed 4 times in 1 ml PBST and developed in 1 ml of a 1∶3 dilution of a 15 ml solution of 50 mM Tris-Cl pH 7 . 5 , 100 mM NaCl , 1 DAB ( 3 , 3′ diaminobenzidine ) tablet ( Sigma D5905 ) , and 15 µl freshly added 30% H2O2 . Staining was monitored visually under a dissecting microscope and terminated via washing in 50 mM Tris-Cl pH 7 . 5 , 100 mM NaCl . Stained embryos were placed in 100% methanol for storage at −20°C and photographed as described above for in situ hybridized embryos . mRNA or morpholinos ( MOs ) were injected into the yolk cell of 1- to 2-cell stage embryos that were subsequently scored at 24 hours post fertilization ( hpf ) for phenotypes . In vitro transcribed mRNAs were generated from linearized plasmid DNA templates using the mMessage mMachine kit ( Ambion ) . mRNAs were diluted to appropriate working concentrations in 100 mM KCl/10% Phenol Red just prior to injection . Templates for ints6 ( WT or mutant ) and ddx26b mRNAs were the same as those used to make in situ riboprobes described above . Other templates included: lft1-pCS2 [64] , sqt-pCS2 [93] , and FLAG-bmp2b [94] . MOs were obtained from GeneTools and prepared as stocks in water according to the manufacturers recommendations . Prior to injection , MOs were diluted to working concentrations in 1× Danieau's Solution ( 58 mM NaCl , 0 . 7 mM KCl , 0 . 4 mM MgSO4 , 0 . 6 mM Ca ( NO3 ) 2 , 5 mM Hepes pH 7 . 1–7 . 3 ) . The sequences of the chordin , noggin1 , and follistatin-like2b translation-blocking MOs are reported elsewhere [54] . We employed anti-chd MO at 0 . 82 ng/µl , anti-nog MO at 4 . 22 ng/µl , and anti-fstl2b MO at 8 . 45 ng/µl and typically injected 1 nl . To target processing of wnt8a transcripts we employed the splice-blocking MOs orf1 E1i1 and orf2 E4i4 , as described [60] . Our fine mapping strategy has been described elsewhere [41] , [95] . Heterozygous p18ahub females and homozygous p18ahub males from the original AB/TU hybrid strain were crossed to obtain females recombinant between SSLP markers flanking the mutation . The initial interval placed p18ahub between z34824 and z6845 . In total we examined 1148 meioses , narrowing the interval to approximately 1 . 35 Mb between z7120 and an SSLP marker that we generated against BAC CR545476 ( forward primer – AGATGTAACTCATCCACTGTCATACACC , reverse primer – AACCGTTGAGAGGTTTCTAGCTAGTAC ) . We then proceeded to sequence all genes within the interval for which a cDNA was reported . We made oligo-dT-primed cDNA from the TU wild-type and the p18ahub mutant strains from total ovary RNA extracted using Trizol Reagent ( Invitrogen 15596-018 ) according to the manufacturer's instructions , and PCR amplified the coding regions of candidate genes . PCR products were purified ( Qiagen PCR clean-up kit ) and both strands were sequenced . The ints6 ORF was sequenced using the primers described above ( without att sites ) ( GenBank Accession number KF700696 ) and a T to A transition at nucleotide 1124 converting valine 375 to an aspartate was discovered in p18ahub-derived ovary cDNA . No mutation was found in genomic DNA derived from the G0 mutagenized male , consistent with the point mutation resulting from ENU-induced mutagenesis . Embryos were mounted in 0 . 3% low-melt agarose in E3 media and photographed with a Leica MZ12 . 5 microscope and Micropublisher 5 . 0 RTV Non-Cooled camera at 5 minute intervals from mid blastula through early somitogenesis stages at room temperature under constant illumination . Images were combined into movies using ImageJ .
|
A complex integration of signaling pathways establishes the body plan of the vertebrate embryo . The dorsal side of the embryo is defined by the organizer , a specialized field of cells that breaks the symmetry of the zebrafish blastula by instructing neighboring cells to adopt dorsal fates based on their proximity . The isolation of mutant genes in the zebrafish has identified many genes required for embryonic development . However , our knowledge of the molecular mechanisms integrating different signaling pathways within a gene regulatory network to properly pattern the embryo is still incomplete . We isolated a recessive maternal effect mutation in the integrator complex subunit 6 ( ints6 ) gene that leads to a circumferential expansion of the organizer and the formation of dorsal tissues at the expense of ventral tissues . Currently , the only reported role for the Integrator Complex is to mediate processing of snRNAs of the spliceosome . Our molecular genetic approach indicates that ints6 confines the organizer to dorsal domains , preventing it from extending around the margin into ventral domains . Thus , we have determined a novel role for a highly conserved component of an RNA processing machine .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
The Integrator Complex Subunit 6 (Ints6) Confines the Dorsal Organizer in Vertebrate Embryogenesis
|
Oral infection by Trypanosoma cruzi has been the primary cause of recent outbreaks of acute Chagas' diseases . This route of infection may involve selective binding of the metacyclic trypomastigote surface molecule gp82 to gastric mucin as a first step towards invasion of the gastric mucosal epithelium and subsequent systemic infection . Here we addressed that question by performing in vitro and in vivo experiments . A recombinant protein containing the complete gp82 sequence ( J18 ) , a construct lacking the gp82 central domain ( J18* ) , and 20-mer synthetic peptides based on the gp82 central domain , were used for gastric mucin binding and HeLa cell invasion assays , or for in vivo experiments . Metacyclic trypomastigotes and J18 bound to gastric mucin whereas J18* failed to bind . Parasite or J18 binding to submaxillary mucin was negligible . HeLa cell invasion by metacyclic forms was not affected by gastric mucin but was inhibited in the presence of submaxillary mucin . Of peptides tested for inhibition of J18 binding to gastric mucin , the inhibitory peptide p7 markedly reduced parasite invasion of HeLa cells in the presence of gastric mucin . Peptide p7* , with the same composition as p7 but with a scrambled sequence , had no effect . Mice fed with peptide p7 before oral infection with metacyclic forms developed lower parasitemias than mice fed with peptide p7* . Our results indicate that selective binding of gp82 to gastric mucin may direct T . cruzi metacyclic trypomastigotes to stomach mucosal epithelium in oral infection .
Orally transmitted infection by the protozoan parasite Trypanosoma cruzi has been responsible for frequent outbreaks of acute cases of Chagas' disease in recent years [1] , [2] . In Brazil , after the elimination of the domiciliary vector Triatoma infestans in many endemic areas , and the control of the blood bank transmission , T . cruzi infection by the oral route constitutes the most important transmission mechanism [2] . The occurrence of Chagas' disease through food contamination , involving triatomine insects other than T . infestans , has been reported in different geographical regions . The incidence is higher in the Brazilian Amazon [1] , where more than half of acute cases of the disease reported in the last 40 years can be attributed to microepidemics of orally transmitted infection [2] , [3] . Studies on oral T . cruzi infection in the mouse model have shown that the insect stage metacyclic trypomastigotes invade the gastric mucosal epithelium and , following intracellular replication as amastigotes , differentiate into trypomastigotes that are subsequently released into circulation [4] , [5] . During oral infection , gastric mucosa is uniquely targeted for metacyclic trypomastigote entry in order to establish a systemic T . cruzi infection , with parasites being undetectable elsewhere within the mucosa of the oropharynx or esophagus [4] . There are several evidences that the metacyclic stage-specific surface glycoprotein gp82 plays a critical role in the establishment of T . cruzi infection by the oral route [6] , [7] . Gp82 is a cell adhesion molecule that mediates metacyclic trypomastigote entry into cultured human epithelial cells , by triggering the signal transduction pathways leading to cytosolic Ca2+ mobilization in both cells [8] , an event essential for parasite internalization [9] , [10] , [11] . In addition to cell invasion-promoting properties , gp82 has the ability to bind to gastric mucin [6] . Through gp82-mediated interaction with gastric mucin , a constituent of the luminal barrier that functions as a first line of defense against invading pathogens , the parasites may effectively be addressed to the target cells . Metacyclic forms of T . cruzi strains deficient in gp82 expression are poorly infective when administered orally into mice , although they efficiently invade host cells in vitro by engaging gp30 , a Ca2+ signal-inducing surface molecule related to gp82 but devoid of gastric mucin-binding property [7] . Unlike gp82-expressing strains , the gp82-deficient strains have reduced capacity to enter cultured epithelial cells in the presence of gastric mucin [7] . This reinforces gp82 binding to gastric mucin as an important requirement for parasites reaching the underlying target cells . Selective binding of gp82 to gastric mucin could explain why parasite invasion is not found anywhere within the oropharynx or esophagus [4] . Shigella dysenteriae , for instance , whose pathogenic potential correlates with its capacity to invade and multiply within cells of the colonic epithelium , adheres preferentially to colonic mucin [12] . Although purified gp82 , either in its native form or as a recombinant protein , binds to gastric mucin , it remains to be demonstrated that T . cruzi metacyclic forms bind selectively to gastric mucin in gp82-dependent manner . Here we aimed at addressing that question , at identifying the gp82 sequences involved in gastric mucin-binding , and at investigating the effect of gp82-based synthetic peptides on metacyclic trypomastigote infection in vitro and on oral infection in mice .
T . cruzi strain CL [13] was used throughout . Parasites were maintained cyclically in mice and in liver infusion tryptose medium . Metacyclic trypomastigotes , generated in Grace's medium , were purified by passage through DEAE-cellulose column , as described [14] . HeLa cells , the human carcinoma-derived epithelial cells , were grown at 37°C in Dulbecco's Minimum Essential Medium , supplemented with 10% fetal calf serum , streptomycin ( 100 µg/ml ) and penicillin ( 100 U/ml ) in a humidified 5% CO2 atmosphere . Cell invasion assays were carried out as detailed elsewhere [15] , by seeding the parasites onto each well of 24-well plates containing 13-mm diameter round glass coverslips coated with 1 . 5×105 HeLa cells . After 1 h incubation with parasites , at multiplicity of infection 10∶1 , the duplicate or triplicate coverslips were fixed in Bouin solution , stained with Giemsa , and sequentially dehydrated in acetone , a graded series of acetone∶xylol ( 9∶1 , 7∶3 , 3∶7 ) and xylol . Cell invasion assays in the presence of gastric mucin were performed with mucin suspended in culture medium . The recombinant protein J18 , containing the full-length T . cruzi gp82 ( GenBank accession number L14824 ) in frame with gluthatione S-transferase ( GST ) , was produced in E . coli DH5-α by transforming the bacteria with a pGEX-3 construct comprising the gp82 gene [16] . All steps for induction of the recombinant protein J18 and its purification are detailed elsewhere [17] . J18b , a construct containing the carboxy-terminal half of gp82 , and J18* , with deletion of 65 amino acids ( residues 257 to 321 ) , were prepared as previously described [18] , [19] and purified in the same manner as J18 . In addition , from a cloned full-length gp82 cDNA named C03 ( GenBank accession number EF445668 ) we generated the recombinant protein C03 containing histidine tail , as previously described [17] . The amount of purified protein was quantified by reaction with Coomassie Plus ( Pierce ) in 96 well plates , and reading at 620 nm . To certify that the desired protein was obtained , the purified samples were analyzed in SDS-PAGE gel stained with Coomassie Blue , and by immunoblotting using anti-GST antibodies . Microtiter plates ( 96 wells ) were coated with mucin from porcine stomach ( Type III , Sigma ) or from bovine submaxillary glands ( Type I-S , Sigma ) in PBS ( 10 µg/well ) . After blocking with PBS containing 2 mg/ml bovine serum albumin ( PBS/BSA ) for 1 h at 37°C , the plates were sequentially incubated at 37°C for 1 h with the recombinant protein J18 , J18* or C03 , with polyclonal monospecific antibody directed to J18 , C03 or GST , and with peroxidase-conjugated anti-mouse IgG , all diluted in PBS/BSA . The final reaction was revealed by o-phenilenediamine and the absorbance at 492 nm read in a Multiscan MS ELISA reader . To check the effective coating with mucin , the mucin-coated microtiter plates ( 10 µg/well ) were blocked with PBS/BSA and incubated at 37°C for 1 h with anti-gastric mucin or anti-submaxillary mucin antisera diluted 1∶100 in PBS/BSA . Reaction proceeded with peroxidase-conjugated anti-mouse IgG , as described above . For coating with gastric mucin preparation at pH 2 . 5 citrate buffer was used . Wells of microtiter plates were coated with gastric mucin ( 10 µg/well ) . After blocking with PBS/BSA , the plates were incubated for 1 h at 37°C with J18 ( 10 µg/ml ) in absence or in the presence of individual synthetic peptides ( 200 µg/ml ) corresponding to the gp82 sequence spanning residues 224–333 , synthesized as described [19] . Following incubation with anti-GST antibodies and peroxidase-conjugated anti-mouse IgG , the reaction was revealed by o-phenilenediamine . Twenty four- well plates , containing 13 mm round glass coverslips , were incubated overnight at 37°C with 100 µl of gastric mucin or submaxillary mucin , at 400 µg/ml in PBS . After washings in PBS , a metacyclic trypomastigote suspension in cell culture medium ( 5×107/ml ) was added to each well and incubation proceeded for 1 h at 37°C . Following washes with PBS , the coverslips were fixed in methanol and Giemsa-stained for microscopic visualization of parasites . ELISA assay for metacyclic trypomastigote binding to mammalian mucin was performed as follows: 96-well microtiter plates coated with gastric mucin or submaxillary mucin , at varying concentrations in PBS , were washed in PBS and then 50 µl of parasite suspension in cell culture medium ( 5×107/ml ) were added . Following 1 h incubation at 37°C and washings in PBS , the parasites were fixed with 3 . 5% formaldehyde for 20 min at room temperature . After washings in PBS , the parasites were sequentially incubated with a monoclonal antibody to T . cruzi surface glycoprotein gp35/50 , and peroxidase-conjugated anti-mouse IgG , all of them diluted in PBS/BSA . The final reaction was revealed by o-phenilenediamine . The same protocol was used for inhibition of parasite binding to gastric mucin , in which microtiter plates coated with gastric mucin ( 10 µg/well ) were incubated with metacyclic forms in absence or in the presence of varying concentrations of the recombinant protein J18 or GST . Polycarbonate transwell filters ( 3 µm pores , 6 . 5 mm diameter , Costar ) were coated with 50 µl of a preparation containing 10 mg/ml gastric mucin or submaxillary mucin in water . T . cruzi metacyclic trypomastigotes , in 600 µl PBS were added to the bottom of 24-well plates ( 7 . 5×107 parasites/well ) and incubated for 1 h at 37°C . Thereafter , the mucin-coated transwell filters were placed onto parasite-containing wells , and 100 µl PBS were added to the filter chamber . At different time points of incubation at 37°C , 10 µl were collected from the filter chamber for determination of parasite number and the volume in this chamber was corrected by adding 10 µl PBS . Four to five week-old female Balb/c mice , bred in the animal facility at Universidade Federal de São Paulo , were used . All procedures and experiments conformed with the regulation of the institutional Ethical Committee for animal experimentation , and the study was approved by the Committee . Mice were infected with T . cruzi metacyclic forms by oral route ( 4×105 parasites per mouse ) , using a plastic tube adapted to a 1 ml syringe . Starting on day 10 post-inoculation , parasitemia was monitored twice a week by examining 5 µl blood samples collected from the tail , at the phase contrast microscope . For detection of parasites in the gastric mucosal epithelium , the stomach of mice inoculated orally with metacyclic forms was collected , fixed with 10% neutral formaldehyde for 24 h . After processing by gradual dehydration in a graded series of ethanol solution , followed by xylene immersion and embedding in parafilm , serial 5 µm tissue sections were cut and stained with hematoxylin and eosin . Six to eight week-old Balb/c mice were used for immunization with J18 , C03 , GST , gastric mucin or submaxillary mucin to generate specific antibodies . Mice received the first dose of antigen ( 10 µg/mouse ) adsorbed in Al ( OH ) 3 ( 0 . 5 mg/mouse ) and after two weeks received three additional doses of the antigen plus the same adjuvant at one week interval . Ten days after the last immunizing dose , the mice were bled by heart puncture and the serum collected . To determine significance of data by Student's t test , the program GraphPad InStat was used .
To address the question why , upon oral inoculation into mice , T . cruzi metacyclic forms invade the gastric mucosa but not the mucosa of the oropharynx [4] , we examined the possibility that metacyclic forms bind to gastric mucin but not to submaxillary mucin . Assays were performed by incubating microtiter plates coated with varying amounts of gastric or submaxillary mucin with parasites at 37°C for 1 h , followed by fixation . Reaction with a monoclonal antibody to T . cruzi surface glycoprotein gp35/50 revealed that metacyclic trypomastigotes bound to gastric mucin in a dose-dependent manner whereas binding to submaxillary mucin was negligible at all concentrations tested ( Fig . 1A ) . Effective coating with mucins was ascertained by reaction with antibodies specific for gastric or submaxillary mucin ( Fig . 1B ) . To determine whether metacyclic surface molecule gp82 was implicated in binding to gastric mucin , assays were also performed in the presence of a recombinant protein containing the full length gp82 sequence fused to GST ( J18 ) . Metacyclic trypomastigote binding to gastric mucin was inhibited in a dose-dependent manner by J18 but not by GST ( Fig . 1C ) , consistent with a role for gp82 . Parasites bound to gastric mucin were visualized by microscopic examination in parallel assays using coverslips coated with gastric mucin ( Fig . 1D ) , as described in the methods section . In addition , the ability of metacyclic forms to traverse a mucin layer was examined in microtiter plates with transwell filters coated with gastric or submaxillary mucin , by counting the number of parasites that translocated through the mucin layer at varying time points . As shown in Fig . 1E , the number of parasites that traversed the gastric mucin layer was significantly higher than the number of parasites that traversed the submaxillary mucin layer . Next , we determined the ability of metacyclic forms to enter host cells in the presence of gastric or submaxillary mucin , in an attempt to mimic the in vivo situation in which the parasites interact with and traverse the mucous layer before reaching the underlying epithelial cells . Mucin , at 2 mg/ml in culture medium , was added to HeLa cells 15 min before addition of parasites . Following 1 h incubations with parasites , the cells were fixed , stained , and the number of intracellular parasites was counted . The rate of metacyclic trypomastigote internalization was comparable in the absence or in the presence of gastric mucin , but was drastically reduced in the presence of submaxillary mucin ( Fig . 1F ) . Even at 20 mg/ml , gastric mucin did not exhibit any inhibitory effect on parasite entry into HeLa cells . To ascertain that the metacyclic trypomastigote surface molecule gp82 bound preferentially to gastric mucin when compared to submaxillary mucin , ELISA assays were carried out using J18 , the recombinant protein containing the full-length gp82 sequence , previously shown to bind to gastric mucin in the same manner as the native molecule [6] . J18 bound to gastric mucin in a dose-dependent and saturable manner whereas binding to submaxillary mucin was minimal ( Fig . 2A ) . As gp82 is encoded by a multigene family [20] , we asked whether another member of the family also possessed gastric mucin-binding property . When compared to J18 , the gastric mucin-binding ability of the recombinant protein C03 , which shares 59 . 1% sequence identity with J18 [17] was negligible ( Fig . 2B ) . Other metacyclic surface molecules , such as gp90 and gp35/50 , were devoid of gastric mucin binding capacity ( data not shown ) . A question that was also examined concerned the effect of pH on J18 binding to gastric mucin . It has recently been reported that pH affects the association behavior of pig gastric mucin in aqueous solutions , with large interchain aggregates being detectable at pH 2 by dynamic light scattering [21] . Gastric mucin preparations at pH 7 . 2 and at pH 2 . 5 , which is close to the pH of the gastric milieu , were used to coat microtiter plates for J18 binding assays . As shown in Fig . 2C , J18 bound efficiently to gastric mucin regardless of the pH . To determine the gastric mucin-binding domain of the gp82 molecule , two other recombinant proteins were used: J18b , containing the carboxy-terminal half of gp82 and J18* , lacking 65 residues ( amino acids 257 to 321 ) corresponding to the gp82 domain required for host cell adhesion ( Fig . 2D ) . J18* did not bind to gastric mucin , in contrast to J18 and J18b ( Fig . 2E ) . To determine more precisely the gp82 sequence involved in binding to gastric mucin , we performed assays of competition between the recombinant protein J18 and synthetic peptides spanning the gp82 central domain . Gastric mucin-coated microtiter plates were incubated with J18 ( 10 µg/ml ) in absence or in the presence of each of the 20-mer peptides with 10 overlapping residues , spanning amino acids 224–333 ( Fig . 3A ) , and the reaction proceeded as in conventional binding assays . Peptides p7 and p10 inhibited J18 binding to gastric mucin by >70% , inhibition by p5 was on the order of ∼46% and by p1<20% , whereas other peptides showed no effect ( Fig . 3B ) . We tested whether the synthetic peptides p7 and p10 , which displayed the highest inhibitory effect on J18 binding to gastric mucin ( Fig . 3B ) could interfere with target cell invasion by metacyclic forms . In absence of gastric mucin , neither of these peptides affected the parasite internalization , but in the presence of gastric mucin peptide p7 , and to a lesser degree p10 , had a significant inhibitory activity while peptides p6 and p8 used as controls had no effect ( Fig . 4A ) . The inhibitory effect of p7 was dose-dependent ( Fig . 4B ) . To ascertain the sequence-specificity of that effect , a peptide with the same amino acid composition of p7 but with a scrambled sequence ( LADLAGWLSPSDVGGAINST ) , designated p7* , was also tested in cell invasion assays . As shown in Fig . 4C , peptide p7* was devoid of inhibitory activity . To examine whether peptide p7 affected the in vivo metacyclic trypomastigote infectivity , Balb/c mice were administered with peptide p7 or p7* ( 20 µg/mouse ) 15 min before oral infection with metacyclic forms ( 4×105 parasites/mouse ) , and the parasitemia levels were monitored . Mice that received peptide p7 developed significantly lower parasitemias than those that received the control peptide p7* ( Fig . 5A ) . To determine whether the difference in the parasitemia levels between the two groups resulted from differential invasion of gastric mucosal epithelium , the stomach of some mice was collected 4 days post-inoculation and processed for histological preparations . As shown in Fig . 5B , the number of parasites replicating in the gastric mucosa , visualized as amastigote nests in the stomach sections , was significantly lower in mice that received peptide p7 as compared to control mice that were given peptide p7* before parasite inoculation .
We had previously suggested that upon oral T . cruzi infection the metacyclic trypomastigote-specific surface molecule gp82 could play a key role in directing the parasites to the gastric mucosa , by promoting adhesion to mucin molecules that are the main constituent of the mucous layer , which would be traversed driven by ATP [22] . The results herein described provide support to that hypothesis . Metacyclic trypomastigotes bound to gastric mucin in vitro and were capable of efficiently traversing the gastric mucin layer to invade the underlying epithelial cells . By contrast , the binding to mucin from submaxillary glands was minimal and the parasite ability to translocate through the submaxillary mucin layer , as well as the infectivity towards host cells in the presence of submaxillary mucin , were reduced . These results , in addition to the observations that metacyclic forms do not invade cells of the oropharynx after oral inoculation into mice [4] , indicate that binding of metacyclic forms to gastric mucin is an important determinant in addressing the parasites to gastric mucosal epithelial cells . Metacyclic trypomastigotes of T . cruzi strains that bind poorly to gastric mucin have poor capacity to invade epithelial cells in vitro in the presence of gastric mucin and in infecting mouse by the oral route [7] . Selective binding to mucin molecules , as a prerequisite for the establishment of infection by other pathogenic microorganisms of the gastrointestinal tract have been reported . Helicobacter pylori , which colonizes gastric mucosa , binds to human gastric mucin [23] , Shigella , which invades and multiplies within cells of the colonic epithelium , binds specifically to human colonic mucin but not to small intestine mucin [24] . Existing evidence suggests that binding of T . cruzi metacyclic trypomastigotes to gastric mucin is mediated by the surface molecule gp82 . Binding of metacyclic forms to gastric mucin was inhibited by a recombinant protein based on gp82 in a dose-dependent manner . This finding , plus the previous observations that invasion of cultured epithelial cells by metacyclic forms of T . cruzi strains deficient in gp82 expression was reduced in the presence of gastric mucin [7] reinforce the idea that the gastric mucin-binding property of gp82 plays a key role in infection . Consistent with the in vitro findings , gp82-deficient metacyclic forms were poorly infective when administered orally into mice [7] . However , gp82 may not be the sole surface molecule that interacts with gastric mucin . We cannot rule out the possibility that other as yet unidentified metacyclic trypomastigote surface molecules bind to gastric mucin , albeit to a lesser extent . The gastric mucin-binding domain of gp82 was found to be localized in the central domain of the molecule , and the peptide sequence PTAGLVGLLSNSASGDAWID was determined as being the most important for invasion of epithelial cells under the gastric mucin coat . In the presence of the synthetic peptide p7 , based on that sequence , not only was the gastric mucin-binding of the recombinant gp82 inhibited but also the metacyclic trypomastigote invasion of epithelial cells coated with gastric mucin was reduced . Although displaying a similar inhibitory effect as p7 on gastric mucin-binding of the recombinant gp82 , the peptide p10 ( NAVKVHDGFKFTGFGSGAIW ) affected parasite entry into target cells to a lesser degree than p7 . One interesting possibility is that metacyclic trypomastigote binding to gastric mucin through the p7 defined region of gp82 facilitates host cell adhesion , but that adhesion is also dependent upon co-operation from other regions of gp82 . The main cell adhesion site of gp82 is the sequence LARLTEELKTIKSVLSTWSK represented by peptide p4 [19] , from which the p7 sequence is separated by 5 residues . Of note is that the isoelectric point of p4 is 9 . 71 and that of p7 is 3 . 23 . The relevance , if any , of such a difference for target cell attachment is not known . In a previous study , a recombinant construct corresponding to the complete gp82 sequence , but without the two glutamic acid residues contained in p4 sequence , had reduced capacity to bind to HeLa cells [19] . Our speculative view is that metacyclic trypomastigotes bound to gastric mucin through the gp82 sequence p7 would have the p4 sequence-mediated interaction with host cells reinforced . If this is the case , the effective gp82-medianted parasite entry into target cells could be more than the result of independent contributions of diverse gp82 functional sites , p7 sequence acting in the early step of gastric mucin-binding and p4 sequence in the subsequent step of cell attachment . It should be pointed out that the central domain of metacyclic trypomastigote gp82 , where the p7 sequence resides , shares considerable sequence similarity with other glycoproteins of gp85 family and trans-sialidase , which are all members of the same superfamily . Analysis using BLASTP to search for sequences homologous to peptide p7 revealed proteins of gp85/trans-sialidase superfamily; therefore , it is possible that a motif similar to that represented by p7 may also be implicated in gastric mucin-binding and be responsible for the low proportion of cases in which cell invasion of metacyclic forms is not inhibited by p7 in the presence of gastric mucin ( Fig . 4C ) . Another interesting possibility , unrelated to p7 sequence , is that metacyclic trypomastigote trans-sialidase binds to gastric mucin sialic residues , an interaction that is not inhibitable by peptide p7 . The importance of gp82-mediated binding to gastric mucin through the p7 sequence in the establishment of T . cruzi infection by the oral route was tested in the mouse model . Mice administered orally with peptide p7 prior to metacyclic trypomastigotes developed low parasitemia levels , in contrast to animals that received the control peptide p7* , with the same composition as p7 but with a scrambled sequence , which developed high parasitemias . The presence of much fewer nests of amastigotes replicating in the stomach epithelium in mice that were given p7 , as compared to the number of amastigote nests in mice that received p7* , is an indication that fewer parasites invaded the gastric mucosal epithelium . What possibly occurs is that the presence of p7 interferes with gp82-binding to gastric mucin , precluding the parasite traversal of the mucus layer towards the target cells . Therefore , the gastric mucin-binding property of gp82 may be as critical for T . cruzi infection by the oral route as its cell-binding capacity .
|
Frequent outbreaks of acute Chagas' disease by food contamination with T . cruzi , characterized by high mortality , have been reported in recent years . In Brazil , oral infection is currently the most important mechanism of T . cruzi transmission . Studies on oral T . cruzi infection in mice have shown that insect-stage metacyclic trypomastigotes invade only the gastric mucosal epithelium and not other areas of mucosal epithelia prior to establishing systemic infection . Here we have shown that metacyclic trypomastigotes bind selectively to gastric mucin , a property also displayed by gp82 , a metacyclic stage-specific surface protein implicated in cell adhesion/invasion process . It is also shown that the gastric mucin-binding property of gp82 resides in the central domain of the molecule and that the synthetic peptide p7 , based on a gastric mucin-binding sequence of gp82 , markedly reduces parasite invasion of cultured human epithelial cells in the presence of gastric mucin . These results , plus the finding that mice that received peptide p7 before oral infection with metacyclic trypomastigotes had fewer parasites replicating in the gastric mucosa and developed lower parasitemias than control mice , lead us to suggest that gp82-mediated interaction with gastric mucin may direct T . cruzi to stomach mucosal epithelium in oral infection .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/neglected",
"tropical",
"diseases"
] |
2010
|
Role of GP82 in the Selective Binding to Gastric Mucin during Oral Infection with Trypanosoma cruzi
|
Kaposi’s sarcoma-associated herpesvirus ( KSHV; human herpesvirus 8 ) belongs to the subfamily of Gammaherpesvirinae and is the etiological agent of Kaposi’s sarcoma as well as of two lymphoproliferative diseases: primary effusion lymphoma and multicentric Castleman disease . The KSHV life cycle is divided into a latent and a lytic phase and is highly regulated by viral immunomodulatory proteins which control the host antiviral immune response . Among them is a group of proteins with homology to cellular interferon regulatory factors , the viral interferon regulatory factors 1–4 . The KSHV vIRFs are known as inhibitors of cellular interferon signaling and are involved in different oncogenic pathways . Here we characterized the role of the second vIRF protein , vIRF2 , during the KSHV life cycle . We found the vIRF2 protein to be expressed in different KSHV positive cells with early lytic kinetics . Importantly , we observed that vIRF2 suppresses the expression of viral early lytic genes in both newly infected and reactivated persistently infected endothelial cells . This vIRF2-dependent regulation of the KSHV life cycle might involve the increased expression of cellular interferon-induced genes such as the IFIT proteins 1 , 2 and 3 , which antagonize the expression of early KSHV lytic proteins . Our findings suggest a model in which the viral protein vIRF2 allows KSHV to harness an IFN-dependent pathway to regulate KSHV early gene expression .
Kaposi’s sarcoma-associated herpesvirus ( KSHV ) or human herpesvirus 8 ( HHV8 ) belongs to the genus Rhadinovirus within the subfamily of Gammaherpesvirinae . It was identified in 1994 when herpesvirus-like DNA sequences were discovered in AIDS-associated Kaposi's sarcoma ( KS ) tissues [1] . Apart from KS , KSHV is also the cause of two lymphoproliferative diseases , primary effusion lymphoma ( PEL ) and multicentric Castleman’s disease ( MCD ) [2 , 3] . The KSHV life cycle is divided into two phases: latency and productive ( ‘lytic’ ) replication . Early after infection , the viral dsDNA enters the nucleus , is circularized and further chromatinized and maintained as a stable episome within the host [4] . During latency no viral particles are produced and only a few latent proteins are expressed from the so called latency transcript cluster under the control of a constitutively active promoter . These latent proteins function as viral regulators enabling the establishment and maintenance of latency as well as the inhibition of the lytic cycle . Furthermore , they are involved in cell proliferation , survival , differentiation and transformation as well as angiogenesis and the induction of interferon stimulated genes ( ISGs ) and thereby contribute to KSHV pathogenesis [4–7] . The latent state can be disrupted by lytic reactivation which is characterized by a distinct pattern of gene expression , involving immediate early , early lytic and late lytic transcripts [8] . Lytic cycle activation can be induced by co-factors like hypoxia , oxidative stress and inflammatory cytokines [8–15] . In addition , viral co-infections with human immunodeficiency virus-1 ( HIV-1 ) , Herpes simplex virus ( HSV ) or Human cytomegalovirus ( HCMV ) are known to induce KSHV reactivation [13 , 16–19] . Several chemical compounds , such as histone deacetylase ( HDAC ) inhibitors like sodium butyrate ( SB ) , or 12-O-tetradecanoylphorbol-13-acetate ( TPA ) can activate the lytic replication cycle [20 , 21] . The immediate early protein , replication and transcription activator ( RTA ) encoded by ORF50 , is a key viral regulator of the KSHV lytic cycle . By trans-activating the expression of other lytic downstream genes as well as its own promoter , RTA is necessary and sufficient to induce the entire viral lytic cycle [22 , 23] . Several viral proteins expressed during the early stages of the lytic replication cycle such as vIL6 , vGPCR , K1 and K15 contribute to KSHV pathogenesis by promoting proliferation , angiogenesis , invasiveness and may counteract the host antiviral immune response [24 , 25] . The KSHV genome also contains four genes with partial sequence homology to cellular interferon regulatory factors ( IRFs ) : an unspliced mRNA transcribed from the viral open reading frame ( ORF ) K9 encodes viral interferon regulatory factor 1 ( vIRF1 ) , while spliced mRNAs generated from ORF K10/K10 . 1 , ORF K10 . 5/10 . 6 and ORF K11/K11 . 1 yield vIRF 4 , 3 and 2 , respectively [26–29] . Apart from KSHV , the only other viruses known for now to encode similar viral IRFs are the rhadinovirus of rhesus macaques ( RRV ) and the retroperitoneal fibromatosis herpesvirus of pig-tailed macaques ( RFHVMn ) [30–32] . In spite of the low degree of protein sequence homology with cellular IRFs and in particular the only partial conservation of five characteristic tryptophan residues found in the N-terminal DNA binding domain ( DBD ) of cellular IRFs , at least vIRF1 and vIRF2 have been shown to bind to viral or cellular DNA in a similar manner as cellular IRFs [33–38] . Most studies on the function of individual KSHV vIRFs suggest that they counteract the cellular interferon ( IFN ) response and inhibit different proliferative , apoptotic and angiogenetic pathways by interacting with and functionally modulating cellular proteins [39] . So far , three vIRFs have been implicated in the regulation of viral lytic replication . Thus , vIRF1 has been shown to promote , and vIRF3 to suppress , KSHV lytic replication by recruitment of USP7 [40] . Interestingly , vIRF3 is expressed as a latent protein only in PEL cells and in MCD tissue , it is important for PEL cell survival and its expression is unchanged after lytic reactivation [41] . Furthermore , vIRF4 has been shown to support lytic reactivation by virtue of its interaction with the Notch downstream effector CBL/CSF1 [42] and to play a crucial role in triggering the switch from KSHV latency to lytic cycle through interfering with the BCL6-vIRF4 axis [43] . In addition to its interaction with USP7 , vIRF3 has been shown to promote the degradation of Promyelocytic leukemia nuclear bodies ( PML NBs ) /nuclear domain 10 ( ND10 ) [44 , 45] . PML NBs are known to restrict the replication of several virus families , which have , in turn , evolved mechanisms to counteract this PML-mediated antiviral restriction . Herpesviruses like HCMV , Varicella-Zoster virus ( VZV ) , Epstein-Barr virus ( EBV ) and HSV1 counteract the antiviral function of PML NB by either interfering with PML NB-associated proteins like Daxx and Sp100 , or with PML SUMOylation , leading to the disruption of these cellular antiviral structures [46–56] . For KSHV , it was shown that vIRF3 , RTA and ORF75 are antagonists of PML NBs leading to their SUMO-dependent and proteasome-mediated degradation or to the disappearance or the dispersion of PML NB-associated proteins like ATRX , Sp100 and Daxx [44 , 45 , 57 , 58] . The present study focuses on the second vIRF protein , vIRF2 , which has previously been shown to interact directly or indirectly with IRF1/2/3/8/9 , p65 , p300 , PKR , Caspase-3 as well as with STAT1 , and which could therefore interfere with the cellular IFN pathway , apoptosis , activation-induced cell death ( AICD ) and the PI3K/Akt signaling pathway [35 , 39 , 59–63] . In addition to the full-length vIRF2 protein translated from a spliced mRNA including the two exons K11 . 1 and K11 , the existence of a smaller protein generated only from the unspliced K11 . 1/K11 mRNA and therefore the K11 . 1 exon has been reported but is controversial [26 , 64 , 65] . Expression of the spliced vIRF2 mRNA has previously been reported to occur only in the lytic replication cycle in different KSHV positive cell lines [26 , 65–68] . Recently , it was shown that vIRF2 is able to bind the promoter regions of different cellular genes by means of a DNA-binding domain ( DBD; aa 7–114 ) ; a nuclear localization signal ( aa146-159 ) has been identified at the N-terminal end of the vIRF2 protein [36 , 69] . Most previous studies on the vIRF2 protein have examined its function by transfecting an expression vector for the spliced vIRF2 mRNA . In this study we investigated the function of vIRF2 during KSHV lytic and latent replication . We show that the vIRF2 protein dampens KSHV early viral gene expression in endothelial cells . Our results suggest that this regulation of the viral life cycle might be connected to the vIRF2-mediated activation of IFN-induced cellular proteins that restrict KSHV lytic gene expression .
When infecting primary human umbilical vein endothelial cells ( HUVEC ) with KSHV we noted that , in contrast to other herpesviruses such as HCMV , which is known to degrade PML NBs [46–48 , 53 , 55 , 56] , and the previously reported ability of KSHV vIRF3 to do the same [44 , 45] , KSHV does not disrupt these nuclear structures ( Fig 1A ) . While HCMV infection leads to a diffuse distribution of PML staining , infection of HUVECs with KSHV significantly increased the number of discrete PML NBs in infected cells in comparison to uninfected cells ( Fig 1A and 1B ) . In accordance with this immunofluorescence experiment , protein levels , as measured on western blots , of PML itself and of the PML NB components Sp100 and Daxx , decreased after HCMV infection , but increased after KSHV infection of primary endothelial cells ( Fig 1C ) . This KSHV-induced increase in PML and Daxx protein levels and in the number of PML NBs is reminiscent of the effect of IFN treatment , which is shown in Fig 1D and has been described in many previous studies [70–72] . Such an ‘IFN-like’ phenotype would be in keeping with the previously noted upregulation of interferon-stimulated cellular genes following KSHV infection [73–75] . However , several KSHV vIRFs have been reported to act as interferon antagonists that counteract the induction of the IFN pathway or the expression of ISGs [35 , 39 , 59 , 61 , 63] , vIRF3 has been shown to mediate the degradation of PML NBs [44 , 45] . We therefore compared the effect of all four KSHV vIRF proteins on PML NB formation by transfecting expression vectors for tagged cDNAs of all vIRFs into HeLa cells and analyzing the number of PML NBs by immunofluorescence . As described for B cells [44 , 45] , overexpression of vIRF3 causes a reduction in the number of PML NBs also in HeLa cells ( Fig 2A and 2B ) . In contrast , while vIRF1 and vIRF4 did not affect the number of PML NB in transfected cells , the overexpression of vIRF2 led to an increase in PML NB formation compared to the control vector ( Fig 2A and 2B ) . We then transduced HUVECs with either a control or a vIRF2 expressing lentiviral vector and found PML protein levels as well as mRNA levels to be increased in the presence of vIRF2 ( Fig 2C and 2D ) . These results suggest that , in contrast to vIRF3 , vIRF2 increases the expression of PML at the transcriptional level and thereby PML NB numbers in a similar manner as is seen following the infection of primary endothelial cells with KSHV ( Fig 1A–1C ) or treatment with IFNα in epithelial cells ( Fig 1D ) . This observation would seem to be at odds with the previously reported ability of transfected vIRF2 to antagonize the activation of the IFNβ promoter and the type I interferon-induced JAK-STAT signal transduction cascade [61 , 63] . We therefore wanted to investigate the role of vIRF2 in KSHV-infected cells . To follow the expression of vIRF2 in KSHV-infected cells we generated a monoclonal antibody against a recombinant vIRF2 fragment corresponding to the entire K11 exon . We expressed the K11 exon as a GST-K11-6xHis fusion protein in E . coli and immunized mice to obtain two hybridoma cell lines producing IgG2b/κ monoclonal antibodies #30F9 and #31A2 . We mapped the epitopes recognized by these two monoclonal antibodies using a peptide array that spanned the K11 protein sequence ( S1 Fig ) . Both monoclonal antibodies recognized the same repeated epitope ( QGPMQSEG ) located at position aa347-354 and aa405-412 within the repeated sequence ELLCETASPQGPMQSEGGEEGSTES in the regions R1 and R2 of vIRF2 , as illustrated in Fig 3B and S1 Fig . In addition , we observed a more weakly reactive third epitope ( S1 Fig , peptides 41–44 ) within the vIRF2 protein , whose sequence is only partially homologous to that of the first two epitopes and which might be detected by our two monoclonal antibodies which a lower affinity . Using the two monoclonal antibodies we were able to detect vIRF2 in KSHV-infected PEL cell lines as well as in in vitro infected endothelial and B cells ( Fig 3C ) . The vIRF2 expression pattern in the PEL-derived cell lines BC1 , BC3 and BCBL1 was similar , with a low basal expression that increased over time following induction of the lytic replication cycle . We also used HuARLT . rKSHV . 219 cells , a conditionally immortalized human endothelial cell line derived from HUVECs expressing the doxycycline-inducible simian virus 40 ( SV40 ) large T antigen ( TAg ) and a human telomerase reverse transcriptase ( hTert ) , as well as the B cell line BJAB . rKSHV . 219 , which are both latently infected with recombinant KSHV . Both cell lines showed vIRF2 expression only after lytic induction . The low basal vIRF2 expression in the three PEL cell lines ( Fig 3C ) may therefore reflect that a small percentage of these cells shows lytic gene expression even in the absence of external stimuli . Interestingly , we could detect several different bands for the vIRF2 protein , of which the full-length form with the highest molecular weight ( ~ 110 kDa ) showed the strongest intensity . We also investigated the subcellular localization of vIRF2 by a fractionation assay on KSHV positive BJAB . rKSHV . 219 cells after lytic cycle induction ( Fig 3D ) . While two adjacent bands of the molecular weight expected for the full-length vIRF2 were detected in the cytoplasm , the nucleus contained mainly the upper of the two bands . These observations confirm that KSHV vIRF2 is expressed at a low basal level in latent PEL cell lines and strongly expressed early after activation of the lytic replication cycle , as predicted from previous studies of its mRNA [26] . Our observation that vIRF2 is found both in the nucleus and cytoplasm of KSHV-infected cells , is in keeping with overexpression studies that used an expression vector for the spliced K11 . 1/K11 cDNA [69] . Having shown that the vIRF2 protein is expressed during lytic viral replication and that it increases the expression of PML and the number of PML NBs , we wanted to investigate its function in KSHV infected cells . As a first step we overexpressed vIRF2 in the KSHV-infected immortalized endothelial cell line HuARLT . rKSHV . 219 [73] via lentiviral transduction and induced the lytic cycle . We could observe a strong inhibition of KSHV early lytic protein expression upon vIRF2 overexpression , as judged by the reduced expression of the KSHV early lytic protein K-bZIP ( Fig 4A ) . In addition , we silenced vIRF2 in the same cell system using three siRNAs targeting different regions of the vIRF2 mRNA , as illustrated in Fig 3B . Using the siRNAs ( 2 ) and ( 3 ) , which target regions within the vIRF2 K11 sequence , the vIRF2 protein expression was suppressed to undetectable levels ( Fig 4B , top panel ) . In contrast , using siRNA ( 1 ) , which targets vIRF2 mRNA at the 5’ end of exon K11 . 1 , resulted only in the silencing of the protein form with the highest molecular weight ( Fig 4B , top panel ) . We hypothesize that siRNA ( 1 ) interferes with the translation of the vIRF2 mRNA from the first translational start codon . Interestingly , the knockdown with any of these three siRNAs led to an increased early lytic gene expression , as shown here for the early lytic proteins K-bZIP and ORF45 ( Fig 4B ) . To verify a possible role of vIRF2 in lytic cycle regulation with another experimental approach , we used the BAC16 vector [76] to generate a vIRF2 deletion mutant ( KSHV . ΔvIRF2 ) by replacing the ORFs K11 and K11 . 1 with a kanamycin resistance cassette ( Fig 4C ) . We sequenced the entire BAC16 KSHV . ΔvIRF2 to verify the correct insertion of the kanamycin cassette and the integrity of the rest of the KSHV genome . We then established stable HEK-293 . BAC16 . KSHV . WT and HEK-293 . BAC16 . KSHV . ΔvIRF2 cell lines by transfecting , respectively , the BAC16 . KSHV . WT and BAC16 . KSHV . ΔvIRF2 genomes into HEK-293 cells and selecting stable hygromycin B resistant bulk populations . Following induction of the lytic replication cycle , expression of the vIRF2 protein was detected in the HEK-293 . BAC16 . KSHV . WT cells but not in the HEK-293 . BAC16 . KSHV . ΔvIRF2 cell line . There was no detectable difference in the expression level of early lytic proteins upon reactivation in the KSHV . ΔvIRF2-infected cells compared to the KSHV . WT-infected cells ( S2 Fig ) . To investigate the function of vIRF2 in the context of the whole viral genome in endothelial cells , we established stable HuARLT . BAC16 . KSHV . WT and HuARLT . BAC16 . KSHV . ΔvIRF2 cells by infecting the immortalized endothelial cell line HuARLT with tissue culture supernatants of induced HEK-293 . BAC16 . KSHV . WT and HEK-293 . BAC16 . KSHV . ΔvIRF2 cells and selecting hygromycin B resistant bulk populations . Following induction of the lytic cycle , vIRF2 protein expression could be detected only in the HuARLT . BAC16 . KSHV . WT cells , but not in the HuARLT . BAC16 . KSHV . ΔvIRF2 cells ( Fig 4D ) . In this endothelial cell lines , we observed an increased expression of the lytic proteins K-bZIP and ORF45 in KSHV . ΔvIRF2- compared to KSHV . WT-infected cells following activation of the lytic cycle ( Fig 4D ) . In addition , the viral titer released from induced KSHV . ΔvIRF2-infected HuARLT cells was about 10 fold higher than from induced KSHV . WT-infected cells ( Fig 4E ) . To confirm that the increased lytic gene expression in KSHV . ΔvIRF2-infected HuARLT cells was due to the absence of the vIRF2 protein we complemented vIRF2 by lentiviral transduction in stable HuAR2T . BAC16 . KSHV . WT and HuAR2T . BAC16 . KSHV . ΔvIRF2 cells and induced the lytic replication cycle . The vIRF2 overexpression led to an inhibition of early viral protein expression in the HuAR2T . BAC16 . WT and also in the HuAR2T . BAC16 . ΔvIRF2 cells ( Fig 4F ) . This experiment showed that overexpressed vIRF2 is able to suppress early viral protein expression not only in KSHV . WT-infected endothelial cells ( as already shown in Fig 4A ) , but also the increased lytic gene expression seen in KSHV . ΔvIRF2-infected endothelial cells . Together , these results show that , in endothelial cells , the deletion of vIRF2 from the viral genome or its suppression by siRNA promotes early viral protein expression and , in the case of the vIRF2 deletion , the production of viral progeny . After infection of the host cell , KSHV initially expresses lytic viral proteins but then switches off their expression as latency is established . Latency represents an important step during the KSHV life cycle and is indispensable for KSHV pathogenesis . In view of our observations in experiments on lytic reactivation from latency , we were curious if the vIRF2 protein is also involved in the regulation of lytic viral protein expression following a new infection . For this , we infected HuARLT cells that had been microporated with two siRNAs targeting ORF K11 ( see above , Fig 3B ) with a recombinant KSHV . We observed an increase in the expression of the lytic proteins K-bZIP and ORF45 upon the transient knockdown of vIRF2 by siRNA ( 3 ) which targets vIRF2 mRNA at the C-terminal end ( Fig 5A and Fig 3B ) . Using another siRNA , directed against a region in the center of the vIRF2 mRNA , we could only observe an increase in ORF45 , but not of K-bZIP expression in comparison to the control siRNA ( Fig 5A ) . To confirm this result , we used cell free KSHV . WT or KSHV . ΔvIRF2 virus , which was generated in reactivated HEK-293 . BAC16 . KSHV . WT and HEK-293 . BAC16 . KSHV . ΔvIRF2 cell lines ( see S2 Fig ) , to infect HuARLT cells . We verified equal infection levels in HuARLT cells by western blot for the latent KSHV LANA protein ( Fig 5B , top panel ) . KSHV . WT-infected cells showed the strongest expression of the early lytic KSHV proteins K-bZIP and ORF45 48 hours after infection , which was followed by a decline at 72 hours , indicating the onset of latency establishment . In contrast , cells which were infected with the KSHV . ΔvIRF2 virus showed an increased and prolonged early lytic protein expression which did not decline at 72 hours after infection . Similarly , the expression of the late lytic KSHV glycoprotein K8 . 1 ( Fig 5B , K8 . 1 A/B panel ) was increased in KSHV . ΔvIRF2- compared to KSHV . WT-infected cells at 72 hours after infection . In addition , we observed again that PML protein expression is strongly increased upon infection with the KSHV . WT virus but much less after infection with the vIRF2 knockout virus ( Fig 5B , PML panel ) . These data indicate that vIRF2 not only inhibits KSHV lytic gene expression during reactivation but also after de novo infection . As mentioned above , two different protein forms of vIRF2 have previously been described . Among them is the short variant encoded by the K11 . 1 exon only and a protein of approximately 110 kDa apparent molecular weight resulting from the translation of the spliced K11 . 1/ K11 mRNA [26 , 35 , 59–61 , 63 , 65 , 77] . Interestingly , by using our newly produced monoclonal antibodies against vIRF2 , we were able to detect four additional protein forms in different KSHV positive cell lines ( Fig 3C ) . To further understand which forms and parts of the vIRF2 protein are necessary and/or sufficient for its regulatory function during KSHV replication , we generated four different KSHV mutants ( Stop#1—Stop#4 ) by introducing double stop codons at different positions within the vIRF2 sequence in the BAC16 backbone using En passant mutagenesis ( Fig 6A ) . We also constructed the corresponding revertants ( Rev#1 , 2 and 4 ) with the exception of revertant #3 , which failed repeatedly for technical reasons . We then generated stable HEK-293 . BAC16 . KSHV cell lines with these mutants and their revertants , as mentioned before . Following reactivation , we found that all three revertant cell lines express the same vIRF2 pattern as KSHV . WT-infected cells ( Fig 6B , lane 9–11 , 3 ) . As already observed before , there was no expression in the KSHV . ΔvIRF2 cell line ( Fig 6B , lane 4 ) . In addition no vIRF2 protein was detected in the case of the KSHV . vIRF2 . Stop#2 cell line ( Fig 6B , lane 6 ) , in keeping with the position of the antibody epitopes E1/E2 ( S1 Fig ) downstream of the inserted stop codons ( Fig 6A ) . It is , however , likely that KSHV . vIRF2 . Stop#2 expresses a C-terminally truncated vIRF2 ( aa1-322 ) as illustrated in Fig 6A . In contrast , the KSHV . vIRF2 . Stop mutants #1 , #3 and #4 showed several bands of different molecular weight that were detected by the vIRF2 antibody ( Fig 6B , lane 5 , 7 , 8 ) . We assume that these bands result from translational initiation at internal methionine codons , indicated by red numbers in the diagram in Fig 6A . Thus , the protein bands with the highest molecular weight observed in the case of the KSHV . vIRF2 . Stop#1 and KSHV . vIRF2 . Stop#4 mutants could initiate at a methionine in position aa47 of the vIRF2 sequence , and bands of lower apparent molecular weight could be translated from other internal methionine codons . The weak immunoreactive band at ~ 30 kDa , detected with the monoclonal antibody in the case of KSHV . vIRF2 . Stop#3 , suggests that either this antibody mainly detects epitope E2 in the second repeat in the context of the entire vIRF2 protein , in spite of reacting with epitopes E1 and E2 on the peptide array ( S1 Fig ) or the vIRF2 protein expressed by the Stop#3 mutant could be unstable . Neither the HEK-293 . KSHV . ΔvIRF2 cell line , nor the HEK-293 . KSHV . vIRF2 . Stop#2-#4 cell lines differed significantly from the HEK-293 . KSHV . WT cell line with regard to the expression of the KSHV lytic protein K-bZIP ( Fig 6B ) . We interpret the lower levels of K-bZIP expression observed with Stop mutant #1 as not being due to the absence of the full-length vIRF2 form , as it is also seen in the corresponding revertant #1 ( Fig 6B , lanes 5 , 9 ) . To investigate if the individual KSHV . vIRF2 stop mutants differed in their reactivation potential in endothelial cells , we established stable HuARLT populations infected with KSHV . WT or the KSHV . vIRF2 stop mutants and the corresponding revertants . Interestingly , while the complete deletion of vIRF2 led to an increased expression of KSHV K-bZIP and ORF45 as noted before ( Fig 4C , Fig 6C lane 2 ) , KSHV . vIRF2 stop mutants #2–4 did not increase the levels of these lytic proteins , while Stop mutant #1 showed an increase in ORF45 but not K-bZIP expression in endothelial cells ( Fig 6C ) . This indicates that the truncated vIRF2 protein forms expressed by stop mutants #2-#4 were still functional and sufficient to maintain the WT phenotype . In particular , these results suggest that the first 460 amino acids of the vIRF2 protein , which are likely produced by KSHV . vIRF2 . Stop#4 , and probably the first 322 amino acids , which are likely produced by KSHV . vIRF2 . Stop#2 but are not detected by our antibody , are sufficient to restrict KSHV lytic gene expression . The phenotype observed with the Stop# 1 ( increase in ORF45 but not K-bZIP expression ) is more difficult to interpret but could suggest that the first 47 amino acids , which are likely lacking in KSHV . vIRF2 . Stop#1 , are required for vIRF2 to restrict lytic gene expression . Having shown that vIRF2 increases the expression of cellular antiviral factors like PML ( Fig 2 ) and that it has an inhibitory effect on KSHV lytic replication ( Fig 4 and Fig 5 ) we wondered if vIRF2 regulates the transcription of cellular genes . To address this question we performed a microarray-based mRNA expression analysis . We used empty HuARLT cells , the HuARLT . BAC16 . KSHV . WT- and the HuARLT . BAC16 . KSHV . ΔvIRF2-infected cells , treated them with the reactivation cocktail for 48 h and extracted mRNA , which was then analyzed using a gene expression microarray . The original microarray data were filtered for up- or downregulated genes that showed a 2 fold or higher difference between KSHV . ΔvIRF2-infected and KSHV . WT-infected cells following lytic reactivation in two independent experiments . This yielded 434 cellular genes that could be regulated , directly or indirectly , by vIRF2 in KSHV-infected cells . This list was sorted for their involvement in biological processes using Gene Ontology . This yielded 51 genes known to be involved in innate or intrinsic defense response ( Table 1 ) . Among these vIRF2-regulated genes were the three IFN-induced proteins with tetratricopeptide repeats ( IFIT1 , 2 and 3 ) , whose expression was downregulated in the KSHV . ΔvIRF2-infected compared to KSHV . WT-infected cells . The average fold differences ranged from 3 . 1 for IFIT1 , 3 . 8 for IFIT3 and 6 . 1 for IFIT2 . The vIRF2-regulated ISGs IFIT1 , 2 and 3 are known to be involved in the cellular antiviral response . Through their IFN-dependent expression , IFIT proteins are induced upon infection with RNA and DNA viruses and exhibit antiviral functions by interfering with the viral life cycle at different steps . It is known that they antagonize RNA viruses like Vesicular stomatitis virus ( VSV ) , Hepatitis C Virus ( HCV ) , West Nile virus ( WNV ) , Sendai Virus ( SeV ) , Japanese encephalitis Virus ( JEV ) as well as two dsDNA viruses , the Human Papillomavirus ( HPV ) and HCMV [78–89] . To confirm the microarray results we induced the lytic cycle in KSHV . WT- and KSHV . ΔvIRF2-infected HuARLT cells and analyzed the IFIT protein expression by western blot . Consistent with the microarray data , we could detect a strong induction for all three IFIT proteins in the KSHV WT-infected cells already 24 h after reactivation . This increased IFIT expression was not detected in the KSHV . ΔvIRF2-infected cell line , in which the IFIT protein levels remained the same as in the absence of lytic induction ( Fig 7A ) . As noted before ( Fig 4C ) , the expression of the KSHV lytic K-bZIP protein was higher in KSHV . ΔvIRF2-infected compared to KSHV . WT-infected HuARLT cells ( Fig 7A ) . In addition , we observed an increase in IFIT 1 and 3 protein levels upon KSHV reactivation in the PEL derived B cell line BC1 ( S3A Fig ) . In this experiment , IFIT1 was not detectable in non-reactivated cells and its expression increases between 24 h and 48 h following reactivation . In contrast , IFIT3 showed a basal level of expression already in unstimulated cells , which increases at 72 h after reactivation ( S3A Fig ) . To investigate if vIRF2 causes the increased IFIT protein expression , we overexpressed vIRF2 by lentiviral transduction in either empty HuARLT or HuARLT . rKSHV . 219 cells with or without reactivation . The vIRF2 overexpression in empty HuARLT cells resulted in a strong increase in the IFIT protein levels , which occurred independently of the reactivation cocktail ( Fig 7B , left ) . The KSHV-infected endothelial cells , HuARLT . rKSHV . 219 showed similar results , although the difference was not as strong as in the uninfected cells , probably because the basal vIRF2 expression in infected cells is already induced by lytic induction ( Fig 7B , right ) . In this experiment we again noted the suppression of K-bZIP expression following overexpression of vIRF2 ( Fig 7B , right ) . We also confirmed this vIRF2-mediated IFIT induction in uninfected primary endothelial cells by transducing human umbilical vein endothelial cells ( HUVECs ) with either the control or the vIRF2-expressing lentivirus . In this experiment , transduction of vIRF2 led to an increase in IFIT1 and 3 protein expression ( S3B Fig ) . Furthermore , we analyzed PML and IFIT1 expression in immortalized endothelial cells ( HuARLT ) stably infected with KSHV . WT , KSHV . ΔvIRF2 or the HuARLT . BAC16 . vIRF2 Stop mutants and their corresponding revertants following the activation of the lytic replication cycle ( S3C Fig ) . We could observe a little variability between the expression of IFIT and PML in the different stop mutants which could be explained by their protein loading level ( S3C Fig , see actin as a control ) . As already observed in the experiment shown in Fig 7A , IFIT1 expression was reduced in KSHV . ΔvIRF2-infected compared to KSHV . WT-infected cells ( S3C Fig ) . In contrast , Stop#1- and Stop#2-infected cells , as well as their revertants , maintained the expression of IFIT1 , indicating that the truncated vIRF2 proteins expressed by these mutants ( Fig 6A ) are sufficient to induce IFIT1 ( S3C Fig ) . As we could show that vIRF2 expression leads to an increased expression of the IFIT proteins , we wondered if this might correlate with the ability of vIRF2 to regulate early lytic viral protein expression during de novo infection . HuARLT cells were infected with either the KSHV . WT or the KSHV . ΔvIRF2 virus and IFIT protein expression was analyzed by western blot . As illustrated in Fig 7C , KSHV . WT strongly activates IFIT1-3 protein expression already 24 h after infection . In contrast , infection with the KSHV . ΔvIRF2 virus resulted in an attenuated , but still noticeable , induction of all three IFIT proteins . As noted before ( Fig 5B ) , infection with KSHV . ΔvIRF2 led to an increased expression of the lytic proteins ORF45 and K-bZIP , whose expression correlated inversely with that of IFIT1-3 ( Fig 7C ) . These results suggest that vIRF2 activates IFIT1-3 expression on its own and in the context of KSHV infection . To understand the connection between the vIRF2-induced IFIT expression and the ability of vIRF2 to inhibit early lytic protein expression , we investigated if the IFIT proteins interfere with KSHV protein expression . We silenced individual IFITs by siRNA and measured the impact on lytic protein expression during reactivation as well as in newly infected cells . Silencing IFIT1 during KSHV lytic reactivation ( Fig 8A , left ) as well as during de novo infection ( Fig 8A , right ) increased the expression of the KSHV early lytic protein K-bZIP . IFIT2 knockdown affected K-bZIP expression only during KSHV de novo infection ( Fig 8B and S4A Fig ) and silencing of IFIT3 increased K-bZIP expression only during lytic reactivation ( Fig 8C and S4B Fig ) . Notwithstanding the possibility that some of our siRNAs might have induced off-target effects , the ensemble of our findings suggest that all three IFIT proteins , but especially IFIT1 , are able to restrict KSHV lytic protein expression during reactivation and/or de novo infection . We also investigated if PML has a similar restrictive effect on KSHV lytic protein expression by silencing PML expression with siRNA during lytic reactivation in the immortalized endothelial cell line HuARLT . rKSHV . 219 . We found that a transient PML knockdown increases KSHV lytic protein expression , as shown for the early lytic protein K-bZIP ( Fig 8D ) . Interestingly , we could observe this effect for PML only during lytic reactivation and not during a de novo infection ( S4C Fig ) which is in accordance with previously reported findings [74] .
Among herpesviruses , vIRFs have so far only been found in Old World primate γ2 herpesviruses [30–32] . A number of studies have shown that the KSHV vIRFs can inhibit IFN-signaling and modulate anti-apoptotic as well as cell proliferation pathways [39] . Only a few studies have addressed the role of the KSHV vIRF proteins during KSHV replication . One study showed that KSHV vIRF4 facilitates lytic replication by targeting the expression of cellular IRF4 and c-myc [90] . In addition , vIRF4 interacts with CSF/CBF1 , a downstream effector of Notch signaling that is also targeted by KSHV RTA and LANA and that is required for efficient lytic reactivation [42 , 91] . Furthermore , vIRF4 has been shown to cooperate with RTA in the activation of several lytic promoters [92] . In contrast , vIRF3 has been shown to suppress , and vIRF1 to promote , lytic replication by recruitment of USP7 [40] or the BH3-only pro-apoptotic Bcl2 family member Bim [93] . However , in apparent contrast , vIRF3 degrades PML NBs in transfected cells , as reported before [44 , 45] and shown in Fig 2 . Our investigation was prompted by the observation that infection of primary endothelial cells with KSHV does not result in the disruption of PML NBs , in contrast to infection with HCMV ( Fig 1A and 1B ) and to what has been noted for many other herpesviruses [46–56] . We noted an increase in the number of PML NBs following KSHV infection of endothelial cells ( Fig 1A and 1B ) , reminiscent of an IFN-induced effect . Others have made a similar observation by showing IFN-induced PML transcription [74] and the increased expression of IFN-induced cellular genes following KSHV infection has been noted by several groups [73–75 , 94] . In view of the ability of vIRF3 to degrade PML NBs [44 , 45] and of several vIRFs to modulate IFN-signaling , we next compared the impact of all four KSHV vIRFs on PML expression and PML NBs . We found that vIRF2 promotes PML NB formation and increased PML protein levels , while vIRF1 and vIRF4 had no effect ( Fig 2 ) . We next investigated the expression of the vIRF2 protein during KSHV latency and lytic replication in different cell lines with the help of two newly generated monoclonal antibodies that recognized the same repeated epitope located in an internal repeat region of vIRF2 ( Fig 3B , S1 Fig ) . Our findings are consistent with previous studies that could detect increased vIRF2 mRNA levels after lytic cycle induction of different PEL as well as endothelial cells [26 , 65 , 67 , 68] . In addition , our results suggest that several vIRF2 protein forms exist ( Fig 3C and 3D ) . Some of them likely result from translational initiation at internal start codons , as they are produced by a KSHV mutant that had two stop codons inserted immediately after the predicted start codon for the vIRF2 ORF in exon K11 . 1 ( Fig 6 , Stop#1 ) . Similarly , the 75 kDa vIRF2 band detected in some KSHV positive cell lines ( Fig 3C ) could also result from internal translation initiation at the beginning of the K11 exon and correspond to a truncated protein expressed only by the K11 exon . This 75 kDa band is not detectable in the nuclear fraction shown in Fig 3D , possibly as a result of lacking the NLS , which is located at the C-terminal end of the K11 . 1 exon ( Fig 3B ) . Since our monoclonal antibodies recognize an epitope in the K11 exon , we cannot address the question whether a small protein that is only expressed from ORF K11 . 1 exists [35 , 67] . In KSHV-infected and reactivated cells , the full-length vIRF2 protein is found both in the cytoplasm and in the nucleus ( Fig 3D ) , as predicted from previous transfection experiments with a vIRF2 expression vector [69] . Using transient knockdown experiments with different vIRF2 siRNAs , as well as a KSHV mutant lacking the vIRF2 gene , we could show that vIRF2 restricts the expression of early lytic KSHV proteins both after reactivation from latency ( Fig 4 ) and following de novo infection ( Fig 5 ) . By generating different KSHV mutants with translational stop codons inserted at different positions in the vIRF2 sequence we could show that the N-terminal part of vIRF2 which contains the DBD as well as an NLS is sufficient to inhibit KSHV lytic gene expression in endothelial cells ( Fig 6 ) . A comparison of the cellular transcriptome in KSHV . WT- and KSHV . ΔvIRF2-infected cells showed that , at protein levels found in KSHV-infected cells , vIRF2 is a transcriptional regulator of different cellular genes ( Table 1 ) . This is in line with the fact that it binds to different cellular promoter regions [36] . Among the cellular genes involved in host defense that are differentially expressed in KSHV . WT- vs . KSHV . ΔvIRF2-infected endothelial cells ( Table 1 ) , we could identify the IFN-induced proteins with tetratricopeptide repeats 1 , 2 and 3 ( IFIT1 , 2 and 3 ) , whose expression we could show to be directly or indirectly induced by vIRF2 ( Fig 7 ) and which restrict KSHV lytic protein expression ( Fig 8 ) . This finding extends the list of IFIT-restricted viruses , which so far only include two DNA viruses , HPV and HCMV [83 , 84 , 89] to another herpesvirus , KSHV . Together our observations suggest that vIRF2 restricts KSHV early lytic protein expression by promoting the expression of IFN-regulated cellular genes that act as antiviral restriction factors . IFIT1-3 and PML are examples of such IFN-induced cellular proteins with the ability to restrict KSHV ( Fig 8 ) , but not necessarily the only ones that could contribute to the ability of vIRF2 to dampen KSHV lytic protein expression . It would therefore appear that vIRF2 enables KSHV to harness an antiviral cellular response to dampen lytic replication and therefore to establish or maintain latency . A role for type I interferons in the maintenance of latency has been reported for cytomegalovirus [95 , 96] . Similarly , IFN-α promotes latency establishment in sensory neurons of the alphaherpesviruses HSV and pseudorabies virus [97] . Epstein-Barr virus latent membrane protein 1 ( LMP1 ) has been shown to prime latently EBV-infected cells for the production of endogenous IFN and the activation of ISGs [98 , 99] . In the case of these three examples , the virus takes advantage of IFN produced by infected or neighboring cells , whereas in the case of KSHV , vIRF2 can mimic or induce an IFN response in the absence of exogenous type I interferons . KSHV infection is known to enhance the expression of several ISGs and KSHV vFLIP has previously been shown to contribute to this process [73–75 , 94] . In addition , a cytoplasmic variant of KSHV LANA has been shown to promote lytic reactivation by antagonizing the cGAS-dependent activation of the IFN pathway [100] and other KSHV proteins can modulate the function of cGAS [101–103] . KSHV vIRF2 may thus provide KSHV with an additional mechanism to harness an antiviral cellular mechanism for the purpose of latent viral persistence .
The use of human umbilical cords was approved by the Hannover Medical School Ethics Committee and experiments were performed in agreement with the Declaration of Helsinki . Written informed consent was obtained from parents of umbilical cord donors . The primary antibodies for western blot and immunofluorescence analysis are listed below . The antibodies goat anti-Calnexin ( sc-6465 ) , goat anti-Lamin A/C ( sc-6215 ) , mouse anti-IFIT2 ( sc-390724 ) , mouse anti-IFIT3 ( sc-393512 ) , mouse anti-KSHV K-bZIP ( sc-69797 ) , mouse anti-KSHV ORF45 ( sc-53883 ) and mouse anti-KSHV K8 . 1 ( sc-65446 ) were purchased from Santa Cruz Biotechnology . The primary antibody mouse anti-β-actin ( A2228 ) was purchased from Sigma Aldrich , the mouse anti-GFP ( 632381 ) from Clontech , the rabbit anti-IFIT1 ( D2X9Z ) from Cell Signaling Technologies , the rabbit anti-PML ( A301-167A ) from Bethyl Laboratories , the rabbit anti-Daxx ( 1094–1 ) from Epitomics and the rabbit anti-Sp100 ( AB1380 ) from Chemicon . For the detection of KSHV LANA we used a rat monoclonal anti-LANA antibody [104] , which we produced from the hybridoma cell line . To detect vIRF2 the newly produced monoclonal mouse antibody clones #30F9 and #31A2 ( see below ) were used for western blot as well as for immunofluorescence ( IFA ) . We used the mouse anti-PML ( PG-M3 , sc-966 ) from Santa Cruz Biotechnology to visualize PML nuclear bodies . The antibody to HCMV pp72 ( IE1 ) was purchased from PerkinElmer ( NEA-9221 ) . The secondary antibodies for western blot and IFA are listed below . The HRP-conjugated secondary antibodies rabbit anti-mouse IgG ( P0447 ) and goat anti-rabbit IgG ( P0448 ) were purchased from Dako . The HRP-conjugated rabbit anti-rat IgG antibody ( 3050–05 ) was purchased from SouthernBiotech . The HRP-conjugated mouse IgGκ light chain binding protein ( sc-516102 ) was purchased from Santa Cruz Biotechnology and the 800CW infrared dye labeled goat anti-mouse IgG ( 926–32210 ) from Odyssey . The secondary antibodies for IFA , the FITC-conjugated goat anti-mouse IgG ( 115-095-146 ) and the Lissamine Rhodamine ( LRSC ) -conjugated goat anti-mouse IgG ( 115-025-072 ) were purchased from Jackson ImmunoResearch . HEK-293 ( CRL-1573 , American Type Culture Collection , ATCC ) , stable HEK-293 . BAC16 . KSHV and HeLa ( CCL-2 , ATCC ) cells were cultured in DMEM ( Gibco ) supplemented with 10% fetal bovine serum ( FBS , Sigma ) . Primary human umbilical vein endothelial cells ( HUVECs ) were isolated from umbilical cords by collagenase digestion as described before [105] and cultured in EGM-2MV BulletKit from Lonza . HuARLT cells , a conditionally immortalized human endothelial cell line derived from HUVECs [106] , expressing the doxycycline-inducible simian virus 40 ( SV40 ) large T antigen ( TAg ) and a human telomerase reverse transcriptase ( hTert ) ( kindly provided by Dagmar Wirth , Helmholtz Centre for Infection Research Braunschweig ) , stable HuARLT . BAC16 . KSHV and the cell line HuARLT . rKSHV . 219 , HuARLT cells latently infected with rKSHV . 219 [73] were cultured in EGM-2MV BulletKit from Lonza or in Microvascular Endothelial Cell Growth Medium enhanced from PeloBiotech supplemented with 1 μg/ml doxycycline . HuARLT . rKSHV . 219 cells were in addition cultured with 5 μg/ml puromycin . The PEL-derived B cell lines BC1 ( CRL-2230 , ATCC ) , BC3 ( CRL-2277 , ATCC ) and BCBL1 ( ACC 683 , German Collection of Microorganisms and Cell Culture , DSMZ ) as well as the B cell lines BJAB ( ACC 757 , DSMZ ) and BJAB . rKSHV . 219 ( BJAB cells latently infected with rKSHV . 219 ) [107 , 108] were cultured in RPMI 1640 ( Gibco ) supplemented with 20% FBS . The BJAB . rKSHV . 219 cells were in addition cultured with 4 . 2 μg/ml puromycin . Additionally , all stable HEK-293 . BAC16 . KSHV and HuARLT . BAC16 . KSHV cells were cultured in the presence of 100 μg/ml hygromycin B . The stable HEK-293 . BAC16 . KSHV cell lines were generated by transfection of HEK-293 cells with 2 μg BAC DNA of a Maxi preparation ( Macherey-Nagel , NuceloBond BAC100 ) with Fugene 6 transfection reagent ( Roche , 11 814 443 001 ) according to the manufacturer’s protocol at a ratio of 3:1 . Three days after transfection the cells were transferred from a 6 well plate into 10 cm dishes and BAC16 containing cells were selected by adding 100 μg/ml hygromycin B . The stable HuARLT . BAC16 . KSHV cell lines were established by infection of HuARLT cells ( at a MOI of 2 ) with the respective BAC16-derived virus produced in HEK-293 . BAC16 . KSHV cells . KSHV positive cells were selected with 100 μg/ml hygromycin B starting three days after infection . HEK-293 cells were treated with IFNα ( Sigma , SRP4596 ) for 72 h . Insect SF9 cells were cultured in spinner flasks at a density between 0 . 5 and 2∙106 cells in grace’s insect media ( Gibco ) . Lytic KSHV reactivation in HuARLT . rKSHV . 219 , HuARLT . BAC16 . KSHV and HEK-293 . BAC16 . KSHV cell lines was induced by treating the cells with 1 . 67 mM sodium butyrate ( SB ) and 10–30% SF9 cell culture supernatant containing KSHV RTA-expressing baculovirus . BJAB . rKSHV . 219 cells were induced by adding 2 . 5 μg/ml anti-human IgM [107 , 108] . The recombinant KSHV . 219 virus was produced from BJAB . rKSHV . 219 cells which were induced for three days with 2 . 5 μg/ml anti-human IgM , while BAC16-derived KSHV or KSHV mutants were produced in HEK-293 . BAC16 cells using 1 . 67 mM SB and 30% tissue culture supernatant containing RTA-expressing baculovirus for three days . All virus containing supernatants used for infection were centrifuged for 5 min at 2500 rpm at 4°C , filtered with a 0 . 45 μM filter and ultracentrifuged for 4–5 h at 15–18 , 000 rpm at 4° C . The pellets were resuspended in medium without FBS and stored at 4°C . To determine the titer of virus stocks , a serial dilution of the virus was prepared and added to 3x104 HEK-293 cells in a 96 well plate , which had been seeded the day before . The plate was centrifuged for 30 min at 32°C and 450xg . After 72 h the viral titers were calculated by counting the number of GFP positive HEK-293 cells . HUVEC or HuARLT cells were infected with either rKSHV . 219 , BAC16 . KSHV . -derived virus or HCMV at the desired multiplicity of infection ( MOI ) in the presence of 8 μl/ml polybrene . The plates were centrifuged at 32°C and 450xg for 30 min . The KSHV RTA-expressing baculovirus , used to induce the KSHV lytic cycle , was produced in SF9 cells . The cells were cultured in spinner flasks to a maximum density of 2x106 cells . SF9 cells were seeded at a final cell density of 0 . 5x106 cells/ml and infected with a baculovirus stock from a previous production . After four days the cells were collected and centrifuged at 1000 rpm for 20 min . The supernatant was filtered with a 0 . 45 μm filter and directly used for lytic cycle induction . Transfection of cells with plasmid DNA was performed with Fugene 6 ( Roche , 11 814 443 001 ) according to the manufacturer’s protocol at a ratio of 3:1 . The vector pCDNA3 . 1 ( + ) was used as a control ( V79020 , Invitrogen ) and the pcDNA3 . 1 ( + ) . vIRF1 . myc/His , pcDNA3 . 1 ( + ) . vIRF3 . myc/His and pcDNA3 . 1 ( + ) . vIRF4/His were kindly provided by Frank Neipel ( Erlangen , Germany ) . To generate a pcDNA3 . 1 vector expressing full length vIRF2 , vIRF2 cDNA was amplified from the PEL derived cell line BC3 after induction of the lytic cycle with 1 . 25 mM sodium butyrate and 10% tissue culture supernatant containing RTA-expressing baculovirus . RNA isolation was performed using the RNeasy Mini Kit ( 74104 , Qiagen ) with an additional on-column DNase digestion . The vIRF2 RNA was reversed transcribed by the expand reverse transcriptase ( 11785826001 , Roche ) and the cDNA was then amplified using 5’-TATGGATCCATGCCTCGCTACACGGAGTCGG-3’ and TATTCTAGATTACAGATCCTCTTCTGAGATGAGTTTTTGTTCGTCTCTGTGGTAAAATGGGGC-3’ followed by cloning into the pcDNA3 . 1 ( + ) vector . To generate the vIRF2 lentiviral vector a T2A element was introduced into the pcDNA3 . 1 ( + ) . vIRF2 . myc vector ( see above ) carrying the vIRF2 cDNA by amplifying the T2A element from a vFLIP vector [73] using the primers 5’-TATGCTAGCAGGGCTCCGGAGAGGGCCGGGGCTCTC-3’ and 5’-TATGGATCCAGGGGCCGGGGTTCTCCTCCACGT-3’ . Thereafter the T2A and vIRF2 . myc fused product was amplified with 5’-TTGCTGTACAAGGGCTCCGGAGAGGGCCGGGGCTCTC-3’ and 5’-TATGTCGACTTACAGATCCTCTTCTGAGATGAGTTTTTGTTCGTCTCTGTGGTAAAATGGGGC-3’ . The fragment was then digested with respective restriction enzymes and ligated into the lentiviral vector pRRL . PPT . SF . GFP ( kindly provided by Axel Schambach , Medical School Hannover ) , previously linearized with BsrgI and SalI enzymes , to generate the pRRL . PPT . SF . vIRF2 . T2A . GFP vector . For lentivirus production , HEK-293 . T cells ( 5∙106 cells/10 cm dish ) seeded the day before were transfected with 10 μg of the pRRL . PPT . SF . vIRF2 . T2A . GFP plasmid or the corresponding control vector pRRL . PPT . SF . GFP as well as the helper plasmids pMDLGg/p ( 6 . 5 μg ) , pRSV–REV ( 2 . 5 μg ) , and pMD2 . G ( 3 . 5 μg ) ( kindly provided by Renata Stripecke , Medical School Hannover ) using the calcium phosphate method . The supernatants containing the lentivirus were collected after 36 h and 48 h . To concentrate the virus stock , the supernatants were filtered with a 0 . 45 μm filter and ultracentrifuged over night at 10 , 000 rpm and 4°C . The next day , the pellet was resuspended in medium without FBS and virus aliquots were stored at -80°C . HuARLT cells were transduced by adding the calculated amount of virus to the cells in the presence of 8 μl/ml polybrene and centrifugation at 32°C and 450xg for 30 min . To produce a vIRF2 protein encoded by the K11 exon , the K11 sequence was cloned by restriction enzyme digestion and ligation into a pGEX-6P-1 . GST vector and a 6xHis tag was fused to the C-terminus using the primers 5’-CGGGATCCAGGGAGGCCGCCAGGAAACAG-3’ and 5’-CCGGAATTCTTAGTGGTGATGGTGATGATGGTCTCTGTGG-3’ . To produce the GST fused vIRF2 protein or the GST control after transformation of E . coli Rosetta cells , the bacteria culture was incubated at 37°C at 220 rpm until it reached a density between OD600 0 . 4–0 . 6 . The protein production was induced by adding 1 mM isopropyl-b-D-thiogalactopyranoside ( IPTG , I6758 , Sigma ) and incubated for 4 h at 30°C . The culture was centrifuged for 10 min at 5000 rpm and 4°C and the pellet was resuspended in 100 ml resuspension buffer ( 1x PBS + protease inhibitors mix: 1 mM Aprotinin , 10 μM Leupeptin , 100 μM Phenylmethylsulfonylfluorid ( PMSF ) , 1 . 46 μM Pepstatin A , 1 mM Benzamidine HCl ) . The cells were sonicated on ice five times for 30 sec and completely lysed by adding 0 . 5% NP40 . After centrifugation for 10 min at 14 , 000 rpm and 4°C , the supernatant was collected and centrifuged again under the same conditions . For pull down of GST fused proteins , 1 ml glutathione-sepharose 4 Fast Flow beads ( 17513202 , GE Healthcare ) were washed in washing buffer ( 1x PBS , 0 . 5% NP40 , 5% Glycerol ) and added to the lysate . After incubation over night at 4°C while shaking gently , beads were washed three times in washing buffer . To elute the proteins from the beads , 2 . 5 ml elution buffer ( 1x PBS , 0 . 5% NP40 , 10% Glycerol , 60 mM Glutathione , protease inhibitors mix , pH 7 . 3 ) were added and incubated for 3 h at 4°C while gently mixing . Afterwards , the solution was centrifuged , the supernatant was collected and recentrifuged . The protein solution was either stored at -80°C after adding 10% glycerol or dialyzed over night in a Slide-A-Lyzer Dialysis cassette ( 66332 , ThermoFisher Scientific , 3 . 500 MWCO , 0 . 5–3 ml capacity ) in dialysis buffer ( 1x PBS , 0 . 5% Glycerol , 100 μM PMSF ) . The next day the solution was isolated from the dialysis cassette and concentrated by using Amicon Ultra centrifugal filter devices ( UFC910024 , Millipore , 100 . 000 MWCO ) . The protein was run on a Coomassie gel ( 0 . 1% Coomassie Brilliant Blue R250 ( Serva ) ) and protein concentration was calculated by Bradford using Roti-Quant ( K015 . 2 , Roth ) . To produce the monoclonal antibodies against vIRF2/K11 C57BL/6J mice were immunized subcutaneously ( s . c . ) and intraperitoneally ( i . p . ) with a mixture of ~ 50 μg purified GST-tagged KSHV K11 protein , 5 nmol CpG ( TIB MOLBIOL ) and an equal volume of incomplete Freund's adjuvant ( Sigma ) . After 6 weeks , a boost without adjuvant was given i . p . and s . c . 3 days before fusion . Fusion of the myeloma cell line P3X63-Ag8 . 653 ( CRL-1580 , ATCC ) with the immune spleen cells was performed according to the standard procedure described by Koehler and Milstein [109] . After fusion , the cells were plated in 96 well plates using RPMI 1640 with 20% fetal calf serum ( FCS ) , 1% penicillin/streptomycin , 1% glutamine , 1% sodium pyruvate , 1% non-essential amino acids , 2% HCS ( Capricorn ) and 1% HT supplement ( Thermo Fisher ) . After 10 days , hybridoma supernatants were tested in an ELISA on plates coated with vIRF2/K11 protein ( 4 μg/ml ) . After blocking with 1x PBS/2% FCS , hybridoma supernatants were added for 30 min . After one wash with 1x PBS , bound antibodies were detected with a cocktail of HRP-conjugated mAbs against the four mouse IgG isotypes . HRP was visualized with ready to use TMB ( 1-StepTM Ultra TMB-ELISA , Thermo Fisher ) and the absorbance was measured at 650 nm with a microplate reader ( Tecan ) . The hybridoma cells of vIRF2/K11-reactive supernatants were cloned at least twice by limiting dilution . Experiments in this study were performed with hybridoma culture supernatant of vIRF2/K11 clone #30F9 and #31A2 ( mouse IgG2b/κ ) at a dilution of 1:10 in PBS-T . The synthesis of the peptide array ( SPOT synthesis ) was carried out with an Intavis MultiPep automated SPOT array synthesizer ( Intavis Bioanalytical Instruments , Cologne , Germany ) on an amino-PEG functionalized SPOT synthesis paper membrane ( AIMS Scientific Products , Berlin , Germany ) with a size of 9x13 cm based on a published procedure [110] . The entire original protein sequence of K11 was divided into 169 overlapping peptides with a length of 15 amino acids and a shift of three amino acids for consecutive sequences . On each peptide position , ß-alanine was first coupled to the paper by using Fmoc-ßAla-OPfp ( 0 . 3 M ) and HOBt ( 0 . 3 M ) in NMP , to which 10 μl/ml diisopropylcarbodiimide was added . The activated amino acid was added at 0 . 2 μl to each position , reaction time was 45 min . Afterwards the free amino groups on the membrane were acetylated for 1 h with 2% acetic acid anhydride in DMF ( capping solution ) . The Fmoc group from the ß-alanine was cleaved by an 8 min treatment with 20% piperidine in DMF . After washing with DMF , the SPOT peptide positions were stained with bromophenol blue ( 2% of the ethanol stock solution in DMF ) , followed by washing of the membrane with ethanol and drying . The corner positions of the array were marked with a pencil for a later identification of the peptide positions . The peptide sequences were assembled by utilizing Fmoc amino acid derivatives ( 0 . 2 M in NMP ) with preactivation for 30 min by equimolar amounts of diisopropylcarbodiimide and hydroxybenzotriazol . 3 μl/ml of a stock solution of 10 mg/ml bromophenol blue in ethanol were added to each amino acid solution , which allowed the monitoring of the progress of the coupling reactions by observing the colour change from blue to green . The activated amino acids were spotted three times to each position at 0 . 2 μl . After the third spotting the reaction was allowed to proceed for a further 30 min , thereafter the membrane was washed with DMF and capping was carried out for 7 min . After final assembly of the peptide chains and cleavage of the terminal Fmoc group , the free N-terminus was acetylated with the capping solution for 15 min . The membranes with the completed arrays were washed with DMF , ethanol and dried . Side chain protection groups were cleaved by two consecutive 2 h treatments with trifluoroacetic acid containing 5% dichloromethane , 3% diisopropyl silane and 2% water . Afterwards , the membranes were washed with dichloromethane and ethanol and finally dried . The ready-to-use peptide arrays were sealed in plastic foil and stored at -20°C until usage . To map the epitope of the two K11 antibody clones #30F9 and #31A2 , the membranes were wetted with a few drops of ethanol and blocked in 5% milk in PBS-T buffer for 2 h at RT . The membranes were incubated with the antibody clones #30F9 or #31A2 ( diluted 1:10 in 5% milk PBS-T ) over night at 4°C gently mixing . After washing the membranes three times in PBS-T , they were incubated with the secondary goat anti-mouse IgG antibody labeled with the IRDye 800CW for 1 h at RT in the dark . The membranes were developed in a LI-COR Odyssey after washing three times in PBS-T . Cells were lysed after washing in 1x PBS in 1x SDS lysis buffer ( 62 . 5 mM Tris-HCl pH 6 . 8 , 2% ( W/V ) SDS , 10% ( V/V ) Glycerol , 50 mM DTT , bromophenol blue ) . To pellet down cell debris , the samples were centrifuged after lysis for 10 min at 13 , 000 rpm at 4°C . If necessary , cell lysates were sonicated for a few seconds before centrifugation . The protein concentrations were measured using the spectrophotometer NanoDrop1000 ( Peqlab ) . Cleared cell lysates were boiled for 5 min at 95°C and loaded on 8–12% SDS polyacrylamide gels . As a protein marker the Precision Plus Protein All Blue Prestained Protein Standards ( 1610373 , Biorad ) was used . After SDS PAGE , the proteins were transferred on a nitrocellulose membrane ( Premium 0 . 45 μm , 10600003 , Amershan ) for 70 min at 350 mA in transfer buffer . Depending on the antibody , unspecific binding was blocked by incubating the membrane in either 5% milk in PBS-T or TBS-T buffer or in 5% bovine serum albumin ( BSA ) in TBS-T buffer . The membranes were incubated with the primary antibody on a roller at 4°C over night or 1 h at RT . After three washing steps in the corresponding buffer , the membranes were incubated for 1 h at RT in the secondary horseradish peroxidase ( HRP ) -conjugated antibody . Primary and secondary antibodies used for western blot are listed above ( section antibodies ) . To visualize the specific detection of proteins , the membranes were developed in a LAS-3000 Imager ( Fujifilm ) using either self-made enhanced chemiluminescence ( ECL ) solution 1 and 2 , which were mixed in a ratio of 1:1 , or the SuperSignal West Femto Maximun Sensitivity Substrate ( 34096 , Thermo Scientific ) . To isolate nuclear and cytoplasmic cellular fractions , the NE-PER Nuclear and Cytoplasmic Extraction Reagents from ThermoScientific ( 78833 ) were used following the manufacturer’s protocol for a packed cell volume of 50 μl . To insert stop codons into the vIRF2 gene in the KSHV genome , competent E . coli GS1783 cells with chromosomally encoded inducible Red- and I-SceI expression were prepared by inoculation of 50 ml LB medium with an overnight culture ( 1:25 ) and incubated for 3 h at 32°C . The culture was shaking at 42°C for 15 min , followed by an incubation of 20 min while shaking in an ice bath . The bacteria were centrifuged for 10 min at 4 , 000 rpm at 4°C and the pellet was washed twice with sterile H2O and once with sterile ice cold 10% glycerol solution . The final pellet was resuspended in 600 μl 10% glycerol , aliquoted at 50 μl and the bacteria were either used directly or stored at -80°C . In the first recombination step of the En passant mutagenesis , first described by Tischer et al . [111] , the PCR amplicon of a kanamycin resistance gene from the pOri6K . I-SceI vector ( kindly provided by Martin Messerle , Medical School Hannover ) with an integrated I-SceI cleavage site with homologous flanking sequences carrying the mutation was inserted in competent E . coli GS1783 by electroporation ( by using 0 . 2 cm Gene Pulser cuvettes ( Bio-Rad ) at 2 . 5 kV , 25 μF and 200 Ohm ) . The primers used for the different mutants are listed in the supplement ( S1 Table ) . For DNA amplification the Phusion high-fidelity polymerase system from NEB ( M0530L ) was used . To amplify the kanamycin cassette for the Tischer mutagenesis , 0 . 2 μM of the forward primer were directly added to a 25 μl reaction whereas the same amount of the reverse primer was added after 17 PCR cycles in a Veriti 96 well thermal cycler from Applied Biosystems . The electroporated bacteria were selected on kanamycin and chloramphenicol and resistant clones were checked with restriction analysis . The second recombination step aims to remove the kanamycin resistance again such as to leave only the introduced mutations . LB medium with chloramphenicol only was inoculated with of an overnight culture ( 1:50 ) and incubated for 3 h at 32°C in a shaker at 220 rpm . After adding 1% L-arabinose to induce I-SceI expression , the culture was incubated for 1 h at 32°C and transferred afterwards into a 42°C waterbath shaker for 25 min . After the heat shock , the culture was returned to 32°C for 3 h . A negative selection process revealed final clones which showed kanamycin sensitivity . The BAC16 . KSHV . ΔvIRF2 mutant was generated differently by using a galK/Kan-based selection in E . coli SW102 [112] essentially as previously described [113] . Briefly , a PCR amplified galK/Kan cassette , carrying 50 bp homologous regions flanking the vIRF2 gene ( primers see S1 Table ) was electroporated into E . coli SW102 carrying the BAC16 . KSHV . WT ( by using a 0 . 1 cm cuvette ( BioRad ) at 25 μF , 1 . 75 kV and 200 Ω ) . The electroporated bacteria were selected on McConkey plates with kanamycin and chloramphenicol . The integrity of all mutants , and the presence of introduced mutations , was confirmed by restriction analysis and final clones were verified by sequencing the entire KSHV BAC by Next Generation Sequencing . Briefly , purified BAC DNA obtained from a maxi preparation was sheared by sonication . To avoid bias by over-amplification , library preparation was performed using the KAPA real-time library preparation kit ( KAPA Biosystems , Wilmington , MA , USA ) with a limited number of PCR cycles . Quality controlled libraries were sequenced on a MiSeq ( Illumina ) using reagent kit v3 to generate 2 x 300 base paired-end reads . Reads were mapped to the KSHV BAC16 parental strain and variants were identified by using the low frequency variant detector function in CLC genomics Workbench v9 . All the KSHV BAC16 mutants used in this study showed only the mutations introduced by mutagenesis and did not contain any additional changes in their genomic sequence . Custom designed siRNAs for vIRF2 and for PML [114] were purchased from Thermo Scientific Dharmacon ( vIRF2 ( 1 ) : CGGAAUGGCUCACGGACUU; vIRF2 ( 2 ) : UUUCGCUGUCACUCGAUUCUU; vIRF2 ( 3 ) : UUCUUCGCGAUGCAUUUCCUU; PML: AGAUGCAGCUGUAUCCAAGUU [114] ) . For IFIT1 and IFIT2 a pool of four siRNAs for each target gene was purchased from Thermo Scientific Dharmacon ( IFIT1: M-019616-01-0005; IFIT2: M-012582-01-0005 ) . For IFIT3 a pool of three siRNAs was purchased from Santa Cruz ( sc-75326 ) . Upon delivery , siRNA pellets were resuspended in nuclease free water ( P1193 , Promega ) following the manufacturer’s protocol ( Thermo Scientific Dharmacon ) to a final concentration of 100 μM . The siRNA was aliquoted and stored at -80°C . A non-targeting siRNA pool was used as a negative control ( Pool 2 D-001206-14-20 , Thermo Scientific Dharmacon ) . For transfection of siRNA into HuARLT . rKSHV . 219 or HUVECs the Neon Transfection System ( ThermoFisher , MPK5000 ) was used . The transfection of 1x105 cells with 150 pmol siRNA was performed using three single pulses of 1350 V each for 30 msec . Total RNA was isolated by using the RNeasy Mini Kit ( 74104 , Qiagen ) with an additional on-column DNase digestion . For qPCR , cDNA was synthesized using the expand reverse transcriptase ( 11785826001 , Roche ) according to manufacturer’s instructions . Briefly , 1 μg of total RNA and 200 pmol of ( dT ) primer ( MWG-Operon ) and PCR grade H2O ( Qiagen ) was mixed in a 11 . 5 μl reaction . The mixture was incubated at 65°C for 10 min in a thermocycler , immediately cooled on ice and afterwards the following components were added: 1x Reverse expand transcriptase buffer , 10 mM DTT , 1 mM of each dNTP , 20 U of recombinant RNAsin ( Promega ) , and 50 U recombinant expand reverse transcriptase to make a total volume of 20 μl . The mixture was then incubated at 43°C for 1 h and the synthesized cDNA was used for subsequent qPCR . To quantify PML mRNA expression in HUVEC , qPCR amplification was performed in a 10 μl reaction volume using Taqman universal PCR master mix according to manufacturer’s recommendations ( Applied Biosystems Cat . No 4364341 ) . The following Taqman gene expression assays were used: PML- Hs 00231241; GAPDHHs 02758991 and qPCR was performed in a Stratagene MX3000P thermocycler using the following thermo profile: hold at 50°C for 2 min , followed by initial denaturation at 95°C for 20 sec , followed by 40 amplification cycles each at 95°C for 3 sec and 60°C for 20 sec period . To detect PML NBs , HeLa cells were plated on glass coverslips ( 2x105 cells per well of a 24 well plate ) transfected with the vIRF constructs or the control vector 24 h later and 48 h later cells were washed in 1x PBS and fixed with 4% paraformaldehyde ( PFA ) for 20 min at RT . For immunofluorescence in HUVECs , cells were plated on glass coverslips ( 2x105 cells per well of a 24 well plate ) and infected the following day with rKSHV . 219 ( MOI 20 ) or HCMV ( MOI 3 ) . After fixation , the cover slips were washed three times with 1x PBS and cells were permeabilized in 0 . 2% triton/PBS for 10 min at RT . Unspecific binding was blocked by incubating the cells in 0 . 5% BSA for 30 min at 37°C after additional washing for three times in 1x PBS . The primary antibody for PML was diluted 1:200 in 0 . 5% BSA/PBS and cells were incubated for 30 min at 37°C . After washing the coverslips three times in 1x PBS , they were incubated with the secondary antibodies diluted 1:200 in 0 . 5% BSA/PBS for 1 h at 37°C . During the incubation with the secondary antibody the DAPI staining was performed in parallel . After this , the cells were washed twice in 1x PBS and once in ddH2O and mounted on slides in 30 μl moviol supplemented with DABCO . The slides were dried overnight in the dark at RT and images were taken with a ZEISS AxioObserver microscope . The Microarray analysis was performed by the Research Core Unit Transcriptomics ( RCUT ) of Hannover Medical School . The Microarray utilized in this study represents a refined version of the Whole Human Genome Oligo Microarray 4x44K v2 ( Design ID 026652 , Agilent Technologies ) , called ‘054261On1M’ ( Design ID 066335 ) developed at the Research Core Unit Transcriptomics ( RCUT ) of Hannover Medical School . Microarray design was created at Agilent’s eArray portal using a 1x1M design format for mRNA expression as template . All non-control probes of design ID 026652 have been printed five times within a region comprising a total of 181560 Features ( 170 columns x 1068 rows ) . Four of such regions were placed within one 1M region giving rise to four microarray fields per slide to be hybridized individually ( Customer Specified Feature Layout ) . Control probes required for proper Feature Extraction software operation were determined and placed automatically by eArray using recommended default settings . Synthesis of cRNA was performed with the ‘Quick Amp Labeling kit , no color’ ( #5190–0447 , Agilent Technologies ) , except that the NTP-mix ( to be used in the T7 reaction ) was exchanged for a mix composed of 15 mM of each ATP , CTP , GTP , 11 . 25 mM of UTP , and 3 . 75 mM of aaUTP ( final concentration of each nucleotide was 1 . 875 mM ) . The labeling of aaUTP-cRNA was performed by use of the Amino Allyl MessageAmp II Kit ( #AM1753; Life Technologies ) and Alexa Fluor 555 Reactive Dye ( #A32756; LifeTechnologies ) . cRNA fragmentation , hybridization and washing steps were carried-out as recommended in the ‘One-Color Microarray-Based Gene Expression Analysis Protocol V5 . 7’ ( Agilent ) , except that 1300 ng of each labelled cRNA population were used for hybridization . Slides were scanned on the Agilent Micro Array Scanner G2565CA ( pixel resolution 3 μm , bit depth 20 ) . Data extraction and processing of raw fluorescence intensity values were performed with the ‘Feature Extraction Software V10 . 7 . 3 . 1’ using the extraction protocol file ‘GE1_107_Sep09 . xml’ , except that ‘Multiplicative detrending’ algorithm was inactivated .
|
The life cycle of Kaposi Sarcoma herpesvirus involves both persistence in a latent form and productive replication to generate new viral particles . How the virus switches between latency and productive ( ‘lytic’ ) replication is only partially understood . Here we show that a viral homologue of interferon regulatory factors , vIRF2 , antagonizes lytic protein expression in endothelial cells . It does this by inducing the expression of cellular interferon-regulated genes such as IFIT 1–3 , which in turn dampens early viral gene expression . This observation suggests that vIRF2 allows KSHV to harness the interferon pathway to regulate early viral gene expression in endothelial cells .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"gene",
"regulation",
"pathogens",
"endothelial",
"cells",
"microbiology",
"epithelial",
"cells",
"viruses",
"protein",
"expression",
"dna",
"viruses",
"molecular",
"biology",
"techniques",
"herpesviruses",
"research",
"and",
"analysis",
"methods",
"small",
"interfering",
"rnas",
"animal",
"cells",
"proteins",
"medical",
"microbiology",
"gene",
"expression",
"microbial",
"pathogens",
"biological",
"tissue",
"kaposi's",
"sarcoma-associated",
"herpesvirus",
"viral",
"replication",
"molecular",
"biology",
"lytic",
"cycle",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"gene",
"expression",
"and",
"vector",
"techniques",
"biochemistry",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"anatomy",
"virology",
"viral",
"pathogens",
"genetics",
"epithelium",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"interferons",
"non-coding",
"rna",
"organisms"
] |
2019
|
Kaposi’s sarcoma-associated herpesvirus vIRF2 protein utilizes an IFN-dependent pathway to regulate viral early gene expression
|
The Arabidopsis fruit mainly consists of a mature ovary that shows three well defined territories that are pattern elements along the mediolateral axis: the replum , located at the medial plane of the flower , and the valve and the valve margin , both of lateral nature . JAG/FIL activity , which includes the combined functions of JAGGED ( JAG ) , FILAMENTOUS FLOWER ( FIL ) , and YABBY3 ( YAB3 ) , contributes to the formation of the two lateral pattern elements , whereas the cooperating genes BREVIPEDICELLUS ( BP ) and REPLUMLESS ( RPL ) promote replum development . A recent model to explain pattern formation along the mediolateral axis hypothesizes that JAG/FIL activity and BP/RPL function as antagonistic lateral and medial factors , respectively , which tend to repress each other . In this work , we demonstrate the existence of mutual exclusion mechanisms between both kinds of factors , and how this determines the formation and size of the three territories . Medial factors autonomously constrain lateral factors so that they only express outside the replum , and lateral factors negatively regulate the medially expressed BP gene in a non-autonomous fashion to ensure correct replum development . We also have found that ASYMMETRIC LEAVES1 ( AS1 ) , previously shown to repress BP both in leaves and ovaries , collaborates with JAG/FIL activity , preventing its repression by BP and showing synergistic interactions with JAG/FIL activity genes . Therefore AS gene function ( the function of the interacting genes AS1 and AS2 ) has been incorporated in the model as a new lateral factor . Our model of antagonistic factors provides explanation for mutant fruit phenotypes in Arabidopsis and also may help to understand natural variation of fruit shape in Brassicaceae and other species , since subtle changes in gene expression may cause conspicuous changes in the size of the different tissue types .
The fruit , a pivotal structure in angiosperms , is the specialized plant organ that develops from the gynoecium after fertilization of the ovules . The very term angiosperm comes from the Greek and means “seeds enclosed in a vessel” ( angion , vessel , and sperma , seed ) , describing the main functions of this organ: seed protection and dispersal . Our present knowledge on fruit development principally derives from research in the crucifer Arabidopsis thaliana , Arabidopsis hereafter [1]–[7] . All the tissues of the Arabidopsis fruit are already present in the bicarpelate pistil , whose development is initiated as a group of cells that form a dome-shaped primordium . Subsequently , polarity is determined along the main axes of symmetry giving rise to pattern elements with specific tissue types . Thus , for instance , along the apical-basal axis both pistils and fruits show , from bottom to top , the basal gynophore , the ovary , the style and the apical stigma ( Figure 1A ) . The dehiscent fruit of Arabidopsis is essentially an expanded ovary encompassing the seeds [8] , and consists of three different territories that constitute the pattern elements along the mediolateral axis . The replum , located at the medial plane of the flower , is a narrow structure that separates two lateral valves . At the valve-replum boundary , the valve margin , another lateral tissue , comprises a few rows of small and rounded cells ( Figure 1A–1C ) . Ripening of the fruit involves the formation of a dehiscence zone in the valve margin and the consequent detachment of the valves from the replum that precedes seed dispersal [9] . The MADS-box gene FRUITFULL ( FUL ) [10] and the homeobox gene REPLUMLESS ( RPL , aka BELLRINGER , PENNYWISE , LARSON and VAAMANA ) [11]–[14] are expressed within the valve and replum tissues , respectively . RPL and FUL , in their corresponding domains of activity , prevent the ectopic expression of the valve margin identity genes SHATTERPROOF1 and 2 ( SHP1 , SHP2 ) [15] , INDEHISCENT ( IND ) [16] and ALCATRAZ ( ALC ) [17] . This regulation ensures the correct formation of valves and replum territories and limits the expression of the valve margin identity genes to the valve-replum boundaries . Thus , in fruits completely lacking both FUL and RPL activities , valves and replum epidermal cells acquire valve margin identity as a consequence of the ectopic expression of valve margin identity genes [13] , [16] , [18] . Early in pistil development , medial tissues form two internal ridges that fuse to form the septum and the placenta , suggesting that presumptive repla have meristematic properties [5] , [19] , [20] . Accordingly , they exhibit expression of meristem genes , as RPL [13] and the class I KNOTTED1-like homeobox ( KNOX ) genes BREVIPEDICELLUS ( BP , aka KNAT1 ) [21]–[24] and SHOOT MERISTEMLESS ( STM ) [25] . Different from the replum , valves show a more leaf-like nature , and consequently , they express genes with crucial roles in leaf development , as the YABBY1 ( YAB1 ) group genes FILAMENTOUS FLOWER ( FIL ) and YABBY3 ( YAB3 ) [26]–[30] , JAGGED ( JAG ) , which codes for a transcription factor with a single C2H2 zinc-finger domain [31] , [32] , and the MYB transcription factor-encoding gene ASYMMETRIC LEAVES1 ( AS1 ) [33] , [34] . FIL , YAB3 and JAG positively regulate the expression of FUL and valve margin identity genes , so that the cooperating activities of these three genes in ovaries have been called JAG/FIL activity . Valve margin identity genes are activated in places close to the presumptive replum where the levels of this activity are low , whereas higher levels activate FUL expression in valves [35] . Replum and valves apparently mirror the antagonistic relationships between meristem and leaves [19] , [20] . Thus , AS1 prevents the ectopic expression of BP in valves while RPL impedes that of JAG/FIL activity genes in the replum [19] , [35] . Based on this antagonism , a model has been proposed to account for mediolateral patterning of the ovary , which puts forward that the different tissue types along the mediolateral axis are determined by the opposing activities of two antagonistic factors: valve factors that basically are the genes involved in the JAG/FIL activity , and replum factors that are composed by BP and RPL [19] , [36] , whose products dimerize to migrate into the nucleus [11] , [14] , [37]–[42] . In accordance to the model , replum and valves would form in territories with high activity of replum and valve factors , respectively , whereas the valve margin would develop in a narrow ridge in which both activities would show low levels and overlap [19] . Nevertheless , recent research has shown that valve and replum factor activities do not overlap , since BP and RPL are not active in the valve margin [36] , [43] ( our unpublished results ) . Therefore , BP and RPL will be hereafter referred to as replum or medial factors , whereas genes involved in JAG/FIL activity ( hereafter referred as JAG/FIL activity genes ) will be called lateral ( valve and valve margin ) factors . In this report , we demonstrate that , indeed , both medial and lateral factors are mutually antagonistic , as they repress each other . We have observed that lateral factors negatively regulate in a non-autonomous fashion BP , thus restricting the size of the medial region , an essential condition for proper replum development , whereas medial factors limit in an autonomous way the expression of JAG/FIL activity genes , whose products only are detected outside the replum . Furthermore , we have also found that AS1 collaborates with lateral factors by preventing downregulation of JAG/FIL activity genes by the ectopic expression of BP in lateral regions . Here , we propose a non-overlapping model whereby the opposing activities of medial and lateral factors determine the specification and size of pattern elements along the mediolateral axis of the Arabidopsis fruit . In accordance with this model , an increase in the expression of medial factors and a decrease in lateral factor activities lead to the overproduction of medial tissues along with a large reduction in the size of the lateral domains .
We have previously demonstrated that the MYB transcription factor AS1 regulates patterning along the mediolateral axis of the Arabidopsis fruit . When compared to wild type , in as1 fruit , the replum contains more epidermal cells , increasing its width . This phenotype is accompanied with a reduction in the final size of the valves as the valve epidermal layer contains fewer cells [19] . We previously found that the as1 fruit phenotype was largely associated with the misregulation of BP , because: 1 ) 35S::BP had the same fruit alterations as seen in as1 plants , 2 ) BP was ectopically expressed in lateral regions of as1 pistils and 3 ) in as1 bp fruits , replum and valves almost completely recovered the wild-type size [19] . However , the increase in the number of replum cells is not the only alteration observed in as1 ( or 35S::BP ) repla . Whereas in the wild-type pistils the replum contains long and narrow cells and no stomata structures form ( Figure 1B and Figure 2A ) , a closer inspection of altered repla in as1 and 35S::BP plants revealed , on the contrary , the presence of extra-large cells , as well as a few interspersed stomata ( Figure S1A , S1B ) . These observations indicate that the negative regulation of BP by AS1 is not only essential in regulating the size of pattern elements along the mediolateral axis , but also for the correct specification of replum identity . As mentioned in the introduction , JAG/FIL activity genes [35] have been postulated to be the valve functions ( that we refer in this work as lateral factors ) patterning the mediolateral axis of the fruit in Arabidopsis . Consequently , similar to as1 mutants , a decrease in this activity drastically affects the valves [7] , [19] , [35] . Furthermore , according to our current model , reduced levels of JAG/FIL activity should not only cause a reduction in the size of the valve territory , but also a mutant replum phenotype consisting in increased width [7] , [19] . Fitting with this hypothesis , we observed that fruits in jag and fil plants , besides their defects in lateral regions [35] , clearly exhibited oversized repla ( Figure 2B , 2C ) . Moreover , a close inspection of the replum surface by SEM revealed the presence of stomata in both fil and jag repla ( Figure S1C , S1D ) . These abnormalities were even more dramatic when the JAG/FIL activity was further reduced , as for example in fil yab3 or fil jag backgrounds ( Figure 2D , 2E and Figure S1E , S1F ) . Because of the similarities between these defects and the ones described before for as1 or 35S::BP repla [19] , we investigated whether the lack of BP was capable of rescuing the fruit phenotypes of mutants affected in the JAG/FIL activity . Indeed , fil bp , jag bp , fil yab3 bp and fil jag bp fruits showed narrow repla and contained no replum stomata ( Figure 2F–2I ) . These observations suggest that JAG/FIL activity regulates the expression of BP in the Arabidopsis fruit and that misregulation of BP is essential to produce the repla defects seen in mutants affected in this activity . The phenotypic similarities of fruits in mutants affected in JAG/FIL activity genes to those of as1 and their rescue by bp led us to investigate whether BP was also negatively regulated by JAG/FIL activity in Arabidopsis ovaries as it is by AS1 [19] . Interestingly two members of this activity , the YAB1 group genes FIL and YAB3 , have been previously described to repress BP in leaves [44] . But so far no evidence indicates that this control also occurs in fruits . We therefore analyzed the expression of the BP::GUS reporter construct in mutant backgrounds affected in JAG/FIL activity . In wild-type ovaries , BP::GUS expression is primarily detected in the medial region , corresponding to the replum ( Figure 3A ) [19] , [43] . When the JAG/FIL activity was compromised , we observed that the intensity of the BP::GUS signal increased and its expression domain expanded , achieving the widest domain in the fil yab3 jag triple mutant ( Figure 3B–3F ) . The exception was the yab3 single mutant , in which the behavior of the BP::GUS reporter was indistinguishable from that of the wild type ( data not shown ) . Nevertheless , we observed by qRT-PCR ( quantitative real-time polymerase chain reaction ) a significant increase in the expression levels of BP transcripts in the pistils of all backgrounds affected in JAG/FIL activity , including yab3 ( Figure 3G ) . Therefore , these results and those shown in the previous section indicate that the JAG/FIL activity , functioning in valves and valve margins , negatively regulates BP expression in medial domains and that this repression is required for the correct specification of the replum . However , these data pose the question of whether the increased expression of BP in mutant backgrounds affected in JAG/FIL activity simply reflects the augmented sizes of the corresponding repla . Contrary to this line of reasoning , the GUS signal in repla of such mutants is not only wider than that of wild type but also more intense ( Figure 3 ) , suggesting that the increase in replum width is not the only cause of the higher levels of BP expression in mutant pistils . To further address this issue , we tested , by qRT-PCR , the expression levels of another replum gene , RPL , in multiple genetic conditions with impaired JAG/FIL activity and lacking BP function ( Figure S2 ) . In the resulting mutants , repla show reduced width as compared to the same backgrounds but with unaltered BP activity ( Figure 2 ) . Levels of RPL transcripts in wild-type and bp pistils were quite similar , indicating that loss of BP function has little effect on RPL expression . However , in pistils of fil yab3 bp and fil jag bp , RPL expression was significantly higher than in those of both the wild type and the bp mutant , despite the moderate width of the repla in the two triple mutants ( Figure S2 ) . Therefore , enhanced expression of RPL , and most likely of BP , in such mutant backgrounds does not exclusively depend on replum size , supporting again the negative regulation of JAG/FIL activity on replum genes . The model for mediolateral fruit patterning hypothesizes that lateral factors repress medial factors and vice versa [19] . Fitting with the model , we have found that JAG/FIL activity negatively regulates BP . Therefore we decided to study if there exists such a reciprocal repression . If that were the case , BP would negatively regulate JAG/FIL activity [19] . To test this prediction of the model , we made use of genetic backgrounds in which BP was misregulated . As BP is ectopically active in fruit valves of as1 mutants [19] we therefore first examined the expression of JAG/FIL activity genes in as1 pistils . We tested by mRNA in situ hybridization the expression pattern of FIL in wild-type and as1 gynoecia . As previously published [28] , [29] , [35] , we found that the FIL mRNA is located in lateral domains of wild-type pistils ( Figure 4E ) . However in as1 pistils the transcript of FIL was detected with less intensity and in a more reduced territory ( Figure 4F ) . This decay in FIL activity was also seen when the FIL::GFP reporter was assayed in as1 pistils ( Figure 4I , 4J and Figure S3A , S3B , S3D , S3E , S3G , S3H ) . A similar behavior was seen when the expression of JAG was monitored using a transgenic GUS-reporter line . In wild-type ovaries JAG::GUS signal is exclusively localized in lateral regions , while in as1 , although the signal is detected in the same region , the levels of GUS activity were conspicuously lower ( Figure 4L , 4M ) . Interestingly , when compared to wild type , 35S::BP plants produced flowers with fewer and narrower petals ( Figure S4A , S4B ) , virtually phenocopying fil mutants ( Figure S4C ) . These observations suggest that FIL activity might be severely compromised in 35S::BP plants . We , therefore , studied the expression pattern of FIL in 35S::BP pistils by in situ hybridization , being unable to detect any signal of FIL transcripts ( Figure 4G ) . We also analyzed the FIL::GFP reporter in 35S::BP gynoecia and observed a drastic reduction in GFP signal when compared to those of wild-type plants ( Figure 4K and Figure S3C , S3F , S3I ) . Unlike the result of the in situ hybridization , in which no FIL expression was detected in 35S::BP ovaries , the reporter provided a slight but perceptible signal , possibly because of a higher sensitivity in the detection of GFP . All together these data strongly suggest that ectopically expressed BP , directly or indirectly , downregulates JAG/FIL activity genes in ovaries . Despite this result , 35S::BP fruits exhibited normal expression of both the ful-1 enhancer trap ( FUL::GUS ) and the SHP2::GUS construct ( Figure S4D , S4E ) To further investigate how ectopic BP expression affects the fruit , we made use of transgenic FIL>>BP plants , in which the BP coding region is transcribed in the FIL expression domain . For this condition , the model predicts that the expression of BP in lateral domains should counteract the JAG/FIL activity , affecting not only this tissue , but also the replum that would acquire a larger size . As expected , FIL>>BP fruits were strikingly similar to those of 35S::BP and as1 plants , with oversized repla and reduced valves ( Figure 4A–4D ) . Accordingly , in FIL>>BP pistils , we were not able to detect FIL transcripts by in situ hybridization ( Figure 4H ) . Our qRT-PCR mRNA quantification also showed that in both 35S::BP and as1 pistils JAG/FIL activity genes were downregulated , according to the results presented above ( Figure 4N ) . Remarkably , higher relative levels of FIL messenger were detected in as1 when compared to 35S::BP , which might be explained by the stronger expression of BP in 35S::BP than in as1 background ( our unpublished results ) . Therefore , all the results presented so far indicate that JAG/FIL activity represses BP , which in turn , negatively regulates the JAG/FIL activity genes . These data further confirm that both sets of factors are mutually antagonistic in the mediolateral axis of the Arabidopsis fruit . The strong similarities between AS1 and JAG/FIL activity in negatively controlling BP expression in fruits led us to generate multiple loss-of-function mutant combinations affected in both functions to reveal the contribution of these genes to mediolateral patterning of fruits . Because of the phenotypic similarities between as1 and plants misexpressing BP , we also crossed 35S::BP plants to mutants affected in JAG/FIL activity . These sets of genetic combinations helped us to test whether the presumable fruit defects generated when as1 and mutations in JAG/FIL activity genes combine are exclusively attributable to BP misexpression . Fruits of 35S::BP jag plants showed a slight increase in replum size and more reduced valves when compared to those of 35S::BP or jag backgrounds ( Figure 5A , 5D , 5F ) . As seen in fil jag fruits , although less frequently , we also found stripes of valve margin tissue at the upper position of the lateral-most region of 35S::BP jag valves ( Figure 5D and Figure S5A , S5E ) . The similarity between 35S::BP jag and fil jag fruit alterations can be explained by the negative regulation of BP on YAB1 group genes . Surprisingly , fruits of as1 jag mutants appeared by far more affected than those of 35S::BP jag plants , showing strong reduction of valve size , as well as enlarged and twisted repla ( Figure 5B , 5C , 5G; Figure S6; and Figure S7D , S7H ) , a phenotype reminiscent to that of ful mutants [10] . However , whereas ful valves show small and rounded epidermal cells , and do not contain any stomata , valves of as1 jag fruits exhibited larger cells and stomata . In line with this phenotype , the activity of FUL::GUS in as1 jag pistils was detected in the reduced valves ( Figure 5E ) , explaining the low levels of FUL expression detected in this background ( Figure S8 ) . Since in ful fruits the valve margin identity genes become ectopically expressed in valve tissue , we studied the activity of the SHP2::GUS reporter in as1 jag fruits . Our previous work showed that this reporter expresses normally in as1 fruits [19] . Whole-mount staining of as1 jag fruits revealed normal expression for the SHP2 reporter in the valve margin , but an expansion of the signal towards the lateral domains was detected at the upper part of the valve ( asterisk in Figure 5H ) , consistent with an enlargement of the valve margin in this area ( asterisk in Figure 5C ) . The phenotypic difference between 35S::BP jag and as1 jag fruits strongly suggests that , besides BP , AS1 and JAG likely cooperate in negatively regulating other genes for mediolateral fruit patterning . It has been previously established that both AS1 and JAG interact to promote sepal and petal development by downregulating the boundary-specifying genes CUP-SHAPED COTYLEDONS1 ( CUC1 ) , CUC2 and PETAL LOSS ( PTL ) [45] . The LATERAL ORGAN FUSION1 ( AtMYB117/LOF1 ) gene [46] was also considered as an additional candidate , since its ectopic expression results in enlargement of the replum [47] , similar to that of as1 , 35S::BP or jag fruits . Therefore , we determined by qRT-PCR the transcript levels of these genes in wild-type , as1 jag and 35S::BP jag pistils ( Figure 5I ) . As previously reported , PTL is not expressed in pistil tissues [48] and basically no transcripts were detected in any of the tested backgrounds ( data not shown ) . Transcript levels of AtMYB117/LOF1 were downregulated in both as1 jag and 35S::BP jag pistils ( Figure 5I ) , which ruled out this candidate . But , interestingly , CUC1 and CUC2 transcripts accumulated at much higher levels in as1 jag than in 35S::BP jag pistils ( Figure 5I ) , suggesting that CUC genes might be involved in the strong phenotype found in as1 jag double mutant fruits . Interestingly , Ishida and coworkers showed that CUC2 is involved in fruit development [49] and , in line with our hypothesis , we have observed that the cuc2 gain-of-function allele ( cuc2-d ) [50] leads to the formation of short fruits that develop enlarged repla ( Figure S9 ) . We next checked the effect of misregulating BP ( as1 and 35S::BP ) in mutant backgrounds affected in YAB1 group genes . Fruits of 35S::BP fil and 35S::BP yab3 showed a similar phenotype , exhibiting stripes of valve margin tissue developing ectopically at the basal region of the valves , whereas the apical region of the ovary lacked valve margin ( Figure 6A–6D ) . Because these defects were reminiscent of those seen in fil yab3 double mutants [35] ( Figure S5C , S5G ) , it is likely that the negative effect of BP on JAG/FIL activity genes could account for this phenotype . However , as1 fil and as1 yab3 fruits exhibited moderate phenotypes when compared to those of 35S::BP fil and 35S::BP yab3 ( Figure 6E , 6F , 6J ) , but still showed a conspicuous reduction in valve size concomitant with an increase in replum width ( Figure S6 ) . In fact , in as1 fil mutants , the replum epidermis contained larger cells and more stomata than in any of the single mutants ( Figure S7C , S7G ) . In fruits of the sesquimutant fil YAB3/yab3 , a stripe of valve margin tissue often appears in the apical region of the valves [35] ( Figure S5B , S5F ) . Interestingly , we observed the formation of ectopic valve margin tissue in the valves of both as1 fil and as1 yab3 fruits ( Figure 6E , 6F , 6J ) , although with smaller size and lower frequency ( 40% and 30% in as1 fil and as1 yab3 fruits , respectively , versus 90% in fil YAB3/yab3 fruits ) . These observations suggest a further reduction of JAG/FIL activity in both double mutants with respect to fil and yab3 single mutants . It has been previously shown that low levels of FUL activity in fil YAB3/yab3 fruits lead to the ectopic expression of valve margin identity genes in valves [35] . Therefore , we analyzed the expression of FUL and the valve margin identity gene SHP2 in as1 fil and as1 yab3 pistils . By qRT-PCR assays in pistils , we found that levels of FUL transcripts in both double mutants were significantly reduced comparing to those in wild type or in as1 pistils ( Figure S8 ) . In line with the phenotypes above described , FUL::GUS signal in as1 fil fruits was detected at lower levels in the apical regions of valves ( Figure 7A ) , just where SHP2::GUS expresses ectopically ( Figure 7D ) and ectopic valve margin is produced ( Figure 6E , 6J ) . When one copy of yab3 was introduced into the as1 fil background ( as1 fil YAB3/yab3 plants ) , the severity of the mutant phenotype was intensified and fruits exhibited smaller valves and larger repla when compared to those of as1 fil and as1 yab3 double mutants ( Figure 6G , Figure S6 , and Figure S7E and S7I ) . We also noticed that as1 fil YAB3/yab3 and as1 jag siliques were very similar , although in as1 fil YAB3/yab3 the replum had fewer cells and was less twisted ( Figure S7D , S7E , S7H , S7I ) . Similar to mutant fruits affected in JAG/FIL activity genes , repla of as1 fil YAB3/yab3 fruits were abnormally wider , showing more and larger epidermal cells , and also presented frequent interspersed stomata , being quite difficult to distinguish them from valves ( Figure S7E , S7I ) . In this scenario , levels of FUL mRNA were drastically reduced ( Figure S8 ) , and FUL reporter signal was restricted to small areas which correspond to the valves ( Figure 7B ) , while the SHP2::GUS marked the position of the valve margins around the reduced valves ( Figure 7E ) . The complete loss of both AS1 and YAB1 group genes in the as1 fil yab3 triple mutant produced dramatic and deleterious defects on mediolateral fruit patterning ( Figure 6H , 6K , 6L and Figure S6 ) . In the basal region of as1 fil yab3 ovaries , the most prevalent phenotype was the presence of two thin stripes of valve margin located at the lateral-most regions of the ovary , both separating what we called two giant “super-repla” ( Figure 6H , 6L ) . We also found fruits with extremely small valves separated from the oversized repla by valve margin tissue ( Figure S10B ) . The aberrant replum of as1 fil yab3 fruits contained wide and large cells and fully developed stomata , making this tissue to adopt a similar appearance to wild-type valves ( Figure S7F ) . In fact , the phenotype was even stronger in the apical region of the ovary where only these wide and large cells and stomata could be observed , completely lacking valve margin tissue ( Figure 6H , 6K ) . Accordingly , as1 fil yab3 pistils showed very low levels of FUL messenger ( Figure S8 ) and FUL reporter activity was only detected in fruits in which valve tissue developed ( Figure 7C and Figure S10A ) . In line with these observations , the expression of the SHP2::GUS reporter was mostly seen forming a stripe in the lateral-most region of as1 fil yab3 ovaries ( Figure 7F ) . These abnormalities make as1 fil yab3 fruits quite different from those of jag fil yab3 triple mutants , since the former are mainly composed of giant replum , while the latter clearly show valve and replum regions , as the signal for the BP::GUS revealed ( Figure S11 ) . In 35S::BP fil yab3 plants , the fruit mutant phenotype was even stronger , and both apical and basal regions of the ovary showed the same aspect as the apical region of ovaries in as1 fil yab3 fruits ( Figure 6I ) . Our model predicts that an increase in the activity ( or misexpression ) of replum factors ( BP ) along with a reduction in the function of lateral factors ( JAG/FIL activity ) should lead to the formation of fruits with an enormous replum territory and very small valves [19] . The fruit phenotypes described for combinations of as1 and mutant alleles in JAG/FIL activity genes are very much in line with these predictions ( Figure 7G and Figure S6 ) . In strong agreement , as1 fil yab3 and 35S::BP fil yab3 plants produced fruits with huge repla that contained abnormal cell types , and an extreme reduction or abolishment of valve development ( Figure 6H , 6I , 6K , 6L and Figure S10B ) . This phenotype is mainly due to ectopic expression of BP in a background with reduced JAG/FIL activity .
Besides their activities in fruit patterning , JAG/FIL activity genes have been previously described by their participation in the formation of other lateral organs . Consequently , they all are expressed in lateral organs but not in meristematic tissues . The YAB1 group genes , FIL and YAB3 , promote leaf development by repressing the expression of class I KNOX meristematic genes in leaves [44] , and specify ventral ( abaxial ) fate [28] , [29] , [51] . Nevertheless , although FIL and YAB3 are not expressed in meristems , by means of a non-cell-autonomous mechanism , they contribute to shoot apical meristem ( SAM ) maintenance by negatively regulating WUSCHEL ( WUS ) and CLAVATA3 ( CLV3 ) genes , both expressed at the central meristem domain [52] . In fact , this mechanism also affects the floral meristem [52] , and fil yab3 mutants exhibit a high frequency of fruits with three valves ( Figure 3E , 3F ) , possibly due to an increase in floral meristem size caused by the expansion of the WUS expression domain . Similarly , the results presented in this work show that , despite FIL and YAB3 , as well as JAG , are active in lateral regions of the ovary and not expressed in the presumptive replum ( medial tissue ) , mutants affected in JAG/FIL activity have oversized repla with extra-large cells and interspersed stomata , indicating that these laterally expressed genes make an important contribution to the correct development of the medial region in the Arabidopsis fruit . Hence , it is most likely that JAG/FIL activity mediate replum development by negatively regulating , also via non-autonomous mechanisms , the expression of meristematic genes , specifically BP , in the replum . This is deduced 1 ) from the enhanced expression of BP in mutants affected in JAG/FIL activity genes , and 2 ) from the rescue of the replum phenotype in jag bp , fil bp , fil yab3 bp and fil jag bp fruits . Altogether , these data provide an additional analogy between meristem and replum , as well as between lateral organs and valves . Nothing is known about how YAB1 group genes control BP expression at the molecular level , and it has been previously shown in vitro that FIL protein binds DNA nonspecifically [53] . In SAM homeostasis , FIL and YAB3 proteins interact with members of the LEUNIG ( LUG ) and SEUSS-like ( SEU-like ) families of transcriptional co-repressors , and the resulting multicompetent protein complexes likely recruit additional transcriptional regulators to acquire then DNA sequence specificity [54] . It is likely that a similar mechanism might be operating during mediolateral patterning of the Arabidopsis fruit to prevent misexpression of medial factors such as BP . The JAG gene , similarly as YAB3 and FIL , controls leaf polarity and , in cooperation with its closest paralog NUBBIN ( NUB ) , inhibits premature tissue differentiation by maintaining cell proliferation [31] , [32] , [55] . Interestingly , the JAG protein contains an EAR ( ERF-associated amphiphilic repression ) -motif [56] near the N-terminus [31] , [57] . This motif is known to be involved in transcriptional repression and critically intervenes during the molecular interaction between transcriptional regulators and co-repressors [58]–[64] . Therefore , it is possible that FIL/YAB3 and LUG/LUH ( LEUNIG HOMOLOG ) -SEU-like complexes might recruit JAG , and perhaps other regulatory proteins , to target specific DNA sequences . A detailed analysis of this possibility might be of interest and would corroborate , at the molecular level , the genetic interactions that occur for both SAM homeostasis and mediolateral fruit patterning . The relationship between replum and valves closely mirrors the antagonism that there exists between meristem and lateral organs [4] , [19] , [20] . One of such antagonistic relationships is established between class I KNOX genes , expressed in meristem , and AS1 expressed in leaves . In the meristem , the class I KNOX gene STM negatively regulates AS1 whereas , in turn , AS1 physically interacts with AS2 to directly repress BP in leaves [33] , [65]–[68] . Similarly , AS1 ( and AS2 ) also negatively regulates BP in pistils , and thus , in as1 mutants BP is ectopically expressed in valves and show higher levels of expression in the replum [19] . Interestingly , as1 and 35S::BP pistils show similar replum defects as those described for mutants affected in JAG/FIL activity genes , in which BP expression is also enhanced in its own medial domain , and replum defects increase when as1 alleles or 35S::BP construct are combined with jag and/or mutant alleles in YAB1 group genes [19] ( this work ) . These findings indicate that JAG/FIL activity and AS genes cooperate to repress the expression of the replum factor BP in the medial region of pistils , and that this regulation is critical to achieve proper replum pattern . Furthermore , we have observed that valve alterations are also drastically enhanced in these mutant combinations , and our genetic and molecular analyses also evidenced that ectopic expression of BP downregulates JAG and YAB1 group genes in lateral tissues . Therefore , we can conclude that BP repression in lateral regions by AS1 ( and AS2 ) plays an important role in valve development by maintaining normal levels of JAG/FIL activity . Nevertheless , although most of the as1 fruit phenotype can be explained by misregulation of BP , the lack of AS1 does not justify all the fruit defects observed in the mutants . This is better seen in as1 bp fruits , which nearly had wild-type appearance but still showed some subtle abnormalities [19] . This observation indicates that , besides controlling BP expression , AS1 plays additional roles in fruit . The existence of such additional AS1 functions is further supported by the stronger phenotype of as1 jag fruits when compared to those of 35S::BP jag plants . Interestingly , AS1 and JAG also interact in the flower to promote petal and sepal development by negatively regulating the boundary-specifying genes CUC1 and CUC2 [45] . In as1 jag flowers , both sepal and petal development is aborted [45] . However , in 35S::BP jag plants these floral organs develop normally ( data not shown ) . In pistils , our qRT-PCR data revealed that both CUC1 and CUC2 are upregulated in as1 jag at much higher levels than in 35S::BP jag . On the other hand , cuc2 gain-of-function allele produced an increase in replum width that resembles that of as1 and jag mutants . All together these data suggest that AS1 and JAG cooperate to negatively regulate CUC function in fruit and that this repression may play an important role in mediolateral patterning . The basis of the model for mediolateral patterning of the Arabidopsis fruit lies on the antagonistic activities of medial factors ( BP and RPL ) and lateral factors ( JAG/FIL activity genes ) [19] . In accordance to the model , the giant “super-replum” phenotype requires both low levels of JAG/FIL activity and ectopic BP expression in valves . This was the case for as1 fil yab3 or 35S::BP fil yab3 siliques ( this work ) . On the other hand , transformation of the replum into a lateral tissue , the valve margin , requires reduction of medial factor activity and increased activity of lateral factors , as in rpl and rpl bp fruits [13] , [19] , [35] . All together support the idea that BP promotes replum fate [19] , [36] and , again , strongly suggest that medial and lateral factors oppose each other to specify pattern elements along the mediolateral axis . Pattern formation by the contribution of antagonistic activities is not uncommon in plant development . For example , leaf adaxial ( dorsal ) /abaxial ( ventral ) polarity is established by antagonistic interactions between genes that specify either abaxial or adaxial identity , such as KANADI and class III HD-Zip genes [69] , [70] . During embryo development , the apical/shoot versus basal/root polarity is determined by the antagonistic relationship between class III HD-Zip and PLETHORA ( PLT ) genes [71] . The model also proposed that lateral and medial factors work through gradients with their minimal activities in the valve margin , where they likely overlap [19] . This easily allowed to explain the low levels of JAG/FIL activity required to produce valve margin [35] . However , recent studies have determined that BP is only expressed and active in the replum , so that it does not overlap in the valve margin with lateral factors [36] , [43] ( our unpublished results ) . This favours a non-overlapping model whereby the medial factors are not required to function through a gradient . Nevertheless , the low levels of JAG/FIL activity needed for promoting valve margin identity suggest a gradient in the activity of lateral factors . Above a certain threshold lateral factors specify valve fate and allow other genes to function ( such as FUL ) and below that threshold valve margin tissue forms [7] , [19] , [20] , [35] . Furthermore , the phenotypes of as1 and 35S::BP also support the existence of such JAG/FIL activity gradient . Misexpression of BP in these backgrounds reduces the expression of lateral factors , shifting to a more lateral position the region of low levels of JAG/FIL activity that produce valve margin . Farther away from the replum , these levels are high enough to activate the expression of FUL and specify valve development . Consequently , when mutations in AS1 , JAG and YAB1 group genes combine , the more JAG/FIL activity is eliminated , the more laterally the valve margin is placed . This can be easily observed in the basal region of as1 fil yab3 ovaries that exhibit a stripe of valve margin in their lateral-most position . In the model , AS function ( AS1 together with AS2 ) is integrated as another lateral factor ( Figure 8 ) . BP overexpression in the repla of ap2 mutants does not affect valve development [36] , suggesting that expressions of JAG/FIL activity genes are not affected in these backgrounds and that medial factors work in a cell-autonomous way to prevent the ectopic expression of lateral factors in the presumptive replum . These observations further suggest that medial factors are not required to generate the gradient of lateral factors . On the other hand , lateral factors restrict medial factor expression to a small area that becomes the replum and , in a non-autonomous fashion , limit the expression levels of BP and RPL in medial tissues to ensure proper replum development ( Figure 8 ) . In replum tissue , AP2 cooperates with lateral factors to negatively modulate the expression of BP and RPL in the medial domain [19] , [36] . Further work will be needed to elucidate how the gradient of lateral factors is generated , although the phytohormone auxin is a possible candidate . In this sense , it has been postulated that a gradient of auxin patterns the apical-basal axis of the Arabidopsis fruit , with the AUXIN RESPONSE FACTOR3 ( ARF3; aka ETTIN , ETT ) in charge of interpreting intermediate levels of auxin to specify the ovary [72] . Interestingly , mutants affected in JAG/FIL activity genes , both with and without as1 , show phenotypic differences along the apical-basal axis , being the phenotype always stronger in the apical region of the ovary [35] ( this work ) , and it has been shown that ETT positively regulates FIL activity during leaf development [73] , [74] . Furthermore , a recent research found that IND creates an auxin minimum , by regulating auxin efflux , necessary for the formation of the separation layer of the valve margin [74] . The mechanism we propose for patterning the mediolateral axis of Arabidopsis fruit ensures a high plasticity , and possibly may help to understand , at least in part , the variability of fruit shapes in Brassicaceae and other related species . It might be possible that subtle changes in the expression of the antagonistic factors involved in this process could produce drastic changes in the size of the different tissue types . According to this line of argument , Arnaud and coworkers have recently discovered that the reduced replum of Brassica plants is due to a single nucleotide change in a cis-regulatory element between the RPL orthologs of Brassica and Arabidopsis , which makes the Brassica wild-type allele less functional [75] .
The mutant lines used in this work were in Landsberg erecta ( Ler ) background and this accession was the wild-type reference . The original 35S::BP line , in No-0 background , was introgressed four times into Ler . In experiments involving reporter genes ( GUS and GFP ) , the references were wild-type segregants showing the er phenotype , as previously described [35] . fil-8 and yab3-2 [44] , jag-1 [31] , [35] , ful-1 [10] , bp-1 [76] , 35S::BP [22] , as1-104 [19] , FIL>>BP [77] , SHP2::GUS [78] , KNAT1::GUS-18 ( BP::GUS ) [65] and FIL::GFP [30] have been described before . JAG::GUS has been generated by J . R . Dinneny . Briefly , to generate the JAG::GUS transgenic line , a JAG promoter fragment was amplified from the T26J14 BAC using the primers oJD196 ( 5′-AAGCTTCCACTGGGCTTGTATTCCCATCC-3′ ) and oJD197 ( 5′-GGATCCAGTGGGAAATGAGAGATTGGCGTGAG-3′ ) , which added HindIII and BamHI restriction sites to the 5′ and 3′ ends , respectively . This fragment was cloned into the pDW294 binary vector to create the construct pJD145 , which was transformed after checking its integrity into Col-0 plants . Plants were grown at 20–22°C with continuous cool-white fluorescent light as previously described [79] . Multiple mutants were identified among the F2 from the characteristic mutant phenotype caused by individual mutations and/or by molecular genotyping . The fil-8 , yab3-2 and jag-1 alleles were genotyped using primers previously published [31] , [35] ( Table S1 ) . Plants with genotypes showing defective development of stamens and poor fertility were hand-pollinated to allow the formation of fruits . Light microscopy analysis and scanning electron microscopy ( SEM ) were performed as previously described [79] . GFP signal was examined under a Nikon SMZ1500 stereo microscope equipped with a mercury UV lamp , and the emitted fluorescence was monitored using a filter permeable for wavelengths over 505 nm . For GUS staining , samples were treated as previously described [19] , [35] . In situ hybridization was carried out as previously described [19] with the following modifications . The DIG-labeled antisense probe for FIL mRNA was obtained from the original plasmid pY1-Y ( provided by J . Bowman ) , amplifying the insert by PCR with M13 forward and reverse primers . The amplified DNA was used as template to transcribe the probe with a T7 RNA polymerase ( Fermentas ) . RNA from pistils at stages 10–13 was extracted using the PureLink RNA Mini Kit ( Invitrogen ) , and DNA contamination was removed by treatment with DNaseI ( Takara ) . Reverse-transcription was performed from 1 µg of total RNA using the RevertAid H Minus M-MuLV Reverse Transcriptase ( Fermentas ) . Real-time PCR was carried out using the LightCycler FastStart DNA MasterPLUS SYBR Green I ( Roche ) in a volume of 20 µl on the LightCycler 1 . 5 instrument ( Roche ) , as previously published [80] with minor modifications . RNA levels were normalized relative to the constitutive OTC gene [81] and to the wild-type levels , and expression results were calculated by an efficiency correction quantification method [82] . All individual experiments were performed by triplicate , and checked twice using new cDNA every time . The reported values are averages of both biological replicates . Primers for qRT-PCR were as previously published for AtMYB117/LOF1 [47] , BP [83] , CUC1 [84] , CUC2 [85] , FUL [86] and OTC [87] . A complete list of primers used in these experiments can be found in Table S1 . A translation of the title , abstract , and author summary into Spanish is provided in Text S1 .
|
There are three main pattern elements in the mediolateral axis of the Arabidopsis fruit . Two of them , the valves and the valve margins , are placed in lateral positions , while the third , called replum , is located in the medial plane of the flower . The replum expresses meristematic genes ( medial factors ) that specify its development , whereas the function of genes that work in leaves ( lateral factors ) determines the development of valves and valve margins . Consequently , medial and lateral pattern elements of fruits apparently mimic the antagonistic relationships between meristem and leaves . According to this , we propose a model for mediolateral patterning of fruits whereby the mutual opposing activities of medial and lateral factors drive the formation of replum , valves , and valve margins . We conclude that medial factors function in an autonomous fashion to prevent the expression of lateral factors in the replum , and that lateral factors repress medial factors by a non-autonomous mechanism to allow normal replum development . Our model provides explanation for changes in fruit shape in Brassicaceae and related organisms either by mutation within a species or by natural variation among different species .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"plant",
"growth",
"and",
"development",
"plant",
"biology",
"developmental",
"biology",
"plant",
"science",
"model",
"organisms",
"plants",
"pattern",
"formation",
"arabidopsis",
"thaliana",
"plant",
"genetics",
"biology",
"fruits",
"plant",
"and",
"algal",
"models",
"genetics",
"genetics",
"and",
"genomics"
] |
2012
|
Antagonistic Gene Activities Determine the Formation of Pattern Elements along the Mediolateral Axis of the Arabidopsis Fruit
|
The hepatitis C virus ( HCV ) p7 protein is required for infectious virus production via its role in assembly and ion channel activity . Although NMR structures of p7 have been reported , the location of secondary structural elements and orientation of the p7 transmembrane domains differ among models . Furthermore , the p7 structure-function relationship remains unclear . Here , extensive mutagenesis , coupled with infectious virus production phenotyping and molecular modeling , demonstrates that the N-terminal helical region plays a previously underappreciated yet critical functional role , especially with respect to E2/p7 cleavage efficiency . Interrogation of specific N-terminal helix residues identified as having p7-specific defects and predicted to point toward the channel pore , in a context of independent E2/p7 cleavage , further supports p7 as a structurally plastic , minimalist ion channel . Together , our findings indicate that the p7 N-terminal helical region is critical for E2/p7 processing , protein-protein interactions , ion channel activity , and infectious HCV production .
Over 130 million people worldwide are at risk for liver fibrosis , cirrhosis , hepatocellular carcinoma , and end stage liver disease as a result of hepatitis C virus ( HCV ) infection [1] . These complications of infection have made hepatitis C the most common indication for liver transplantation [2] . Further , while novel direct-acting antivirals targeting HCV have dramatically improved clinical outcomes , no vaccine exists to date , and the disease burden is expected to increase over the next decade [3] . HCV is a hepatotropic , plus-strand RNA virus of the Hepacivirus genus and Flaviviridae family [4 , 5] . IRES-mediated translation of the 9 . 6 kb HCV genome yields a single polyprotein that is proteolytically cleaved to produce 10 mature viral proteins that participate in viral replication and assembly of nascent virions [6] . The p7 protein , located at the junction between the structural and non-structural proteins [7] , is a small , 63 amino acid integral membrane protein [8] , predominantly localized to the endoplasmic reticulum ( ER ) [9] . In the context of the HCV life cycle , p7 is dispensable for viral RNA replication [10] but required for infectious virus production [11 , 12] , although it does not appear to be a structural component of the virion nor is it required for HCV glycoprotein-mediated entry [9 , 13 , 14] . Accumulating evidence suggests that p7 orchestrates intracellular viral protein distribution [15–17] , at least in part , via an ( direct or indirect ) interaction with the viral NS2 protein [16 , 18–22] . Additional interactions have been suggested with core at the genetic level [23] and with E2 by immunofluorescence-based colocalization and FACS-FRET methods , although coimmunoprecipitation of p7 with HCV glycoproteins in HCV-replicating cells has yielded disparate results [9 , 24] . Further , yeast two-hybrid and bioinformatically-predicted cellular binding partners have not been further validated [25–28] . Based on the ability of p7 to alter membrane permeability , it has been classified as a viroporin along with HIV-1 vpu and influenza virus M2 , among others ( reviewed in [29] ) . p7 ion channels are sensitive to hexamethylene amiloride [30] , long-alkyl-chain iminosugar derivatives [31] , and–depending on genotype [32 , 33]–amantidine [34] , all of which inhibit cation channel activity in artificial membranes [34 , 35] . The importance of p7 ion channel function for HCV has been demonstrated by correlation of intravesicular pH modulation and infectious virus production in cell culture [36] . This activity has been hypothesized to enable proper glycoprotein folding , protect against premature degradation [37] , or guard against acid-induced conformational changes [14 , 36 , 38 , 39] . Structurally , initial computational modeling predictions [18 , 40] , refined by NMR experiments [22 , 41 , 42] , indicate that p7 monomers adopt a “hairpin-like” topology consisting of an N-terminal helix and “turn” sequence upstream of two transmembrane segments that are connected by a hydrophilic , positively-charged cytosolic loop containing two highly conserved basic resides . The N- and C-termini are oriented towards the ER lumen and may provide a platform for interactions with viral or host proteins [18 , 43] . The intricacy of p7 structure is further complicated by p7 homo-oligomerization . Based on the typical oligomeric structures of viroporins , p7 subunits reside side-by-side in classical hexameric and heptameric models [40 , 42 , 44 , 45] . Molecular dynamic simulation of p7 oligomers , based on the monomeric model put forth by Montserret et al . [41] , suggest that multiple oligomeric states are feasible and that p7 is structurally plastic and may adopt multiple conformations during oligomerization and/or as a function of its lipid environment [44 , 46] . In contrast , the recent NMR structure of hexameric p7 [47] exhibits an unusual architecture where part of each p7 subunit crosses over to interact with all the five other p7 subunits . The resulting rigid structure is reminiscent–albeit comparatively inverted–of single-particle electron micrographs of p7 that depicted a “flower-shaped” architecture [43] . Despite the increasing amount of structural data on p7 , there is no consensus on which conformation ( s ) exist during a natural infection or how structural elements relate to p7 protein-protein interactions , cation selectivity , and ion channel gating . Influenza virus M2 and HCV p7 can partially functionally complement each other [36 , 48] , yet analogy to HIV-1 vpu or influenza virus M2 provides limited mechanistic insight given the divergent structural features and diverse functions described [29] . Modeling of homo-oligomeric assembly [35 , 40 , 44] and electrophysiology experiments [49] indicate that the first transmembrane helix of p7 lines the pore , and the C-terminus ( including TMD2 and unstructured termini ) has been proposed to interact with other proteins [41] . Further , residues potentially involved in cation selectively and gating or intra/intermolecular stability have been postulated [12 , 41 , 47] . However , while mutation of two basic residues , K33 and R35 , within the cytosolic loop , supported their role in ion channel function [36] , none of the putative pore-lining residues studied to date by mutagenesis are essential for p7 ion channeling in vitro [12 , 41 , 49–51] . A comprehensive analysis of residues in key structural regions has not been performed . While the amino acid sequence of p7 is not highly conserved , extensive physico-chemical conservation [41] suggests that the overall p7 structure is similar across genotypes despite variability among individual amino acids . Here , we aimed to probe p7 plasticity and functionality using a combination of mutagenesis and molecular modeling approaches . Our data indicate a critical role for the N-terminal helix region of p7 in modulating E2/p7 cleavage and further support p7 as a structurally plastic , minimalist ion channel through interrogation of specific N-terminal helix residues predicted to point toward the channel pore .
Previous reports indicate that p7 is not required for viral RNA replication but is required for infectious virus production . Modeling data indicate that in addition to the hexameric and heptameric forms of p7 demonstrated experimentally , tetrameric and pentameric oligomers may also exist , at least transiently [44] . To provide biological evidence of p7 structural features and define regions important for functionality , we generated two p7 mutant panels in the context of the J6/JFH infectious clone–one in which an alanine was inserted after every third amino acid throughout the entire length of the protein to perturb p7 structure and a second in which tryptophan substitutions were made throughout the transmembrane domain regions at residues 19–29 , 31–32 , and 36–43 to probe intra- and intermolecular interactions as well as amino acid hydrophilic pore- vs . hydrophobic bilayer-facing orientation ( Fig 1A ) . Mutation of conserved basic residues K33 and R35 in the cytosolic loop was previously shown to impede ion channel activity and block infectious virus production both in vitro and in vivo [11 , 12 , 17 , 36 , 48 , 52]; thus , we excluded these from our analysis . Quantification of cell-associated HCV RNA at 8 and 72 hours post-electroporation indicated that over this time frame all mutants replicated with wild-type ( WT ) efficiency ( Fig 1B ) , exhibiting a mean 495-fold increase in RNA copies per 50 ng of total RNA . Several mutants exhibited marked reductions ( >1 log decrease vs . WT ) in extracellular infectious titers ( A6 , A27 , A39 , A51 , A60 and F26W , A28W , V36W ) ( Fig 1C ) , and these data correlated with slightly lower levels of HCV RNA at 72 hours post-electroporation ( Fig 1B ) , likely due to reduced virus spread within the culture . Of these , A6 and A60 , located near the N- and C-termini , failed to produce any infectious virus . Surprisingly , the majority of p7 mutants were competent for infectious virus production , including mutants with alanine insertions within the first transmembrane domain , a region considered important for ion channeling . Further , several mutations ( e . g . A21 , F25W and Y31W ) even yielded titers above those obtained for WT virus indicating that p7 can accommodate these genetic changes ( S1 Fig ) . These results support a model of p7 structural plasticity in human hepatoma cells replicating full-length HCV genomes . The first structure of monomeric p7 was obtained by combining NMR experiments performed in a 2 , 2 , 2-trifluoroethanol ( TFE ) / water mixture with molecular dynamics ( MD ) simulations [41] . A full-length , FLAG-tagged monomeric p7 structure was later determined in methanol [42] , and recently , structures of p7 from two different genotypes were determined in 1 , 2-Dihexanoyl-sn-glycero-3-phosphocholine ( DHPC ) micelles [22] and dodecylphosphocholine ( DPC ) [47] , illustrating the monomeric and hexameric p7 forms , respectively . Importantly , these p7 NMR structure models differ on the location of secondary structural elements and orientation of p7 transmembrane domain regions ( Fig 2A ) , most notably for segment 33–47 . These discrepancies may be due in part to differences in the HCV genotype tested ( 1b vs . 5a ) and/or the lipid-mimicking environment used ( TFE , DHPC , DPC , or methanol ) , the latter of which has been shown to impact p7 activity [46] . To better visualize reported p7 structural elements , we used hexameric p7 models in DPC as described by OuYang and colleagues [47] ( model 1; Fig 2B ) and in 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine ( POPC ) as described by Chandler et al . [44] that employs the monomeric p7 structure put forth by Montserret et al . [41] ( model 2; Fig 2B ) . These two models were selected based on the availability of hexamer structure coordinates and our effort to compare divergent models . Interestingly , in model 1 , the p7 subunits are crossed such that part of each monomer interacts with all other subunits , while in model 2 , they reside side-by-side , illustrating the topology typical of two transmembrane helical proteins ( Fig 2B ) . However , despite the huge differences in the organization of the central part of p7 subunits between these two models , the N-terminal helix ( 1–18 ) is close to the C-terminus segment of p7 subunits in both models and forms a hexameric helix bundle with a similar organization of residues; notably , the side chains of amino acids 9 and 12 in both models point to the pore lumen ( Fig 2C ) . Importantly , this organization is also observed in other theoretical and NMR-based models [40 , 42] . Given these similarities in N-terminal helical packing and the conserved hydropathic pattern in this region , we extended our tryptophan mutagenesis of p7 to screen positions 1–18 ( Fig 3A ) . Similar to the mutants tested in Fig 1 , all of the N-terminal mutants replicated to similar wild-type levels ( Fig 3B ) , yet strikingly , mutagenesis in this region had a more profound impact on infectious virus production illustrated by a greater than 1 log decrease in titer for half of the mutants tested including A1W , L2W , E3W , K4W , V6W , H9W , A10W , A11W , and S12W . While varying levels of infectious HCV were detected for genomes harboring mutations at positions 5 , 7 , 8 , 11 , and 13 , six of the viruses tested ( A1W , E3W , K4W , H9W , A10W , and S12W ) failed to produce any detectible virus by 72 hours post-electroporation ( Fig 3C ) . Previous studies have shown that mutations in the N-terminal region of p7 can modulate the partial cleavage at the E2/p7 and p7/NS2 junctions [55] . To assess a potential impact of Trp substitution on host signal peptidase cleavage efficiency , we probed for E2 antibody-reactive proteins in Huh-7 . 5 cells replicating WT or p7 mutant genomes by western blot . To demonstrate E2 antibody specificity and distinguish between incompletely processed E2-containing protein species ( E2p7NS2 , E2p7 , and E2 ) , we utilized monocistronic , wild-type J6/JFH1 and ΔE1E2 genomes along with bicistronic genomes that contain an IRES between E2 and p7 or p7 and NS2 to remove the requirement for protein processing at these junctions . Parallel analysis of N-terminal helix mutants illustrated an E2/p7 processing defect for A1W , E3W , and K4W . Surprisingly , these data also suggested a similar defect may contribute to the deleterious phenotypes of other downstream N-terminal helix mutants , most notably for H9W through S12W ( Fig 3D ) . To identify second site amino acid changes that could compensate for E2/p7 cleavage or p7-specific defects introduced by tryptophan substitutions , we serially passaged Huh-7 . 5 cells harboring deleterious N-terminal mutant genomes to allow for the emergence of variants that are competent for infectious virus production ( Fig 4 ) . After three to seven passages , virus was detected in the supernatant for all genomes except H9W . Despite several attempts at electroporating hepatoma cells with this mutant genome , we were unable to select for a virus capable of spread . This was not due to a high genetic barrier ( i . e . the requirement of multiple nucleotide changes to obtain a viable virus ) , as both serine and glycine ( amino acids that are one nucleotide change away from tryptophan ) function in this position ( Fig 7; see also S1 Fig and S6 Fig for models ) . Further analysis by titrating WT RNA into a constant amount of either control ( Δp7 ) or H9W RNA at the time of electroporation resulted in 3 . 7-fold less infectious virus production at a WT:H9W RNA ratio of 1:2 ( compared to WT:Δp7 at the same ratio ) . While these data hinted that the mutant p7 might act via a dominant negative mechanism to suppress WT p7 and the production of infectious virus ( S2 Fig ) , the effect was less dramatic when the ratio was further increased in favor of the mutant . We next sequenced the p7 region of HCV RNA extracted from naïve Huh-7 . 5 cells inoculated with supernatant from passaged cells replicating mutant HCV genomes and identified conservative same-site changes in five of the ten viruses analyzed ( Fig 4A ) . These viruses–with mutations at positions 1 , 3 , 4 , 10 , and 12 –failed to produce any detectable infectious virus in our original characterization , and together , these data suggest that certain physico-chemical characteristics of the amino acid side chain at these positions are critical . Replacement of tryptophan at positions 1 and 10 with cysteine represents a reversion to a small residue , while glycine at polar positions 3 , 4 , and 12 likely represents a release of hydrophobic steric constrains ( S3 Fig ) . Re-engineering of these amino acid changes into the original mutant genome confirmed their ability to rescue WT-levels of virus production ( Fig 4C and 4D ) . L2W , V6W , and A11W revealed putative second-site mutations in E2; however , as we chose to focus our analysis on p7 , whether these mutations are responsible for rescuing infectious virus production remains to be determined . To investigate whether rescue of infectious virus production correlated with enhanced E2/p7 cleavage , we probed for E2 by western blot , comparing original mutant genomes with those re-engineered to contain p7 mutations identified after passage . Strikingly , all viruses harboring p7 mutations identified after passage yielded a marked increase in the amount of ‘free’ E2 relative to E2p7 compared to their original mutation counterparts ( Fig 4B ) . These data suggest that diminished cleavage at this junction contributed to our original deleterious phenotypes for these N-terminal helix mutant viruses and consequently impede our ability to evaluate their impact with respect to p7-specific functions . To evaluate the impact of tryptophan substitutions in the N-terminal helix independent of E2/p7 cleavage , we engineered these mutations at positions 1–13 into a bicistronic genome containing the EMCV IRES between E2 and p7 ( J6/JFH E2-IRES-p7 [11]; Fig 5A ) , thus eliminating the need for polyprotein processing at this junction . We then phenotyped these bicistronic N-terminal tryptophan substitution mutants with respect to replication and infectious virus production after electroporation into Huh-7 . 5 cells . As expected , all mutant viruses replicated to similar wild-type levels ( Fig 5B ) . Furthermore , as predicted by our western blot data indicating defects in E2/p7 cleavage were at least partially responsible for abrogation of infectious virus production , the majority of bicistronic N-terminal helix mutants now yielded infectious virus titers that were comparable to wild-type ( Fig 5C ) . Nonetheless , several mutants , including A1W , V6W , H9W , A10W and S12W , remained impaired , suggesting a cleavage-independent , p7-specific defect also impacts infectious particle production for these viruses . These phenotypes indicated a deleterious impact of tryptophan on p7 function at positions 1 , 6 , 9 , 10 and 12 but do not provide evidence for a rational hypothesis regarding the mechanism of the defect . Thus , to gain insight into the impact of these mutations on p7 structure , we modeled these tryptophan substitutions via homology molecular modeling . Because the structural impact , and hence , proposed functional consequence of our mutations , may differ depending on the 3D model , we aimed to develop hypotheses based on both model 1 and model 2 ( Fig 2B ) . Comparing the sequence of the J6 ( genotype 2a ) p7 used in our study to the genotype 1b and 5a p7 used to study the p7 structure by NMR indicated sufficient similarity at the amino acid level ( Fig 2A ) to enable the generation of J6 p7 models by homology using Swiss-Model facilities [56] . ( Coordinates of homology models 1 and 2 for p7 HCV J6 strain are available as supplementary . pdb files; S1 File and S2 File ) . Introduction of any of our tryptophan mutations in these homology models yielded energetically stable hexamer structures without significant structural changes indicating that p7 structure models 1 and 2 readily accommodate these mutations ( S1 Fig ) . We then closely examined each model to evaluate/predict the structural/functional consequence of the tryptophan substitution ( Fig 6A ) . Not surprisingly , given the similarity of these two models at the N-terminus ( Fig 2 ) , our predictions were generally consistent between model 1 and model 2 ( S1 Fig ) . Indeed , the orientation of the tryptophan side chain toward the lumen of the pore in both models suggests a likely ion channel defect for H9W and S12W . The A10W mutation could disturb p7 intramolecular interactions or interrupt p7 interactions with binding partners , as also predicted for the A1W mutant because of its N-terminal position ( S1 Fig ) . Interestingly , model 1 and model 2 did give rise to incongruent hypotheses for some mutants ( e . g . V6W ) , suggesting these residues may impact multiple aspects of p7 function . Alternatively , such mutants could be used as tools to test the accuracy of one model over the other by directly assessing the functional defect in cell culture . We next sought to corroborate the hypothesized p7 functional consequences of tryptophan substitution based on our homology models by further assessing selected mutants in cell culture . Specifically , we aimed to rescue infectious virus production by putative ion channel defective mutants in Huh-7 . 5 cells using bafilomycin A1 . Bafilomycin A1 prevents vesicular acidification and thus may compensate for a loss of p7 channel activity . Further , this inhibitor was previously shown to compensate for a defective p7 mutant harboring mutations K33A and R35A [36] . In our initial experiments , 8 nM bafilomycin was found to be both relatively non-toxic to cells and extremely effective in alkalinizing cellular compartments and over a 24 hr time period , retaining 80% cellular viability with complete loss of acidic organelle labeling with LysoTracker Red DND-99 ( S4 Fig ) . Since bafilomycin A1 can also prevent endosomal acidification and thus impede HCV entry into cells used for subsequent infectivity analysis , we further optimized methods to concentrate virus- and bafilomycin A1-containing supernatants 5-fold while simultaneously removing a sufficient amount of the inhibitor to enable infectious virus quantification by limiting dilution assay ( S4 Fig ) . We selected V6W , H9W , and S12W bicistronic p7 mutant viruses specifically for analysis based on our homology models that suggested an ion channel defect for both H9W and S12W . For V6W , the interpretation differed between model 1 and model 2 , providing a potential opportunity to decipher between them . Following electroporation and incubation of Huh-7 . 5 cells with selected viral genomes , cells were provided with media containing either bafilomycin A1 or DMSO . Cell culture supernatants were collected 24 hours later and infectivity was assessed ( Fig 6B ) . Notably , bafilomycin A1 treatment resulted in a boost of viral titers for all genomes tested that were capable of making detectible levels of infectious particles under DMSO conditions . However , only in the case of the KRAA mutant and our S12W mutant was this increase significant ( Fig 6C ) . These data support our structure model-based hypothesis that tryptophan substitution at position 12 abrogates ion channeling , and also indicate that altering intracellular pH via bafilomycin A1 treatment is insufficient to ‘rescue’ the impact of tryptophan substitution at position 6 . Interestingly , despite both models pointing to an ion channel defect for H9W , we were unable to recover any infectious virus for this mutant in our assay . One explanation is that our methods are not conducive to detection of low levels of infectious virus production; indeed , the limit of quantification for our limiting dilution assay in this context is 10-fold higher than previous experiments due to some residue bafilomycin A1 carryover in the supernatant . Thus , a small , but significant increase in infectious virus production , as was previously shown in a similar experiment with the KRAA mutant [36] , may not be uncovered . In both models interrogated in this study , residues 9 and 12 point to the pore formed by p7 oligomerization ( Fig 2C ) . These structural data , supported by our ability to significantly increase S12W mutant infectious viral titers by altering vesicular pH , indicate the amino acids at these positions could contribute to cation selectivity and flux . To further analyze the requirements at these residues , as well the residue at position 6 , which is oriented towards the pore in model 1 , we expanded the amino acid repertoire at these positions and analyzed the impact of polarity , charge , and hydrophobicity on viral replication and infectious virus production ( Fig 7 ) . Amino acids were analyzed in both monocistronic and bicistronic viral genetic backgrounds in order to segregate between amino acids that impact E2/p7 cleavage versus those that influence p7-specific functions . Our data indicate that although position 6 tolerates both hydrophilic and hydrophobic residues , bulky residues ( Leu , Phe and Trp ) are more detrimental in the monocistronic context , suggesting these amino acids have a negative impact on E2/p7 cleavage , whereas residues with smaller side chains ( Ala , Ser , Thr ) have almost no effect on infectious virus production ( Fig 7A and 7D , and S5 Fig ) . Specifically , we observed a correlation between the increasing size of hydrophobic residue side chains ( S6 Fig ) and the inhibition of virus production . Extending our analysis to position 9 , we observed that polar amino acids Gln , Asn , and Ser all function at this position to support infectious virus production while hydrophobic residues ( Ala , Cys , Leu and Tyr ) do not ( Fig 7B ) . In agreement with these data , H9A mutation in JFH-1 p7 ( genotype 2a ) was previously shown to reduce channel conductance by ~70% [47] . However , the impact of these hydrophobic residues was less significant in Huh-7 . 5 cells replicating bicistronic genomes , indicating the primary impact of these substitutions is on E2/p7 cleavage ( Fig 7E ) . Surprisingly , both acidic ( Asp ) and basic ( Arg ) residues support infectious virus production , albeit to low levels . Similar to position 9 , we also observed that amino acids at position 12 with bulky hydrophobic side chains inhibited virus production in the monocistronic context , but this was again less apparent for bicistronic genomes harboring the same amino acid changes ( Fig 7C and 7F ) , indicating again that the primary impact of these substitutions is on E2/p7 cleavage . In both cases , amino acids at positions 9 and 12 with polar character supported infectious virus production ( Fig 7B , 7C , 7E and 7F ) . Interestingly , substitution with negatively charged Asp at position 12 yielded an increase in viral titers above those obtained for WT , potentially via enhanced cation recruitment at the pore entry . However , positively charged Arg also functions at this position while Ala and Ser are the only natural amino acids found at this position . Together these data indicate that E2/p7 cleavage efficiency is sensitive to downstream mutations within the N-terminal helix region of p7 while the tolerance of positions 6 , 9 , and 12 to amino acids of different nature in a bicistronic context further supports p7 as a structurally plastic , minimalist ion channel .
In this report , we have extensively interrogated the HCV p7 protein via mutagenesis and determined the effects of these mutations on virus replication and infection in cell culture . In addition , we have modeled these mutations using p7 structure information based on previous NMR experiments . Our data confirm previous reports that p7 is not required for viral replication , as all p7 mutants tested replicated with wild-type efficiency . Importantly , our large-scale , structure-function analyses illustrate a global tolerance for amino acid sequence alterations , either by insertion or individual amino acid substitution in the J6/JFH background . These results underscore the structural flexibility of p7 [44] that has been similarly described for other viroporins such as HIV-1 Vpu [57] . Our data are further in line with the conservation of p7 amino acid physico-chemical properties and hydropathic character but not precise sequence across genotypes [41] . Notably , two of the nine conserved amino acids ( G18 and Y42 ) were directly assessed in this study by Trp substitution and resulted in an increase and decrease in infectious virus titer , respectively , although these phenotypes were not the most dramatic in our panel . Interestingly , structure models give rise to incongruent hypotheses regarding the impact of G18W mutation ( S1 Fig ) , suggesting this residue may provide another opportunity to further probe these structure models and test p7 function in cell culture and p7 ion channel activity after reconstitution in artificial membranes . Together our data indicate that escape mutants with significant fitness could be readily generated in the context of p7-targeting antiviral compounds , potentially limiting the efficacy of this class of inhibitors in the clinic . Still , there were several positions tested that did show a marked impact on infectious virus production; this was most pronounced when the residues within the first eighteen amino acids , comprising the N-terminal helical region , were interrogated . Our functional predictions are based on available hexameric p7 models; however , previous studies indicate that the oligomerization of seven p7 subunits is also feasible [35 , 44] , possibly even resulting in a mix of oligomeric states within the infected cell . Nonetheless , computational analyses for models where p7 subunits reside side-by-side [35 , 44] indicate that amino acid positions are similar and pore lining residues are retained with the addition of the 7th monomer; thus , our data interpretation would likely be consistent in this context . Interestingly , molecular dynamics simulations in a hydrated POPC bilayer showed that hexameric p7 model 2 formed a pore that was transiently permissive to solvent , potentially linked to a hydrophobic barrier formed by F25 [44] . In our initial mutant panels we observed a decrease in titer following Trp substitution at position F26 and an increase in titer after F25W mutation . Both hexameric models assayed here suggest residue 26 has multiple contacts within p7 , thus likely playing a role in stabilizing the protein , whereas the amino acid at position 25 lines the pore . Naturally occurring residues at both of these positions across genotypes are invariably hydrophobic , and while the higher polarity of the Trp side chain may facilitate the passage of ions at position 25 , resulting in a higher production of virus particles , the same substitution at position 26 may destabilize p7 assembly . The ability of Trp substitution to boost titers at several positions ( including residue 25 , as well as 29 and 31 ) is interesting given that this amino acid does not naturally occur at any of these positions . This suggests that the increased titers we observed in Huh-7 . 5 cells are not advantageous in a more physiologically relevant system ( e . g . primary human hepatocytes ) and may negatively impact viral fitness in vivo , although this was not directly tested in this report . A side-by-side structural comparison of the models proposed by OuYang et al . [47] ( model 1 ) and Chandler et al . [44] ( model 2 ) revealed similar helical packing at the N-terminus in these otherwise incongruent models; hence , we focused our studies on this region of p7 . Surprisingly , several lines of investigation , including western blot analyses of N-terminal helix mutants and related pseudorevertants , as well as infectious virus production phenotyping of bicistronic mutant genomes , all indicated that mutation within this region has a significant , detrimental impact on E2/p7 processing . This supports previous studies that have implicated this region in modulating the partial cleavage at E2/p7 and p7/NS2 junctions [58] . Interestingly , A13W , a mutant that originally demonstrated a 1-log attenuation compared to WT , acquired an additional mutation at position 17 ( N17D ) after passage that correlated with enhanced E2/p7 processing and increase in infectious virus production . In model 1 , residue 13 points towards the pore , while 17 lies at the p7 protein surface within the hydrophobic region of the membrane–making the identification of negatively charged aspartic acid at this position energetically counter intuitive ( S3 Fig ) . In model 2 ( as well as the model presented by Foster et al . [42] ) , however , residues 13 and 17 both point to the p7 pore , one directly above the other , indicating these residues could be related in function . Overall , our work suggests that tryptophan substitution ( or potentially bulky or hydrophobic residues in general , as demonstrated for residues 6 , 9 , and 12 ) negatively modulates important interactions between the C-terminus of E2 and the N-terminus of p7 that play a regulatory role in cleavage efficiency of the E2/p7 junction , possibly through mediating correct presentation of the cleavage site to the signal peptidase . In uncleaved E2p7 species , the topology of p7 may be inverted [59] , and while a specific role for E2p7 in the HCV life-cycle remains questionable , it has been hypothesized that the proper timing of E2/p7 cleavage may be critical to avoid spontaneous ion channel formation in the ER membrane and to promote the assembly process [44] . Nonetheless , the separation of E2 and p7 is absolutely required for infectious virus production , specifically for proper NS2 localization near assembly complexes [58] and presumably p7 oligomerization . Beyond deficiencies in protein processing , our mutagenesis data using bicistronic constructs , further informed by homology modeling , identified several N-terminal helix mutants with p7-specific defects , including positions 6 , 9 , and 12 . Interestingly , modeling of V6W ( Fig 6A ) , which is located within a conserved hydrophobic cluster spanning from position 5–8 and is either a Val or Ile in all HCV genotypes , indicated this mutation would face the pore in model 1 and has been proposed to play a role in closing the pore via the formation of a hydrophobic ring [47] . In contrast , this V6W mutation would likely affect helix-helix interactions ( i . e . oligomeric structural stability ) in model 2 . In fact , our inability to rescue this bicistronic mutant using bafilomycin A1 , in addition to the fact that position 6 tolerates both hydrophilic and hydrophobic residues suggests this position does not play a critical direct role in ion channeling . Position 9 is an Asn or , in genotype 2 viruses , a His–both of which have an affinity for monovalent and divalent cations . The hydrophilic nature of this position , as well as its location at the pore entry in both models , has implicated this residue in cation selectivity [41] . OuYang et al . [47] further propose it serves as a “filter” to dehydrate cations , allowing them to pass through the hydrophobic ring formed by position 6 , the more narrow part of the channel in model 1 . In accordance with these hypotheses , polar residues supported virus production in our study , while substitution with hydrophobic or charged residues resulted in significantly decreased infectious virus titers compared to wild-type . Our results for this residue were reminiscent of those obtained for position 17 , which is located within the turn sequence and found to line the pore . Position 17 naturally occurs as a histidine or asparagine ( like position 9 ) as well as glutamine–all of which share polar characteristics–and has been implicated in ion channel function [49] , as our present data suggest for position 9 . Similar to our results at position 9 , mutation of H17 to A or G by Meshkat et al . resulted in a decrease in titer while H17E ( polar ) boosted infectious virus in the supernatant [51] . Interestingly , we did not see a major phenotype following Trp substitution at position 17 ( N17W ) in the context of the J6 p7 sequence , perhaps due to conservation of some polarity by this substitution . Beyond positions 6 and 9 , we also investigated the previously overlooked residue at position 12 , which is only Ser or Ala in natural variants and points to the pore in both models . Our data show a striking complete loss of infectious virus production upon Trp substitution–a phenotype that was only partially rescued when this mutation was subsequently engineered into a bicistronic background . Modeling data indicated aromatic ring-mediated pore obstruction while subsequent rescue with bafilomycin A1 further suggest a novel role for this position in ion channel function . Still , the global tolerance of both position 9 and position 12 for amino acid substitution of different characteristics supports p7 as a structurally plastic , minimalist ion channel . Our present study identified N-terminal mutants that are defective for infectious virus production but did not distinguish between a defect in particle assembly versus infectivity . The identification of p7 mutants that are competent for particle assembly , but exhibit a profound defect in specific infectivity would provide a unique tool to probe the impact of p7 function on the viral particle at later stages of the viral life cycle . Further , deleterious mutants for which our models generated incongruent functional defect hypotheses offer additional opportunities to delineate between these models by probing p7 functional defects in cell culture or reconstituted in artificial membranes and correlating these with structure-based predictions to support or refute the available structural data . Importantly , our study highlights a potential regulatory role of the p7 N-terminal helix residues in the cleavage efficiency of the E2/p7 junction , although the precise underlying mechanisms remain elusive . In sum , our work illustrates the convergence of current p7 models at the N-terminal helix and demonstrates the biological impact of amino acid perturbation in this region , offering extensive insight into the relationship between p7 structure and function in the context of HCVcc .
J6/JFH [60] , the K33A/R35A p7 mutant and bicistronic genomes J6/JFH E2-IRES-p7 and p7-IRES-NS2 [11] have all been previously described . To facilitate the generation of p7 mutant viruses , two silent restriction sites ( NotI and BglII ) were engineered into the wild type monocistronic J6/JFH sequence in E2 and NS2 , respectively . The resulting genome is termed J6/JFH 1 . 1 and referred to here simply as J6/JFH or wild-type . Mutations in p7 were introduced by overlap PCR using standard procedures and engineered into monocistronic or bicistronic constructs using NotI and BgIII or MluI and NotI , respectively . Plasmid and primer sequences are available upon request . All constructs were confirmed by sequencing . Huh-7 . 5 cells [61] were propagated in Dulbecco’s modified minimal essential medium ( DMEM ) supplemented with 10% heat-inactivated fetal bovine serum ( FBS ) and 0 . 1 mM nonessential amino acids ( NEAA ) . Cells were grown at 37°C in a humidified 5% CO2 atmosphere . Viral cDNAs were linearized with XBa1 and purified using a MinElute PCR purification kit ( Qiagen ) . In vitro RNA transcription was performed using a T7 RiboMAX Express large-scale RNA production system ( Promega ) and newly synthesized RNAs were isolated with an RNeasy RNA isolation kit with a second DNase I digestion ( Qiagen ) according to the manufacturer’s protocol . RNAs were eluted in nuclease free water and integrity and concentration were determined by agarose gel electrophoresis and absorbance at 260 nm , respectively . In vitro-transcribed RNAs were electroporated into cells using a 4 mm gap 96-well plate format ( BTX ElectroSquare Porator ECM830 with Plate Handler; Harvard Apparatus ) . Briefly , Huh-7 . 5 cells were trypsinized , washed in cold , RNase-free Dulbecco’s phosphate-buffered saline ( D-PBS ) without Ca2+ / Mg2+ ( Gibco–Invitrogen ) and resuspended at a concentration of 1 . 5 x 107 cells / ml in D-PBS . Two hundred microliters ( 3 x 106 cells ) was then mixed with 5 ug RNA and loaded into the cuvette . Electroporation was performed using the following settings: 0 . 80 kV , 99 ms , 5 pulses . Pulsed cells were transferred into 1 . 5 ml DMEM with 10% FBS and 0 . 1 mM NEAA before plating . Cells were plated in 24-well plates at a density of 5 . 3 x 104 cells / well in a final volume of 0 . 5 ml . Eight hours post-electroporation , cells were washed twice with D-PBS . One well from each electroporation was then harvested in 0 . 35 ml RLT buffer containing 0 . 01 ml beta-mercaptoethanol ( βME ) per ml , applied to a Qiashredder and spun at 16 , 300 x g for 2 min before storage at -80°C . A second well from each electroporation was provided with 0 . 5 ml fresh complete medium and returned to the incubator until 72 hours post-electroporation when the cell culture supernatant was collected and stored at -80°C until analysis . The cells were then washed and collected in RLT buffer as described above . HCV infectious titers in the supernatants were determined by a limiting dilution assay on naïve Huh-7 . 5 cells as previously described [60] . Total cellular RNA was isolated using an RNeasy kit ( Qiagen ) and 50 ng of total RNA was then assayed for HCV genomes using a one-step quantitative RT-PCR assay ( Multicode-RTx HCV RNA kit , Luminex Corp . ) targeting the 3’ UTR of the viral genome and a Roche LC480 light cycler , according to manufacture’s instructions . Selected p7 genomes shown to be defective for infectious virus production were electroporated into Huh-7 . 5 cells as described above . Cells were plated ( 1 . 3 x 106 cells ) in 100 mm dishes and maintained in 10 ml DMEM 10% FBS with 0 . 1 mM NEAA . Supernatants were collected before each passage and stored at -80°C until analysis . Supernatants found to contain infectious virus were then applied to naïve Huh-7 . 5 cells ( 300 , 000 cells / 100 mm dish plated 24 hrs prior to inoculation ) , split once , and harvested in 0 . 6 ml RLT containing βME for RNA extraction and HCV RNA sequencing or fixed in 4% paraformaldehyde ( PFA ) and stained with anti-NS5A antibody ( 9E10 [60]-alexafluor 647 ) to determine the frequency of HCV antigen-positive cells by flow cytometry . The relatively high amino acid sequence similarities between p7 of HCV strain J6 and that of strains EUH1480 ( 40% identity , 80% overall similarity ) and HC-J4 ( 62% identity , 92% overall similarity ) allowed us to construct three-dimensional homology models 1 and 2 for p7 hexamers , respectively , using the NMR structure of HCV p7 of OuYang et al . [47] as template ( PDB accession number 2M6X ) for model 1 , and the NMR/MD model of Chandler et al . [44] as template for model 2 . Models of p7 were constructed with the Swiss-Model automated protein structure homology modeling server ( http://www . expasy . org/spdbv/ [56] ) using the p7 HCV strain J6 sequence as input . p7 model 1 was directly obtained as a hexamer by the automated procedure . For model 2 , raw amino acid sequence of p7 from strain J6 was first loaded in Swiss-PdbViewer software [56] and fitted to the NMR/MD p7 hexamer model of Chandler et al . [44] before submission for model building to Swiss-Model using the SwissModel Project Mode . All p7 mutants were constructed using the latter protocol , i . e . , fitting of the raw amino acid sequence of p7 mutants to wild type hexamer models 1 and 2 from the J6 strain . Coordinates of homology models 1 and 2 for p7 HCV J6 strain are available as supplementary . pdb files . These coordinates are derived directly from the automated model building with no further minimization or manual manipulation . Electroporated Huh-7 . 5 cells were plated in 6-well plates and lysed 72 hpe using modified radioimmunoprecipitation assay ( RIPA ) buffer ( 50 mM Tris-HCl ( pH 8 . 0 ) , 1% ( v/v ) nonyl phenoxypolyethoxylethanol , 0 . 5% ( w/v ) Na-deoxycholate , 150 mM NaCl , and 0 . 1% sodium dodecyl sulfate ) . Protein ( 10 μg ) was then denatured and subsequently deglycosylated using PNGase F according to the manufacturer’s protocol ( New England BioLabs , Inc . ) before being separated on 4–12% Bis-Tris NuPAGE polyacrylamide gels ( ThermoFisher Scientific ) and transferred to 0 . 2 micron nitrocellulose membranes . Membranes were blocked with 5% milk in Tris-buffered saline with 0 . 1% Tween-20 and E2-containing protein species were detected using rat anti-E2 antibody ( clone 3/11 [62]; 2 μg/ml final concentration ) . Following secondary antibody staining with Peroxidase AffiniPure donkey anti-rat IgG ( H+L; 1:10 , 000 ) , blots were visualized using SuperSignal West Dura reagent ( Thermo Scientific ) . To establish the bafilomycin concentration to be used in subsequent virus rescue experiments , bafilomycin A1 ( Sigma Aldrich; or DMSO vehicle control ) was titrated onto mock-electroporated cells and both viability and intracellular pH assessed 24 hrs post-treatment . Cellular viability was determined using CellTiter-Glo luminescent cell viability assay ( Promega ) according to the manufacturer’s protocol . Parallel wells were washed in HEPES buffer [36] and loaded with 50 nM LysoTracker Red DND-99 ( ThermoFisher Scientific ) diluted in HEPES buffer for 30 min at 37°C to label acid organelles . Cells were then washed with PBS , trypsinized , and LysoTracker Red content analyzed by flow cytometry after gating on live cell singlets . For rescue experiments , selected p7 mutant genomes were electroporated into Huh-7 . 5 cells as described above . Forty-eight hours post-electroporation , the cell culture media was removed and replaced with media containing bafilomycin A1 ( 8 nM final concentration ) or DMSO ( vehicle control ) . Supernatants were harvested 24 hours post-treatment , pooled across identical wells and applied to Millipore centrifugal filters ( 100 MW cutoff ) . Samples were centrifuged at 930 x g for 15 min at 4°C and then dialyzed with 4 ml serum free medium by centrifuging again at 930 x g for 12 min to remove bafilomycin A1 . The remaining sample volume ( 250–500 μl ) was brought up to 600 μl with DMEM containing 10% FBS and 0 . 1 mM nonessential amino acids and infectious virus was quantified by standard limiting dilution assay performed on naïve Huh-7 . 5 cells . Statistical analysis of virological data was performed with GraphPad Prism 5 . Specific tests are noted in figure legends .
|
Hepatitis C virus ( HCV ) infection can lead to significant liver disease and , without a vaccine , continues to pose a significant public health threat . The viral p7 protein is a multifunctional protein that is required for infectious virus production via its role in orchestrating virion assembly and its activity as an ion channel . However , while there is accumulating structural information on p7 , there is no consensus on which conformation ( s ) exist during a natural infection or how structural elements relate to p7 functions . By comparing two prominent , yet highly divergent models of p7 , we identified one region of structural similarity–the N-terminal helical region . While mutagenesis screening of other regions of the protein are in keeping with p7 conformational flexibility , mutations within the N-terminal helical region had a significant impact on infectious virus production , due in part to a loss of efficient E2/p7 cleavage . We further postulated the precise functional impact of mutations throughout p7 by homology modeling and demonstrated tolerance for diverse amino acid substitutions for specific N-terminal helix residues with putative ion channel defects . Together , these data not only support p7 as a structurally plastic , minimalistic ion channel , but also provide extensive insight into the p7 structure-function relationship and highlight the importance of the N-terminal helical region in E2/p7 processing , protein-protein interactions , ion channel activity , and infectious HCV production .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
The N-terminal Helical Region of the Hepatitis C Virus p7 Ion Channel Protein Is Critical for Infectious Virus Production
|
Ichthyoses are a heterogeneous group of inherited cornification disorders characterized by generalized dry skin , scaling and/or hyperkeratosis . Ichthyosis vulgaris is the most common form of ichthyosis in humans and caused by genetic variants in the FLG gene encoding filaggrin . Filaggrin is a key player in the formation of the stratum corneum , the uppermost layer of the epidermis and therefore crucial for barrier function . During terminal differentiation of keratinocytes , the precursor profilaggrin is cleaved by several proteases into filaggrin monomers and eventually processed into free amino acids contributing to the hydration of the cornified layer . We studied a German Shepherd dog with a novel form of ichthyosis . Comparing the genome sequence of the affected dog with 288 genomes from genetically diverse non-affected dogs we identified a private heterozygous variant in the ASPRV1 gene encoding “aspartic peptidase , retroviral-like 1” , which is also known as skin aspartic protease ( SASPase ) . The variant was absent in both parents and therefore due to a de novo mutation event . It was a missense variant , c . 1052T>C , affecting a conserved residue close to an autoprocessing cleavage site , p . ( Leu351Pro ) . ASPRV1 encodes a retroviral-like protease involved in profilaggrin-to-filaggrin processing . By immunofluorescence staining we showed that the filaggrin expression pattern was altered in the affected dog . Thus , our findings provide strong evidence that the identified de novo variant is causative for the ichthyosis in the affected dog and that ASPRV1 plays an essential role in skin barrier formation . ASPRV1 is thus a novel candidate gene for unexplained human forms of ichthyoses .
The skin and in particular the epidermis provide both an outward and inward barrier function , which is essential for survival . Aberrant skin development or homeostasis can impair this barrier function and may result in skin disorders . Ichthyoses are a heterogeneous group of skin disorders characterized by dry skin , scaling and/or hyperkeratosis , often associated with erythroderma [1 , 2] . These clinical signs are caused by a defect in the terminal differentiation of keratinocytes and subsequent desquamation taking place in the uppermost layer of the epidermis , the stratum corneum . Ichthyoses are primarily inherited skin disorders that can either be non-syndromic , when clinical findings are limited to the skin , or syndromic in case additional organs are involved [2] . Non-syndromic forms of ichthyoses are further sub-classified into common ichthyoses , autosomal recessive congenital ichthyoses , keratinopathic ichthyoses caused by variants in different keratin genes , and other forms of ichthyoses [1 , 2] . The common ichthyoses consist of ichthyosis vulgaris ( IV ) and recessive X-linked ichthyosis ( RXLI ) . IV is the most common and mildest form of ichthyosis with an incidence of approximately 1:250 to 1:1000 in humans [2 , 3] . IV is caused by different semidominant genetic variants in the FLG gene encoding filaggrin [4] . Filaggrin ( filament aggregating protein ) is a keratin bundling protein and a key player in the formation of the stratum corneum [5] . The precursor of filaggrin is the >400 kDa protein profilaggrin , which is the major component of keratohyalin granules in the granular layer of the skin [6 , 7] . Profilaggrin consists of a unique N-terminus and a series of filaggrin units separated by short linker peptides . The initially highly phosphorylated profilaggrin is dephosphorylated and cleaved by proteases into individual filaggrin molecules during the cornification process . In addition to its role in aggregating keratin intermediate filaments into bundles , filaggrin is also degraded into free amino acids that contribute to the hydration of the cornified layer [8] . RXLI , sometimes also called X-linked ichthyosis ( XLI ) , is clinically more severe and characterized by dark brown scales and generalized dry skin . It is caused by variants , mainly large deletions , affecting the STS gene , which encodes steroid sulfatase [9] . Autosomal recessive congenital ichthyoses ( ARCI ) are the second category of non-syndromic ichthyoses . They may be caused by variants in at least 9 different genes: ABCA12 , ALOX12B , ALOXE3 , CERS3 , CYP4F22 , LIPN , NIPAL4 , PNPLA1 and TGM1 [2] . Finally , keratinopathic ichthyoses , the third category of non-syndromic ichthyoses , are caused by variants in the KRT1 , KRT2 , or KRT10 genes [2] . Thus , there are currently 14 human genes implicated in different forms of non-syndromic ichthyoses [2 , 10] . Dogs represent valuable models for many human hereditary diseases and enabled for example the discovery of PNPLA1 as an ichthyosis gene . The first pathogenic PNPLA1 variant was identified in Golden Retriever ichthyosis , which is characterized by a mild phenotype . Interestingly , this canine genodermatosis currently has an extremely high prevalence in the breed [11 , 12] . Other dog models for human non-syndromic ichthyoses include Norfolk Terriers with an epidermolytic ichthyosis caused by a KRT10 variant [13] , Bulldogs with ARCI caused by a NIPAL4 variant [14] , and Jack Russell Terriers with another form of ARCI caused by a TGM1 variant [15] . Further cases of canine ichthyoses have been reported , but the underlying genetic defects have not been solved [16] . Thus , dogs might help to identify additional ichthyosis genes , which might be of relevance for unsolved human forms of ichthyoses . In the present study , we describe a novel non-epidermolytic form of ichthyosis in a German Shepherd . In this breed , until now , no ichthyosis cases have been reported in the scientific literature . We therefore applied a whole genome sequencing approach to unravel the causative genetic variant .
An intact female German Shepherd was presented at 10 months of age with a history of severe scaling of the skin with mild pruritus . According to the owner , the lesions started to develop shortly after birth . Dermatological examination revealed generalized hypotrichosis and focal areas of alopecia with generalized severe exfoliation of greyish scales and mild erythema . Comedones were seen on the ventral abdomen and in the perivulvar area ( Fig 1 ) . The owner reported that this phenotype had not been seen in the six littermates or the parents of the affected German Shepherd . The skin condition improved under topical treatment with a rehydrating , anti-seborrheic spray and shampoo . Histopathological analysis of four skin biopsies from different body regions revealed a severe laminar to compact orthokeratotic hyperkeratosis extending into the follicular infundibula in all biopsies . The keratin layers were multifocally exfoliating as large scales . The underlying epidermis was mildly hyperplastic . In the biopsy from the inguinal region , the infundibula of the hair follicles were moderately dilated . The histological findings were consistent with a cornification disorder and an inherited non-epidermolytic ichthyosis as possible cause ( Fig 2 ) . We sequenced the genome of the affected dog at 31x coverage and called SNVs and small indel variants with respect to the reference genome ( CanFam 3 . 1 ) . We then compared these variants to whole genome sequence data of 288 control dogs of various breeds including 13 German Shepherds not closely related to the affected dog ( Table 1 ) . As we did not find any protein-changing variants in the 14 known ichthyosis-associated genes , we hypothesized that the affected dog represented an isolated case of a novel form of ichthyosis . Consequently , we considered both a recessive and a dominant mode of inheritance for the hypothetical mutant allele . As purebred dogs are maintained in closed populations with a small effective population size and a considerable degree of inbreeding , recessive genetic defects within a breed typically can be traced back to single founders and are mostly found in homozygous state in affected dogs . In a first analysis , assuming a recessive mode of inheritance , we therefore searched for homozygous private protein-changing variants in the affected dog . Our automated pipeline detected 4 such homozygous protein-changing variants . However , upon individual visual inspection all 4 variants turned out to be sequencing artifacts . They were either located close to gaps in the reference assembly , in highly repetitive sequences , or in regions with low read coverage ( S1 Table ) . Given that the described dog was the only case in a litter of seven and ichthyosis had never before been reported in the German Shepherd breed , we hypothesized that a dominant mode of inheritance due to a de novo mutation event was more likely than a recessive mode of inheritance . In a second analysis , we therefore filtered for heterozygous private protein-changing variants . Our automated pipeline identified 19 such variants in 13 genes . None of the identified variants was located in a known ichthyosis gene . In order to identify potential de novo variants , we obtained whole genome sequences from both parents of the affected dogs . Inspection of the sequencing data for each of the 19 heterozygous candidate variants revealed that only one of them was indeed a de novo variant ( S1 Table ) . This de novo variant was a missense variant , c . 1052T>C , located in the ASPRV1 gene encoding “aspartic peptidase , retroviral-like 1” also known as skin aspartic protease ( SASPase ) , which is involved in profilaggrin-to-filaggrin processing [17 , 18 , 22] . We performed Sanger sequencing in the affected dog and both parents and confirmed that the variant was absent in both parents ( Fig 3A ) . We experimentally confirmed the correct parentage by an analysis of microsatellite and SNV genotypes in the trio . The ASPRV1:c . 1052T>C variant is predicted to result in the amino acid substitution p . ( Leu351Pro ) . The leucine at position 351 is strictly conserved among different species of placental mammals and only one residue away from one of the major cleavage sites required for auto-activation of the protein ( Fig 3B ) [17] . To assess the putative impact of the ASPRV1 missense variant we performed immunofluorescence staining with anti-ASPRV1 antibodies on skin sections of the affected and a control dog . The ASPRV1 signal in the affected dog showed the expected localization , mainly in the stratum granulosum , but was stronger than in the control dog ( Fig 4 ) . As this experiment could not assess whether the detected ASPRV1 protein is functional , we also investigated filaggrin processing by immunofluorescence staining with anti-filaggrin antibodies . This experiment demonstrated an abnormal filaggrin expression pattern in the affected dog ( Fig 4 ) . In the affected dog , diffuse staining across epidermal layers ( from stratum basale through stratum spinosum to stratum granulosum ) and some nuclear staining indicated defective processing of profilaggrin .
In the present study we identified a de novo missense variant in the canine ASPRV1 gene in a dog with a novel form of ichthyosis . We provide five arguments supporting the causality of the ASPRV1:c1052T>C variant for the observed ichthyosis . First , the c . 1052T>C variant leads to a non-conservative amino acid exchange p . ( Leu351Pro ) close to the functionally important auto-cleavage site of the ASPRV1 protease . It is thus conceivable that this specific genetic variant might affect the protein function . Second , the c . 1052T>C was absent from 288 non-affected dogs of different breeds and thus perfectly associated with the disease status . Third , the affected dog was heterozygous for the variant , but the mutant allele was absent in blood leukocytes of both parents . We therefore confirmed that the variant had arisen by a de novo mutation event that must have occurred in either one of the parental germlines or during early embryonic development of the affected dog . While exact numbers on the frequencies of de novo mutation events in dogs were not available at the time of this study , an analysis of 10 human trios reported 73 de novo mutation events on average per trio [19] . In another study on human de novo mutation events , it was shown that only ~1 . 3% of the de novo variants actually represented protein-changing variants [20] . Similar numbers of de novo mutation events were observed in cattle [21] . If we assume that these numbers are similar in dogs , one might expect roughly one de novo protein-changing variant in any dog on average , which exactly matches our data with one identified protein-changing variant in the affected dog . We deem it unlikely that such an event would coincidentally affect a gene with a known role in filaggrin processing without being causative for the ichthyosis phenotype . Fourth , we observed a difference in the ASPRV1 protein expression between the affected dog and a control dog . Somewhat surprisingly , the ASPRV1 protein expression was upregulated in the affected dog . Such an upregulation might have been caused by a compensatory mechanism , if the expressed ASPRV1 protein is non-functional as has been reported for other missense variants [22] . Thus , the increased ASPRV1 quantity is at least compatible with a causal role of the ASPRV1 missense variant in the observed ichthyosis . Finally and as fifth argument , we experimentally confirmed that the filaggrin protein expression pattern was altered in the affected dog suggesting a defect in profilaggrin-to-filaggrin processing . In our opinion and taken together , these arguments prove the causality of the ASPRV1:c1052T>C variant for the observed ichthyosis beyond any reasonable doubt . The ASPRV1 gene and its encoded protease were initially identified in humans and shortly afterwards in mice . ASPRV1 protein expression was only detected in stratified epithelia and was restricted to the stratum granulosum [17 , 18] . Further studies suggested that ASPRV1 cleaves the linker sequence in profilaggrin . A deficiency of ASPRV1 resulted in a lower level of stratum corneum hydration [23] . Furthermore , high ASPRV1 protein levels were present in several non-neoplastic skin disorders [17 , 24] . In transgenic mice , aberrant Asprv1 expression resulted in delayed wound closure [25] . Asprv1 deficient mice ( Asprv1-/- ) in a C57BL/6J background showed characteristic parallel skin wrinkles or lined grooves parallel to the body axis , but were reported to have normal skin histology and did not show any signs of ichthyosis [18] . The skin of Asprv1-/- mice in a hairless background ( Hos:HR-1 ) displayed more fine wrinkles and was drier and rougher than in Asprv1+/- or Asprv1+/+ mice [23] . In addition to an increased number of epidermal cell layers and a lower level of stratum corneum hydration , a decreased amount of filaggrin monomers together with an accumulation of aberrantly processed profilaggrin in the lower stratum corneum was observed in these mice . The total amount and the composition of free amino acids was however not significantly different from control mice [23] . Our results clearly showed an altered filaggrin expression pattern in the skin of an affected dog with the ASPRV1 variant . Thus , similar to Asprv1 deficient mice , profilaggrin processing appeared to be defective in the affected dog . In contrast to the findings by Matsui et al . [23] , we did however not detect an accumulation of incompletely processed filaggrin in the stratum corneum , but rather in the stratum spinosum and stratum granulosum . It remains unclear why the Asprv1-/- mice did not show an ichthyosis phenotype as the ASPRV1 mutant dog . Potential explanations include a gain of function effect of the specific canine missense variant or general differences in the homeostasis of the epidermis between mice and dogs . According to our knowledge , the ASPRV1 gene has not been associated with any form of ichthyosis in humans . Missense variants in the human ASPRV1 gene were reported in 5 of 196 Japanese patients with atopic eczema and 2 of 28 control subjects [23] . Two of the identified variants , p . V243A ( identified in a control subject ) and p . V187I ( identified in 3 atopic eczema patients ) led to absence or reduction of ASPRV1 activity in vitro [23] . Another study failed to find significant associations between ASPRV1 genetic variants and atopic eczema or clinically dry skin in different cohorts of Caucasian ancestry [26] . In conclusion , with the identification of a dominant de novo missense variant in the ASPRV1 gene of an ichthyotic dog , we present a new candidate gene for ichthyosis . It seems possible that ASPRV1 variants might also contribute to unsolved human ichthyoses .
All animal experiments were performed according to the local regulations . The dogs in this study were examined with the consent of their owners . The study was approved by the “Cantonal Committee For Animal Experiments” ( Canton of Bern; permits 22/07 , 23/10 , and 75/16 ) . The affected dog was examined by a board certified veterinary dermatologist in a private specialist clinic and followed up after initiating treatment with a rehydrating , anti-seborrheic spray ( Ermidra , Ufamed AG , Sursee , Switzerland ) and shampoo ( Sebomild P , Virbac AG , Glattbrugg , Switzerland ) . The absence of a similar phenotype in littermates , parents and ancestors was reported by the owner . Skin biopsies ( 6 mm ) of the case were taken from the flank , thigh , shoulder , and inguinal region and fixed in 10% buffered formalin for 24 hours . Biopsies were processed , embedded in paraffin and sectioned at 4 μm . Skin sections were stained with hematoxylin and eosin . The histopathology was performed by board certified pathologists . We isolated genomic DNA from EDTA blood samples of the affected dog and its parents . We confirmed the parentage by two different approaches: A multiplex PCR with 7 fluorescently labeled microsatellites primer pairs was performed for both parents and the case . Allele sizes were determined on an ABI 3730 capillary sequencer ( Life Technologies ) and analyzed using the GeneMapper software ( Life Technologies ) . Mendelian transmission of the alleles was confirmed at all 7 loci . The three dogs were additionally genotyped for 173 , 662 SNVs on the illumina canine_HD chip by GeneSeek/Neogen . We evaluated the parentage using the --genome and --mendel commands in plink 1 . 07 [27] . Both analyses were in agreement with the assumed parentage ( <0 . 001% genome regions with IBD = 0 for mother and father; 26 Mendel errors for the trio ) . Illumina PCR-free TruSeq fragment libraries with insert sizes of 350 bp—400 bp were prepared . For the affected dog , 276 million 2 x 150 bp paired-end reads or 31x coverage were obtained on a HiSeq3000 instrument . The parents were sequenced at 21x coverage . The reads were mapped to the dog reference genome assembly CanFam3 . 1 and aligned using Burrows-Wheeler Aligner ( BWA ) version 0 . 7 . 5a [28] with default settings . The generated SAM file was converted to a BAM file and the reads were sorted using samtools [29] . Picard tools ( http://sourceforge . net/projects/picard/ ) was used to mark PCR duplicates . To perform local realignments and to produce a cleaned BAM file , we used the Genome Analysis Tool Kit ( GATK version 2 . 4 . 9 , 50 ) [30] . GATK was also used for base quality recalibration with canine dbsnp data as training set . The sequence data were deposited under the study accession PRJEB16012 at the European Nucleotide Archive . The sample accessions are SAMEA4506895 for the case ( DS043 ) , SAMEA72802918 for the sire ( DS053 ) and SAMEA72802168 for the dam ( DS051 ) . Putative SNVs were identified in each sample individually using GATK HaplotypeCaller in gVCF mode , and subsequently genotyped per-chromosome and genotyped across all samples simultaneously [31] . Filtering was performed using the variant filtration module of GATK . To predict the functional effects of the called variants , SnpEFF [32] software together with the ENSEMBL ( version 80 ) annotation CanFam 3 . 1 was used . We additionally visually inspected the short read alignments of the functional candidate genes ABCA12 , ALOX12B , ALOXE3 , CERS3 , CYP4F22 , FLG , KRT1 , KRT10 , KRT2 , LIPN , NIPAL4 , PNPLA1 , STS , and TGM1 in the integrative genome viewer [33] to exclude any structural variants in these genes . We also inspected the CLDN1 gene in the same manner . For variant filtering we used 288 control genomes , which were either publicly available [34] or produced during other projects of our group . A detailed list of these control genomes is given in S2 Table . We used the dog CanFam 3 . 1 reference genome assembly for all analyses . Numbering within the canine ASPRV1 gene corresponds to the accessions XM_014117456 . 1 ( mRNA ) and XP_013972931 . 1 ( protein ) . Numbering within the human ASPRV1 gene corresponds to the accessions NM_152792 . 2 ( mRNA ) and NP_690005 . 2 ( protein ) . We used Sanger sequencing to confirm the candidate variant c . 1052T>C in ASPRV1 in the affected dog and its absence in both parents . A 370 bp fragment containing the variant was PCR amplified from genomic DNA using AmpliTaq Gold 360 Master Mix ( Life Technologies ) and the primers ACCCCAGGGACAGATTAAGG and AGCTGAAGCTGAAGGCAGAG . After treatment with shrimp alkaline phosphatase and endonuclease I , PCR products were directly sequenced on an ABI 3730 capillary sequencer ( Life Technologies ) . We analyzed the Sanger sequence data using the software Sequencher 5 . 1 ( GeneCodes ) . Immunofluorescence staining was performed on formalin-fixed paraffin-embedded skin sections of an age- and sex-matched control dog and the affected dog with some adaptations as described previously [35] . Briefly , tissue sections were deparaffinized using xylene . For ASPRV1 staining , antigens were retrieved in a microwave oven for 20 min at 95°C in sodium citrate buffer ( 10 mM sodium citrate , pH 6 . 0 ) . Blocking was performed for 1 . 5 hours at room temperature ( 10% goat serum , 1% BSA , 0 . 1% Triton X-100; 5% cold fish gelatin in PBS ) . Tissue sections were incubated with a polyclonal rabbit anti-ASPRV1 antibody ( 1:250 , NBP2-33981 , Novus Biological ) overnight at 4°C and with the secondary goat anti-rabbit Alexa Fluor 488 nm antibody ( 1:1000 , Abcam ) for 1 hour at room temperature . DNA was stained with 4’ , 6-diamidino-2-phenylindole ( DAPI ) contained in Vectashield Antifade Mounting Medium ( Vector Laboratories ) . For filaggrin staining , antigens were retrieved in a pressure cooker for 15 min in Tris buffer ( 100 mM ) with 5% urea . Blocking was performed for 1 . 5 hours at room temperature ( 10% goat serum , 1% BSA , 0 . 1% Triton X-100; 5% cold fish gelatin in PBS ) . Tissue sections were incubated with a polyclonal rabbit anti-filaggrin antibody ( 1:250 , PRB-417P-100 , Covance ) overnight at 4°C and with the secondary goat anti-rabbit Alexa Fluor 488 nm antibody ( 1:1000 , Abcam ) for 1 hour at room temperature . DNA was stained with DAPI ( 1:1000 , Sigma Aldrich ) . Tissue sections serving as negative controls were incubated with rabbit IgG serum . Images were taken with a Nikon Eclipse 80i fluorescence microscope using a Plan Flour x40/10 oil-immersion objective and excitation wavelength of 393 and 488nm . Pictures were captured and further processed using Improvision Open Lab 5 . 5 . 2 . software .
|
The skin undergoes a constant process of self-renewing and keratinocytes migrate from the basal layer of the epidermis to the uppermost layer , the stratum corneum , as they differentiate . A defect in the differentiation of keratinocytes can lead to cornification disorders such as ichthyosis . The most common form of this disorder in humans is ichthyosis vulgaris caused by variants in the filaggrin gene . Filaggrin is required for bundling intermediate filaments resulting in the flattening of keratinocytes . Filaggrin is produced from profilaggrin and the processing steps involve several enzymes including proteases . In the present study , we sequenced the genome of a dog with a novel form of ichthyosis . By comparing this sequence to 288 control genomes , we identified a private missense variant in the ASPRV1 gene encoding the retroviral-like aspartic protease 1 , also known as SASPase , which is involved in the processing of profilaggrin . The variant was due to a de novo mutation event , which is consistent with the patient being an isolated single case of a novel form of ichthyosis . Filaggrin protein expression was altered in the skin of the affected dog . Thus , our results strongly suggest that genetic variants in ASPRV1 can cause ichthyosis by altering filaggrin processing .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"skin",
"medicine",
"and",
"health",
"sciences",
"integumentary",
"system",
"immunofluorescence",
"staining",
"enzymes",
"enzymology",
"vertebrates",
"dogs",
"animals",
"mammals",
"protein",
"expression",
"genome",
"analysis",
"epidermis",
"mammalian",
"genomics",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"proteins",
"comparative",
"genomics",
"molecular",
"biology",
"animal",
"genomics",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"gene",
"expression",
"and",
"vector",
"techniques",
"biochemistry",
"anatomy",
"genetics",
"biology",
"and",
"life",
"sciences",
"proteases",
"genomics",
"amniotes",
"computational",
"biology",
"organisms"
] |
2017
|
A de novo variant in the ASPRV1 gene in a dog with ichthyosis
|
Pleuropulmonary Blastoma ( PPB ) is the primary neoplastic manifestation of a pediatric cancer predisposition syndrome that is associated with several diseases including cystic nephroma , Wilms tumor , neuroblastoma , rhabdomyosarcoma , medulloblastoma , and ovarian Sertoli-Leydig cell tumor . The primary pathology of PPB , epithelial cysts with stromal hyperplasia and risk for progression to a complex primitive sarcoma , is associated with familial heterozygosity and lesion-associated epithelial loss-of-heterozygosity of DICER1 . It has been hypothesized that loss of heterozygosity of DICER1 in lung epithelium is a non-cell autonomous etiology of PPB and a critical pathway that regulates lung development; however , there are no known direct targets of epithelial microRNAs ( miRNAs ) in the lung . Fibroblast Growth Factor 9 ( FGF9 ) is expressed in the mesothelium and epithelium during lung development and primarily functions to regulate lung mesenchyme; however , there are no known mechanisms that regulate FGF9 expression during lung development . Using mouse genetics and molecular phenotyping of human PPB tissue , we show that FGF9 is overexpressed in lung epithelium in the initial multicystic stage of Type I PPB and that in mice lacking epithelial Dicer1 , or induced to overexpress epithelial Fgf9 , increased Fgf9 expression results in pulmonary mesenchymal hyperplasia and a multicystic architecture that is histologically and molecularly indistinguishable from Type I PPB . We further show that miR-140 is expressed in lung epithelium , regulates epithelial Fgf9 expression , and regulates pseudoglandular stages of lung development . These studies identify an essential miRNA-FGF9 pathway for lung development and a non-cell autonomous signaling mechanism that contributes to the mesenchymal hyperplasia that is characteristic of Type I PPB .
Fibroblast Growth Factor 9 ( FGF9 ) is required during lung development for mesenchymal growth and epithelial branching , and inactivation of Fgf9 in mice results in perinatal death due to respiratory insufficiency [1–3] . Overexpression of FGF9 in embryonic mouse lung epithelium results in cystic expansion of the small airspaces , increased mesenchymal proliferation , and diminished mesenchymal differentiation [2 , 3] . Interestingly , the phenotype of lungs that overexpress FGF9 during development closely resemble those seen in mouse lungs lacking epithelial Dicer1 , an RNase III protein that is required for the biogenesis of microRNAs ( miRNAs ) , small RNA molecules that most commonly bind to specific sequences in the 3’ UTR of mRNA molecules targeting them for destruction or inhibiting their translation [2 , 4–6] . The phenotypic similarity between overexpression of Fgf9 and loss of epithelial Dicer1 suggested that microRNA modulation of Fgf9 expression could be an essential mechanism regulating lung development and that deregulation of Fgf9 expression could lead to developmental abnormalities or other diseases . Pleuropulmonary blastoma ( PPB ) , the most common primary malignancy of the lung in children , is either solitary or multifocal , is often familial , and is seen in association with several nonpulmonary neoplasms including cystic nephroma , Wilms tumor , neuroblastoma , rhabdomyosarcoma , medulloblastoma , ovarian Sertoli-Leydig cell tumor , intestinal polyps , and thyroid nodules [7–15] . The earliest morphologic changes in the affected lung consist of a localized area ( s ) of cystic expansion of alveolar spaces and uncommitted small mesenchymal cells within the expanded alveolar septa ( Type I PPB ) . The interstitial mesenchyme in these early lesions resembles that of the late pseudoglandular stage of lung development [16] . The initial multicystic lung lesion in Type I PPB can progress to a complex primitive sarcoma , which is first recognized by the presence of primitive mesenchymal cells , often with rhabdomyosarcomatous features , arranged in a dense layer ( cambium ) beneath a benign appearing epithelium . The risk for progression in a purely cystic Type I PPB peaks in the first 5 years of life [16 , 17] . Genetic linkage analysis of familial PPB and related disorders identified loss-of-function mutations in DICER1 [8 , 18] . DICER1 is normally expressed in both epithelial and mesenchymal tissues of human ( and mouse ) lung [18 , 19] . Immunostaining for DICER1 showed loss of , or decreased , staining in PPB-associated lung epithelium in a proportion of Type I PPBs but consistent expression in the underlying mesenchyme [18] . We hypothesized that the epithelial and mesenchymal phenotype of early PPB results from focal loss of functional DICER1 in lung epithelium and that deregulation of an epithelial factor would non cell-autonomously affect subepithelial mesenchyme . Of several secreted signaling molecules that are important for lung development , we considered FGF9 a plausible candidate as it is expressed in lung epithelium and mesothelium in early embryonic development and has the capacity to signal to both mesenchyme , where it regulates proliferation and inhibits differentiation , and epithelium where it affects branching and directly induces epithelial dilation [1–3 , 20–22] . Here , we show that the lung phenotype caused by loss of epithelial Dicer1 is dependent on Fgf9 , as it can be partially rescued by reducing the gene copy number of Fgf9 . We also show that the Fgf9 3’ UTR is responsive to conserved miRNA-140 , miRNA-328 , and miR-182 , and that miRNA-140 ( and miR-328 ) is an important regulator of lung development . Strikingly , we find that FGF9 is highly expressed in the epithelium of Type I PPBs in humans and in mouse embryonic lung epithelium that conditionally lack Dicer1 . These studies thus identify FGF9 as a developmentally essential downstream target of epithelial DICER1-cleaved miRNAs during lung development and as a candidate “tumor promoting factor” for PPB .
To determine whether lung epithelial microRNAs regulate molecules that affect lung mesenchyme development , we used a ShhCre knockin allele [23] to inactivate a floxed allele of Dicer1 [24] specifically in lung epithelium . At embryonic day 10 . 5 ( E10 . 5 ) , ShhCre/+ , Dicer1f/f lungs were histologically and phenotypically normal; however , E12 . 5 , ShhCre/+Cre/+/+ , Dicer1f/f lungs were larger than controls , with dilated epithelial ducts , reduced branching , and substantially expanded mesenchyme ( Fig 1A , 1C , 1E , and 1G ) . E14 . 5 and E16 . 5 , ShhCre/+ , Dicer1f/f lungs were of similar size to controls but revealed marked cystic expansion of the epithelial ducts and decreased branching ( S1 Fig ) . Because of the early inactivation of Dicer1 with ShhCre/+ and the rapid progression of the phenotype , we focused most subsequent analyses on the E12 . 5 time point . The striking expansion of mesenchyme in lungs lacking epithelial Dicer1 suggested activation of a non-cell autonomous epithelial-derived signal . Because of the established role of FGF9 signaling to lung mesenchyme [1 , 2] , we considered Fgf9 as a plausible candidate for this signal . In situ hybridization and qRT-PCR examination of E12 . 5 lung showed increased Fgf9 expression , primarily localized to epithelium , in ShhCre/+ , Dicer1f/f compared to control lungs ( Fig 1O–1Q ) . To compare phenotypes resulting from ectopic overexpression of epithelial Fgf9 with epithelial loss of Dicer1 , we forced expression of FGF9 in lung epithelium by crossing the Sftpc-rtTA transgenic mouse line with mice containing a doxycycline inducible Fgf9 transgene , Tre-Fgf9-Ires-Gfp ( Tre-Fgf9 ) . These double transgenic mice could be induced to express FGF9 in lung epithelium in the presence of doxycycline [2] . Induction of FGF9 expression from E10 . 5 to E12 . 5 showed a similar phenotype to lungs lacking epithelial Dicer1 ( Fig 1B , 1C , 1F , and 1G ) . Consistent with induction of a primary epithelial to mesenchymal signal in both ShhCre/+ , Dicer1f/f lungs and doxycycline-induced Sftpc-rtTA , Tre-Fgf9-Ires-eGfp lungs , analysis of cell proliferation showed no significant change in epithelial proliferation at E12 . 5 , but significantly increased mesenchymal proliferation ( Fig 1I–1K , 1M , and 1N ) . Previous studies on lung development identified a feed forward signaling network that linked FGF9 with mesenchymal FGFR and Wnt/β-Catenin signaling and Fgfr1 and Fgfr2 expression [3 , 25] . In this network , FGF9 and mesenchymal FGFR signaling regulated the expression of the canonical Wnt ligand , Wnt2a , and activated mesenchymal Wnt/β-Catenin signaling . Mesenchymal Wnt/β-Catenin signaling , driven by mesenchymal Wnt2a and epithelial Wnt7b , was required for the expression of mesenchymal Fgfr1 and Fgfr2 [3 , 21] . To determine if this feed forward loop was activated in lung tissue lacking epithelial Dicer1 , we examined expression of Wnt2a and the WNT-responsive transcription factor , Lef1 . Compared to control lung ( Fig 2A and 2D ) , ShhCre/+ , Dicer1f/f lungs showed increased Wnt2a and Lef1 expression ( Fig 2B and 2E ) . These observations support a model in which upregulation of Fgf9 in ShhCre/+ , Dicer1f/f lungs activate mesenchymal FGF-Wnt/β-Catenin signaling . If increased Fgf9 expression were a primary factor mediating the phenotype resulting from epithelial inactivation of Dicer1 , partial rescue would be expected following epithelial-specific reduction of Fgf9 gene dosage . To test this , we generated mouse embryos with the genotype ShhCre/+ , Dicer1f/f , Fgf9f/+ in which one allele of Fgf9 was inactivated specifically in lung epithelium that also lacked both copies of Dicer1 . By itself , heterozygosity for Fgf9 has no effect on development [1–3] . Compared to ShhCre/+ , Dicer1f/f littermates , genetic inactivation of one allele of Fgf9 in lung epithelium significantly reduced lung size and epithelial dilation ( Fig 1C and 1D ) . Examination of lung histology revealed reduction in both epithelial airspace dilation and mesenchymal thickness ( Fig 1G and 1H ) . Immunostaining for phospho-Histone H3 ( pHH3 ) positive cells showed a reduction in mesenchymal proliferation in ShhCre/+ , Dicer1f/f , Fgf9f/+ compared to ShhCre/+ , Dicer1f/f , Fgf9+/+ lungs ( Fig 1K , 1L and 1N ) . Consistent with the normalized phenotype resulting from inactivation of one allele of epithelial Fgf9 , Wnt2a and Lef1 expression levels were also normalized in ShhCre/+ , Dicer1f/f , Fgf9f/+ lung tissue ( Fig 2A–2F ) . We also examined potential phenotypic rescue of lung tissue lacking both alleles of Fgf9 in epithelium of mice lacking epithelial Dicer1 ( ShhCre/+ , Dicer1f/f , Fgf9f/f ) . At E14 . 5 , epithelial inactivation of both alleles of Fgf9 ( in the ShhCre/+ , Dicer1f/f background ) resulted in a smaller lung with reduced cystic dilation of epithelial ducts ( S2 Fig ) . This finding is consistent with inactivation of epithelial FGF9 compensating for the epithelial Dicer1 loss phenotype and endogenous mesothelial Fgf9 ( which is not affected in this model ) having a primary role in regulating lung mesenchymal development [21] . To further assess the contribution of FGF9 signaling to early stages of epithelial differentiation , we examined expression of Sftpc , an epithelial differentiation marker , which is first expressed in the epithelial branching tips at ~E12 . 5 . Compared to control lung , ShhCre/+ , Dicer1f/f lungs showed significantly reduced expression of Sftpc at E12 . 5 ( Fig 2G and 2H ) . However , in ShhCre/+ , Dicer1f/f , Fgf9f/+ lung , low-level expression of Sftpc was detected ( Fig 2I ) consistent with partial rescue of lung development . Quantitative RT-PCR on whole E12 . 5 lung also demonstrated a significant ( P<0 . 04 ) increase in the expression of Sftpc in ShhCre/+ , Dicer1f/f , Fgf9f/+ ( or Fgf9f/f ) lung compared to control lung ( Fig 2J ) . Sequence analysis of the FGF9 3’ UTR identified multiple potential miRNA binding sites that were highly conserved between human , chimp , mouse , and pig ( Fig 3A and S3 Fig ) . Importantly , several of these miRNAs ( miR-24 , miR-140 , miR-182 , miR-183 , miR-328 ) are expressed in fetal or neonatal lung and their relative expression levels are modulated during lung development [26 , 27] or in lung cancer [28–30] . Analysis of expression of these miRs at different stages of lung development showed that miR-140-5p and miR-328-3p were expressed at relatively low levels during the pseudoglandular stage of lung development and at relatively higher levels during the saccular stage , while miR-182-5p showed the opposite profile ( Fig 3B–3D ) . By comparison , Fgf9 expression levels were relatively high during pseudoglandular stage lung development and lower at late developmental stages ( Fig 3E ) . To functionally assay the FGF9 3’ UTR , we cloned the mouse and human UTR’s into the dual luciferase vectors psiCHECK-2 ( pFgf9UTR ) and pEZX-MT01 ( pFGF9UTR ) , respectively , and transfected into HEK293 cells . Mature microRNA mimics for miR-24 , miR-140 , miR-182 , miR-183 , and miR-328 were then screened for their ability to regulate luciferase activity of the human or mouse FGF9 3’ UTR . Of these , miR-140 , miR-183 , and miR-328 suppressed luciferase activity , while miR24 and miR-182 increased luciferase activity ( Fig 3F , mouse , and S4A Fig and S4B Fig , human ) . Because miR-140 demonstrated the strongest repressive effect on luciferase activity , this miRNA was investigated further . To establish specificity of miR-140 , we engineered mutations in the mouse and human FGF9 3’ UTR miR-140 seed sequences . Luciferase activity assays showed that the miR-140 mutant 3’ UTRs no longer responded to co-transfection with the respective mature miR mimic ( Fig 3G and S4C Fig ) . To determine the primary cell-type expressing miR-140 , we hybridized locked nucleic acid ( LNA ) in situ probes to E12 . 5 whole lungs and E18 . 5 lung sections . Consistent with regulation of epithelial Fgf9 mRNA expression , miR-140 was prominently expressed in E12 . 5 lung epithelial ducts and E18 . 5 distal conducting airway epithelium and Type II pneumocytes ( Fig 3H–3M ) . miR-328 was also prominently expressed in E18 . 5 lung epithelium ( S5 Fig ) . Collectively , expression patterns and in vitro suppression of the Fgf9 3’ UTR identified miR-140 and miR-328 as candidate miRNAs that could function in vivo to suppress Fgf9 as lung development progresses from pseudoglandular to canalicular stages . To establish whether miR-140 and miR-328 functionally regulate lung development , lung explant cultures were treated with seed-targeting 8-mer LNA oligonucleotides or single mismatch control LNA oligonucleotides ( tiny LNAs ) [31] . To demonstrate efficacy of tiny LNAs , HEK293 cells , transfected with pFgf9 UTR and miR-140 or miR-328 mimics , were co-transfected with tiny LNA antagomers . At a concentration of 10 nM , tiny LNAs effectively blocked miR-140 or miR-328 ability to suppress Fgf9 3’ UTR activity in vitro ( Fig 4A and 4B ) . Using 6-FAM-labeled miR-140 tiny LNA , we also demonstrated efficient uptake into lung explant tissue 48 hr following exposure to media containing 100 nM tiny LNA ( S6 Fig ) , consistent with efficient uptake of other types of oligonucleotides into lung explant cultures [21] . Embryonic lungs , explanted at E10 . 5 , showed robust mesenchymal growth and epithelial branching over 48 hr in culture . In response to treatment with FGF9 , lung explants revealed increased mesenchymal thickness and epithelial airspace dilation [3 , 21 , 32] . Similar to lung explants treated with FGF9 , treatment of E10 . 5 explants with 100 nM anti-miR-140 or 50 nM of each , anti-miR-140 and anti-miR-328 , showed a significant ( P<0 . 01 ) increase in mesenchymal thickness and a decrease in epithelial branching compared to explants treated with mismatch control LNA oligonucleotides ( Fig 4C–4H ) . Finally , Fgf9 expression was evaluated in lung explant cultures treated with tiny LNAs . Lungs treated with the LNA-140 ( n = 4 of 4 ) or LNA-140 and LNA-328 ( n = 4 of 4 ) demonstrated increased expression of Fgf9 compared to treatment with control LNA in which only one of seven explants showed Fgf9 expression ( Fig 4I and 4J ) . These data indicated that during ex vivo lung development , miR-140 is sufficient to regulate Fgf9 expression in lung epithelium and regulate mesenchymal growth and epithelial branching . Identification of DICER1 loss in Type1 PPB-associated lung epithelium suggests that deregulation of a non-cell autonomous factor could initiate the pathological process leading to abnormal mesenchymal proliferation and subsequent oncogenic transformation . To test this , human Type I PPB tissue was immunostained for FGF9 , the proliferation marker , Ki67 and p-Erk . Robust epithelial FGF9 expression was observed in 13 of 16 cases ( 81% ) of Type I PPBs examined ( Fig 5A and 5B ) . Consistent with FGF9 signaling to mesenchyme , cell proliferation , as determined by Ki67 immunostaining and p-ERK expression was increased in subepithelial mesenchyme in all ( n = 10 ) cases assessed ( Fig 5C–5F , 5H and 5J ) . Examination of PPB-associated epithelium showed increased proliferation , but reduced p-ERK immunostaining ( Fig 5G and 5I ) in nine of eleven ( 81% ) cases assessed , suggesting direct consequences of DICER1 loss in epithelium and possible indirect effects of FGF9 secondary to increased mesenchymal proliferation . For comparison with similarly staged mouse lung , epithelial FGF9 expression was induced from E16 . 5 to E18 . 5 in Sftpc-rtTA , Tre-Fgf9-Ires-Gfp double transgenic embryos ( Fig 5K and 5L ) . Immunostaining for Ki67 showed increased proliferation compared to uninduced control mouse lung , within mesenchymal and epithelial compartments ( Fig 5M , 5N , 5Q and 5R ) . Similar to what was observed in the tissues of human PPBs , p-ERK expression was significantly increased in lung mesenchyme , but absent in associated epithelium ( Fig 5O , 5P , 5S and 5T ) . The pathologic findings in Type I PPB showed a characteristic expansion of primitive or uncommitted mesenchyme ( Fig 6A ) associated with a benign-appearing Nkx2 . 1 positive epithelium ( Fig 6B ) . In the mouse , induction of epithelial FGF9 expression from E16 . 5 to E18 . 5 resulted in mesenchymal hyperplasia beneath a benign-appearing Nkx2 . 1 positive epithelium ( Fig 6G and 6H ) with histological features that were virtually identical to those observed in Type I or cystic PPB ( Fig 6A and 6B ) . Control double transgenic embryos that were not exposed to doxycycline were phenotypically normal . Both in human Type I PPB and in mouse lung induced to express FGF9 , the continued expression of Nkx2 . 1 indicates that lung epithelial identity is retained . Progression towards malignancy often involves loss of cellular terminal differentiation . Examination of mesenchymal differentiation into peribronchiolar smooth muscle showed reduced expression of smooth muscle actin ( SMA ) in peribronchiolar locations in both Type I PPB tissue and mouse lung tissue induced to express FGF9 ( Fig 6C and 6I , white arrow ) . However , vascular SMA expression appeared normal in both mouse and human tissue ( Fig 6I , black arrowhead ) . Markers of proximal-distal epithelial differentiation were similarly altered in both human Type I PPB and FGF9-induced mouse lung . The proximal Club cell secretory protein , CC10 , expression was decreased in bronchiolar epithelium ( Fig 6D and 6J , white arrow ) , the distal alveolar Type II cell marker , Sftpc ( Fig 6E and 6K , white arrow ) and the alveolar Type I cell marker Aquaporin 5 ( Aq5 ) [33] or T1α [34] ( Fig 6F and 6L , white arrow ) were increased and expanded proximally . Thus , loss of DICER1 in human lung epithelial tissue or overexpression of FGF9 in late stage fetal mouse lung epithelium not only affects mesenchymal growth and differentiation , but also results in distal differentiation of epithelial cell types . The decreased expression of Sftpc in E12 . 5 lungs that lack Dicer1 ( Fig 2G , 2H and 2J ) likely represents a delay in epithelial differentiation at this stage of development .
We have shown that several conserved miRNAs can regulate the human and mouse FGF9 3’ UTR and that miR-140 regulates Fgf9 expression in developing lung . However , the relationship between miRNAs and Fgf9 may also have a role in the development and pathogenesis of other tissues . Mice that lack miR-140 are viable but exhibit decreased growth of long bones , attributed to reduced chondrocyte proliferation [35] . miR-140 is contained within intron 16 of the ubiquitin ligase , Wwp2 , which is expressed in chondrocytes and in epithelial tissues , including lung [36 , 37] . Although Fgf9 was not identified as a target of miR-140 in chondrocytes , FGF9 is known to functionally regulate bone growth in part by suppressing chondrocyte proliferation [38] and could therefore be a functional miR-140 target in developing bone . Interestingly , miR-140 and Wwp2 are both directly induced by Sox9 in chondrocytes , ATDC5 cells , and 293T cells [37] . Sox9 expression in distal lung epithelium [39] and the established role for Sox9 in chondrogenesis , suggests potentially interesting parallels between skeletal and lung development . miR-140 is also involved in the pathogenesis of several human malignancies , including breast , ovarian , non-small cell lung , basal cell , colon , osteosarcoma , and hepatocellular carcinoma [40–46] . In hepatocellular carcinoma , miR-140 functions as a tumor suppressor , where it directly suppresses Fgf9 expression [42] . In non-small cell lung carcinoma , miR-140 suppresses tumor growth and metastasis by downregulating IGF1R [41] , and in breast cancer , miR-140 targets Sox2 [44] . Interestingly , Fgf9 is expressed in 10% of human non-small cell lung carcinomas and induced expression of Fgf9 in adult mouse lung epithelium leads to the rapid formation of adenocarcinomas [47 , 48] . Thus , miR-140 suppression of Fgf9 may not only be important for the development of lung and other tissues , but it may also function as an important tumor suppressor to ensure the quiescence of Fgf9 in adult tissues . Increasing evidence suggests that miR-328 also functions as a tumor suppressor in several types of cancers , including malignant glioma , breast , and colorectal carcinomas [49–52] . In malignant gliomas ( World Health Organization grade IV astrocytic glioblastomas ) , miR-328 expression is decreased and is associated with worse prognosis [50] . Additionally , miR-328 showed reduced expression when comparing levels in grades II and III astrocytoma to those in secondary grade IV glioblastomas [49] . Fgf9 is a potent growth factor for glial cells and was originally isolated from a glioma cell line [53] . Insufficient miR-328 in glioblastomas could lead to increased FGF9 expression and thus provide a mechanism to promote disease progression . The human and mouse FGF9 3’ UTR are highly conserved and are similarly regulated by miR-140 , miR-182 , miR-183 , miR-328 . However , the human FGF9 3’ UTR differs from the mouse UTR in that it contains a microsatellite sequence and binding site for the RNA binding proteins FUBP3 and HuR . FUBP3 has been shown to potentiate FGF9 mRNA levels [54] . Although the mouse Fgf9 3’ UTR does not contain an HuR binding motif , FGF9 was shown to regulate HuR expression and HuR was shown to regulate lung branching morphogenesis through regulation of Fgf10 and Tbx4 expression [55] . Thus , these RNA binding proteins , miRNAs , and the FGF9 gene ( including its protein product and 3’ UTR ) may be involved in a common gene regulatory network that controls human lung development . Sarcomatous progression of the mesenchymal cells in Type I or cystic PPB appears to require bi-allelic mutations in DICER1 . These mesenchymal cells typically have one allelic loss of function mutation and one somatic RNase IIIb missense mutation , leading to an inability to process mature 5p miRNAs , but preservation of 3p miRNAs [56] . Additionally , evidence was found for TP53 inactivation occurring as a third genetic event in PPB in the solid sarcomatous foci of the Type II and Type III neoplasms [56 , 57] . The mesenchymal hyperplasia and either the increased proliferative index or increased number of mesenchymal cells resulting from Fgf9 activation in PPB-associated epithelium , coupled with second hit DICER1 RNase IIIb point mutations , could further enhance the oncogenic transformation of these mesenchymal cells . Inactivation of Dicer1 in developing mouse lung using the Shh-Cre/+ driver effectively models the earliest stages of PPB . Similarities include increased Fgf9 expression , mesenchymal hyperplasia , and cystic expansion of epithelial ducts . However , this mouse model does not recapitulate the disease progression seen in some examples of human PPB . This is likely due to the severity of the phenotype of the mouse model after E14 . 5 and the non-viability of these mice after birth . To examine the effects of FGF9 expression at later stages of development that better match the more advanced stages of human PPB , we used an inducible Fgf9 transgenic system . Activation of Fgf9 from E16 . 5 to E18 . 5 showed marked similarities to human PPB at both the histological and molecular levels . An additional difference between the Shh-Cre/+ , Dicer1f/f mouse model and PPB is that in familial PPB , DICER1 is haploinsufficient in all cells and lost in lesion-associated epithelium , whereas in the mouse model , Dicer1 is only inactivated in lung epithelium . Future refinements of the mouse model will be needed to reflect these differences in Dicer1 genetics , the multifocal nature of human PPB , and the ability to observe disease progression beyond initial disease stages . Manipulation of miRNA expression as a therapeutic target is under consideration for a wide range of human cancers [58] . Early stage PPB may be a particularly good target for miRNA directed therapy because sarcomatous progression in PPB , when it occurs , typically does so in the first five years of life [16] . Thus , inhibition of key targets of miRNAs during early childhood could slow progression or prevent events in the development of the cystic stage of PPB until after this developmental window of susceptibility . In the case of Fgf9 , it appears that miRNAs serve to downregulate Fgf9 expression during the transition from pseudoglandular to canalicular stages of development . In adult lung tissue , Fgf9 expression is very low and may be maintained in this low stage independent of miRNA regulation . Consistent with this model , a recent study showed that loss of lung epithelial Dicer1 at later stages of development does not result in PPB-like cystic morphology [59] . Furthermore , ectopic activation of Fgf9 in adult lung results in the rapid formation of adenocarcinoma , without associated mesenchymal hyperplasia [47] . This suggests that adult mouse lung mesenchyme becomes non-responsive to FGF9 . Our demonstration that miR-140 and miR-328 mimics can directly suppress the Fgf9 3’ UTR , shows the therapeutic potential of supplying critical microRNAs directly to lung epithelium during the period of childhood susceptibility to PPB .
All mouse strains , including Fgf9f/f , Dicer1f/f , ShhCre/+ , Tre-Fgf9-Ires-eGfp , Sftpc-rtTA , ( f , floxed allele ) , have been previously described [2 , 23 , 24 , 60 , 61] . For conditional inactivation of Dicer1 and Fgf9 in lung epithelium , mice were generated with the genotype , ShhCre/+ , Dicer1f/f , Fgf9+/+ and ShhCre/+ , Dicer1f/f , Fgf9f/+ . Control mice were of the genotype ShhCre/+; ShhCre/+ , Fgf9f/+; ShhCre/+ , Dicer1f/+ , Fgf9f/+; Dicer1f/+ , Fgf9f/+; or Fgf9f/+ , all of which showed no phenotypic differences from wild type mice . All loss of function mice were maintained on a mixed 129SV/J-C57BL6/J background . Transgenic strains , used for gain-of-function experiments , were maintained on the FVB background . The human Type I Pleuropulmonary Blastoma tissue samples , formalin fixed and paraffin embedded , were obtained through the genetic studies tissue bank of the International PPB registry ( http://www . ppbregistry . org/enrollment/genetic-studytissue-bank ) . Mouse embryo tissues were collected in ice cold PBS , fixed in 4% PFA overnight at 4°C , washed with 1X PBS , photographed , and embedded in paraffin prior to sectioning at 5 μm . For histology , mouse and human sample slides were stained with hematoxylin and eosin ( H&E ) . For immunohistochemistry , paraffin section or cryo-sections were rehydrated and treated with 0 . 3% hydrogen peroxide in methanol for 15 min to suppress the endogenous peroxidase activity . Antigen retrieval was achieved by microwaving the sections in 10 mM citrate buffer for 10 min followed by gradual cooling to room temperature . Sections were incubated overnight at 4°C with the following primary antibodies: NKX2 . 1 ( M3575 , DAKO , 1:200 ) ; FGF9 ( AF-273-NA , R&D , 1:100 ) ; Ki67 ( VP-K451 , VECTOR Laboratories , Inc . , 1:200 ) ; p-ERK ( 4370s , Cell Signaling Technology , Inc , 1:200 ) ; Surfactant Protein C ( Sftpc , AB3786 , EMD Millipore Corporation , 1:1 , 000 ) ; CC10 ( sc9722 , Santa Cruz , 1:200 ) ; Aquaporin 5 ( AQ5 , AB92320 , Abcam , 1:200 ) ; T1α ( 128370 , SDHB , 1:200 ) and pHH3 ( H9908 , Sigma , 1:200 ) . The anti-goat ( BA9500 , 1:200 ) and anti-syrian hamster biotin-conjugated ( 107065–142 , 1:200 ) antibody were from VECTOR and Jackson ImmunoResearch Lab , Inc . , respectively . All other antibodies were visualized using Broad Spectrum ( AEC ) Kit ( 95–9743 , Zymed Laboratories Inc . ) for mouse samples ( staining with red color ) and Broad Spectrum ( DAB ) Kit ( 95–9643 , Zymed Laboratories Inc . ) for the human samples ( staining with brown color ) . All staining patterns are representative of at least three cases of human samples or three mouse embryos . For quantification of Ki67 and p-ERK immunostaining , at least three individual tissue samples were included . For each tissue sample , three different slides were stained and analyzed , and for each slide , three 10X fields were counted for immunostained cells per 100 epithelial or mesenchymal cells . Statistical analysis was based on the three original tissue samples . In situ hybridization probes were from the following sources: Fgf9 [62] , Lef1 [63] , Wnt2a ( A . McMahon , Harvard University , Cambridge , MA , USA ) , Sftpc [64] . Digoxigenin-labeled LNA miRNA detection probes were obtained from Exiqon Inc . ( Scrambled-miR , #99004–01 ) , has-miR140-5p ( #21309–05 ) , has-miR-328 ( #38156–05 ) . miRNA in situ hybridizations were performed according manufacturer instructions ( http://www . exiqon . com/ls/Documents/Scientific/miRCURY-LNA-miRNA-ISH-Optimization-Kit-manual . pdf ) . cDNA-based probes were synthesized and labeled with a kit from Roche Applied Science . Whole mount in situ hybridization was performed as described [2 , 3] . Following color reaction and methanol dehydration , tissues were photographed and then cryo-sectioned ( 5 μm ) , mounted on slides and rephotographed . In situ hybridizations of tissue sections were performed as previously described [65] . All staining patterns are representative of at least three cases of human samples or three mouse embryos . Lung explant cultures were performed as described [2] . E10 . 5 embryonic lungs were dissected and cultured on Transwell filters ( Costar , Corning ) for 48 hours at 37°C , 5% CO2 . For miR inhibition with locked nucleic acids ( LNA ) , E10 . 5 lung explants were cultured with a total final concentration of 100 nM LNA in culture media . To quantify mesenchymal thickness , explants were photographed and mesenchymal thickness was measured using Canvas X software . Data shown is representative of at least three independent experiments . p values were calculated using the Student’s t-test and plotted as mean ± SD . For whole mount in situ hybridization , explants were cultured for 48 hr with LNAs , the lung explants were then lifted from the filters , fixed with 4% PFA over night at 4°C , and then processed for whole mount in situ hybridization . Total RNA was purified from Lung explant cultures or HEK 293T cells using Trizol Reagent ( #10296–010 , Life Technologies Corporation , USA ) or RNeasy Plus Micro Kit ( #74034 , Qiagen Inc . USA ) . cDNA was synthesized using the iScriptTMSelect cDNA synthesis Kit ( #170–8841 , BIO-RAD Laboratories , USA ) . mRNA expression was measured using TaqMan Fast Advanced Master Mix ( #4444557 , Life Technologies Corporation , USA ) and TaqMan assay probes . miRNA were purified using the miRVana miRNA Isolation kit ( AM1561 , Life Technologies Corporation , USA ) and the TaqMan miRNA Reverse Transcription kit ( #4366596 , Life Technologies Corporation , USA ) . mRNA expression was normalized to either HPRT or GAPDH . miRNA expression was measured using TaqMan assay probes ( mmu-miR-140-5p , #001187; mmu-miR-182 5p , #002599; mmu-miR-183 , 5p 002269; mmu-mir-328-3p , #000543 ) . Expression was normalized to endogenous U6 NA ( #001973 , Life Technologies Corporation , USA ) . All assays were run on an ABI 7500 Fast Real-Time PCR System . Technical triplicates were run for each sample . Data was analyzed using the ΔΔCT method . The human FGF9 3’ UTR in the pEZX-MT01vector was purchased from Genecopoeia ( Rockville , MD USA ) . The mouse Fgf9 3’ UTR ( nt 997–1538 from clone NM_013518 ) was excised from a T7 vector using SacI ( blunted with Klenow ) and NotI enzymes , and cloned in the psiCHECK-2 vector at the PmeI and NotI sites . The inserted 3’ UTR was confirmed by DNA sequencing . HEK293T cells where grown to 70% confluence in 12 well tissue culture plates and transfected with 50 ng plasmid DNA and 10nM microRNA mimics in Optimem medium ( 2 ml ) following Lipofectamine 2000 instructions . After 6 hr , the media was replaced with fresh media ( DMEM , 10% FBS ) . After 48 hr , cells were harvested and Luciferase activity was assessed on a Lumat LB 9507 luminometer ( Berthold Technologies ) using the Dual-Luciferase Reporter 1000 Assay System ( E1980 , Promega ) according to the manufacturer instructions . Each condition was assayed in triplicate and all experiments where performed at least two times . A mutant version of the Fgf9 3’ UTR , in which the miR-140 seed sequences was deleted , was generated using the QuikChange XL Site-Directed Mutagenesis kit ( Agilent Technologies ) using primers listed in S1 Table . miRIDIAN microRNA mimics ( Dharmacon ) were used to increase mature microRNA expression in HEK293 cells . Mimics were added to culture medium at a final concentration of 10 nM . MicroRNA mimics used are listed in S2 Table . Tiny LNA antimiR oligonucleotides were custom designed to target the seed sequence of microRNAs . Tiny LNAs were synthesized with a phosphorothioate backbone ( Exiqon , See S2 Table for sequences ) . Tiny LNAs were transfected in HEK293T cells as described previously [31 , 66] at a final concentration of 10 nM for in vitro validation experiments . For expression in lung explant cultures , tiny LNAs were added directly to the culture media at a total final concentration of 100 nM . The data are reported as the mean ± SD and changes with p values less than 0 . 05 were considered to be statistically significant . Data was analyzed using the unpaired Student’s t test . Numbers of mice used per group per experiment are stated in the figure legends . This study was carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . The protocol was approved by the Washington University Division of Comparative Medicine Animal Studies Committee ( Protocol Number 20130201 ) . All efforts were made to minimize animal suffering . Human tissues were obtained from the International Pleuropulmonary Blastoma ( PPB ) Registry ( http://www . ppbregistry . org/ ) . Human tissues were obtained from the International PPB registry with IRB approval from Children′s Research Institute , Children′s National Medical Center Human Research Protection Office; IRB #4603 , renewed with IRB electronic study #Pro00000315 .
|
Pleuropulmonary Blastoma ( PPB ) is a pediatric disease that presents with multifocal cystic lung lesions . Familial mutations in DICER1 , an essential gene for microRNA synthesis , predisposes to PPB and other related diseases . Loss or mutation of the second copy of DICER1 in developing airway epithelium is thought to initiate cyst formation and increase growth of the underlying mesenchyme . In later stages , additional genetic events in PPB mesenchyme ( mutations in the DICER1 RNase IIIb domain or in TP53 ) can lead to the formation of high-grade sarcomas . We hypothesized that loss of DICER1 function in lung epithelium leads to persistent overgrowth of mesenchyme ( and subsequent risk for malignancy ) , implicating an indirect tumor initiation mechanism . In this study , we show histological and molecular similarity in Type I PPB and mice lacking epithelial Dicer1 or overexpressing epithelial Fibroblast Growth Factor 9 ( Fgf9 ) , demonstrate increased FGF9 expression in Type I PPB and in Dicer1-deficient mouse lung epithelium , and show that Fgf9 mediates at least some of the pathology resulting from Dicer1 inactivation in lung epithelium . Finally , we show that specific lung epithelial microRNAs regulate Fgf9 . These studies identify FGF9 as a target of DICER1 in lung epithelium that functions as an initiating factor for PPB .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Fibroblast Growth Factor 9 Regulation by MicroRNAs Controls Lung Development and Links DICER1 Loss to the Pathogenesis of Pleuropulmonary Blastoma
|
The cell cycle is the fundamental process of cell populations , it is regulated by environmental cues and by intracellular checkpoints . Cell cycle variability in clonal cell population is caused by stochastic processes such as random partitioning of cellular components to progeny cells at division and random interactions among biomolecules in cells . One of the important biological questions is how the dynamics at the cell cycle scale , which is related to family dependencies between the cell and its descendants , affects cell population behavior in the long-run . We address this question using a “mechanistic” model , built based on observations of single cells over several cell generations , and then extrapolated in time . We used cell pedigree observations of NIH 3T3 cells including FUCCI markers , to determine patterns of inheritance of cell-cycle phase durations and single-cell protein dynamics . Based on that information we developed a hybrid mathematical model , involving bifurcating autoregression to describe stochasticity of partitioning and inheritance of cell-cycle-phase times , and an ordinary differential equation system to capture single-cell protein dynamics . Long-term simulations , concordant with in vitro experiments , demonstrated the model reproduced the main features of our data and had homeostatic properties . Moreover , heterogeneity of cell cycle may have important consequences during population development . We discovered an effect similar to genetic drift , amplified by family relationships among cells . In consequence , the progeny of a single cell with a short cell cycle time had a high probability of eventually dominating the population , due to the heritability of cell-cycle phases . Patterns of epigenetic heritability in proliferating cells are important for understanding long-term trends of cell populations which are either required to provide the influx of maturing cells ( such as hematopoietic stem cells ) or which started proliferating uncontrollably ( such as cancer cells ) .
The cell cycle is a process leading to cell division . It plays a critical role in tissue growth , development and regeneration of multicellular organisms . It consists of two critical phases: the S phase , in which the cell replicates its DNA , and the M phase where it divides in two progeny cells ( mitosis ) . These phases follow the G1 and G2 phases , respectively . After division , progeny cells usually re-enter the cell cycle and return to the G1 phase [1 , 2] . Depending on a variety of factors , they may become quiescent ( pass to the dormant G0 phase ) . One of the important biological questions is how the dynamics at the cell cycle scale , which is related to family dependencies between the cell and its descendants , affects cell population behavior in the long-run . We address this question using a “mechanistic” model , built based on observations of single cells over several cell generation , and then extrapolated in time . We follow a paradigm recently expressed among others by Sandler et al . [3] and Dolbniak et al . [4] stating that stochastic processes in cells are associated with fluctuations in mRNA [5] , protein production and degradation [6 , 7] , noisy partition of cellular components at division [8] , and other cell processes . Variability within a clonal population of cells originates from such stochastic processes , which may be amplified or reduced by deterministic factors [9] . Independently of recent approaches , our work has been inspired by earlier work of Darzynkiewicz et al . [10] , who analyzed cycling Chinese hamster ovary ( CHO ) cells using flow cytometry . They reported variability in G1 phase caused mainly by unequal division of cytoplasmic constituents into progeny cells , and the main conclusion was that the cell-cycle heterogeneity was generated mostly during cytokinesis and to a lesser degree during the G2 phase . These data influenced the mathematical models of Kimmel et al . [11] , and Arino and Kimmel [12] . In these models , the heterogeneity has been generated only by unequal division or RNA or cytoplasm , with cell growth and the cell cycle duration being deterministic functions of the birth-size of cell . Models involving cell cycle duration stochasticity followed , with the most recent one being ref . [4] . The latter model is a precursor of the present one , yet with a more limited scope and based only on literature data . Understanding of the complexity of cell-cycle dynamics and of the specific patterns of cell-cycle progression remains incomplete . Quantitative dynamic imaging combined with mathematical modelling has become an essential approach to understanding such complex dynamics [13] . The recently developed experimental FUCCI ( fluorescent ubiquitination-based cell-cycle indicator ) reporter system [14 , 15] allows continuous imaging of cell-cycle progression in single live cells . In this system , two distinct proteins CDT1 and GEMININ , fused to fluorescent markers , indicate the G0/G1 and S/G2/M phases of the cell cycle , respectively . FUCCI system has been used to investigate inheritance mechanisms in non-stimulated dividing mammalian cells [3] , as well as in reoxygenated [16] and X-ray-irradiated cells [17] . In ref . [3] the authors analyzed the correlation of cell-cycle phase durations between family members . Variability in cell-cycle duration is ubiquitous , and sources of heterogeneity such as extrinsic and intrinsic noise [7] or unequal division [18] have been reported . Division times may also be epigenetically regulated [19] . In the present paper , we analyzed experimental and modelled cell pedigrees to determine patterns of inheritance of cell-cycle phase durations for aggregated G1 and S/G2/M phases , based on dynamic imaging of live NIH 3T3 cells . Based on this , we developed an integrated model involving bifurcation autoregression to describe cell proliferation and cell-cycle phase durations , and an ordinary differential equation ( ODE ) system to describe single-cell protein dynamics . The idea of bifurcating autoregression is that each line of descent from an ancestral cell follows the autoregression model ( descendants inherit certain properties from the ancestor ) , while the inherited and environmental effects in progeny are correlated . We developed estimates of the parameters of bifurcating autoregression under lognormally distributed noise , given observed cell-cycle phase durations , and fitted single-cell protein trajectories to the ODE model . In this way we found correlations among parameters for single cells . We validated the model , using the cell pedigrees from dynamic imaging data . Using the validated model , we employed long-term simulations to address the long-term behavior of the population , including homeostasis , memory of initial conditions and heritability . Specifically , we were interested in how regulation mechanisms of the cell cycle may contribute to propagation of new genetic or epigenetic variants in the cell population . This seems important because it has been established that many disease processes in living organisms were caused by the replacement of original cell diversity by clones which either proliferate without control ( as in cancer ) , or dominate tissue-specific stem cells , limiting their resilience and ability to regenerate , as in aging bone marrow ( see [20] and references therein ) . This observation has been explored in a number of deterministic and stochastic models ( see review [21] ) . We summarized our findings using a version of the classical population-genetics Wright-Fisher model , with variable population size; examples and references can be found in [22 , 23] . This approach is also related to the branching process paradigm , although our cell proliferation model is not a classical branching process [24] . Remark We employ the following vocabulary convention for family relationships in cell pedigrees . Suppose cell A divides into cells X and Y; then X divides into L and M , while Y divides into N and P . We call the progeny cells of the same parent cell sibling pairs . X and Y , L and M , and N and P , are sibling pairs . Cells whose parents are siblings we call cousin pairs . Thus , L and N , and M and P , are cousin pairs .
The data at our disposal include single-cell observation of NIH3T3 cells using the FUCCI-2A system ( see Methods section ( Experimental procedure ) for details ) , under two different Fetal Bovine Serum ( FBS ) concentrations . Cells were grown in constant conditions for 72 hours . During the experiment , films were recorded in randomly selected areas . We collected data from 123 cell lineages , including 890 individual cells from eight recorded films for 15% FBS , and 69 cell lineages including 224 individual cells from five recorded movies for 10% FBS . Based on experimental results we discovered that cell-cycle duration is shorter when higher FBS concentration is used . Faster progression of cell cycle is caused mainly by speeding up of S/G2/M phases progression . Differences between these two experiments are mainly visible ( S7E Fig ) in the fraction of cells entering dormancy ( G0 ) . Also see Supporting Information ( S1 Text , Sensitivity of durations of cell-cycle phases to serum stimulation ) for a more detailed description of these results . Further analysis was performed for 15% FBS data , since the sample size was significantly higher than in the 10% FBS experiment , and our model is not focused on dormant cells . Interpretation of results obtained by cell-cycle model may be difficult because it is not known which effects are caused by population growth and which by correlations among family members . To separate these two effects we used a simplified model of cell cycle which does not include correlations among family members . In the model , cell cycle phase duration of each individual was drawn from lognormal distribution and the parameters were estimated based on experimental data . Long-term experiments were designed for different initial numbers of ancestors . Results for four cases introduced in Table 1 are presented in Table 2 . Estimated effective population size ( K ) is almost twice as high as that obtained using the complete model , which includes correlations among family members , and close to harmonic mean value calculated from simulated genealogies . To verify how the value of effective population size depends on correlations among family members we used the model presented in this paper , which can reproduce behavior of populations with varying correlations among family members . Parameters of the model were estimated , simulations were performed and K values were estimated for all possible cases for three different numbers of ancestor cells ( Fig 5 ) .
Many models of cell-cycle progression have been proposed in the literature: age-structured cell population models [34] , branching processes [4 , 34] , transition probability models [35–37] and other novel models [38] , many of them based on experimental data [4 , 13 , 34 , 38 , 39] . The importance of developing a fully integrated model with different sources of noise and heterogeneity was discussed in ref . [40] . This motivated us to develop heterogeneous population-growth models with protein dynamics included , such as the model in the present paper , and an earlier model in ref . [4] . The important feature of our model is that it reproduces most of the characteristics observed in experimental data , as is evident from Figs 1 , 2 and 3 . Our model is based in part on bifurcating autoregression [26] applied to cell-cycle phases and on ideas concerning cell-cycle regulation and unequal division that were developed by many authors; specifically , we refer here to the model by Kimmel et al . [11] . In the model , chance and deterministic elements contribute to its ability to accurately fit multiple facets of cell-cycle kinetics in a heterogeneous cell population . This type of model may be important not only for understanding the kinetics of cell proliferation but also for testing of the individual responses of the cell to stimuli , especially when such a response is cell cycle-dependent . The models should also be useful for predicting the growth rates of populations consisting of subgroups with different properties and/or in which epigenetic effects are strong . We know that tumor growth is the consequence of competition among a few cell populations . It seems that even a small difference in cell characteristics , such as the cell-cycle proliferation time [41] , may increase the ability of a cancer to survive chemotherapy and re-enter the cell cycle . One of the features of our approach is integration of in vitro experiments with statistics and in silico simulations . The function of the latter is to understand the long-term behavior of cell population given the set of rules ( i . e . the mathematical model ) inferred using statistical tools , based on a limited-time in vitro experiment . Questions that can be answered in this way include the homeostatic properties of the population growth . Specifically , do the rules of cell growth and division lead under constant environmental conditions to stabilization of the distributions of important cell characteristics , such as cell-cycle time and durations of cell-cycle phases , as well as concentrations of cell proteins ? A related question concerns the nature of the transients that emerge after a cell with extreme parameters becomes an ancestor of its own population . If this cell has a short cell cycle , will its progeny tend to dominate the population ? Based on simulations , in two extreme cases of initial cell-cycle time 13 . 6 h and 61 . 3 h , large differences in the population growth rate have been observed . Within the interval from 0 to 200 h , during which the cell-cycle duration in both populations returned to the equilibrium distributions , the descendants of the cell with the short cell-cycle length have formed a subpopulation n = 40 times larger than the descendants of the cell with the long cell-cycle duration . If these two sub-populations were mixed , the one originating from the ancestor with the shorter cell-cycle length would dominate the other . In bacterial cells , the importance of long-term dynamics of cellular populations was considered in recent studies [42 , 43] , in which mathematical models were supported by biological experiments using E . coli strains . In the first paper [42] , authors discovered that ( 1 ) condition-dependent change of mean cell-cycle time is strongly correlated with variability in cell-cycle durations; and ( 2 ) increase of the heterogeneity of generation times in a population may be the method to evolve to a higher population growth rate in a constant environment , which is partly parallel to our conclusions . In our case a higher variability of cell cycle times was observed in the population with shorter mean cell-cycle times ( 10% FBS , mean cell cycle time 21 . 6 h , MAD = 0 . 17 , CV = 0 . 25; 15% FBS , mean cell cycle time 20 . 4 h , MAD = 0 . 21 , CV = 0 . 30 ) . In the second paper [43] , a method to predict histories of single cells in an exponentially growing population was proposed . Analysis revealed that physiological differences in sister cells have a significant impact on individual cell histories and their contribution to the overall population-growth . As stated previously , the models of cell proliferation and the models of genetic change in populations have been historically based on two apparently contradictory hypotheses , i . e . the unlimited branching process and a completely constant population size , respectively . By necessity , when it became clear that somatic mutations in proliferating eukaryotic cells are important for growth rates , the population constancy assumption in genetic models has been relaxed . A seminal paper concerns the Wright-Fisher coalescent under exponentially growing population [44] , followed by a number of models developed for other growth patterns , such as the model in ref . [45] , where linear growth has been considered . Polanski and Kimmel in ref . [22] , developed computable expressions for a Wright-Fisher coalescent with arbitrary growth pattern ( originally due to [46] ) . We analyzed the experimental and modelled cell pedigrees to determine patterns of inheritance of cell-cycle phase durations for aggregated G1 and S/G2/M phases , based on dynamic imaging of live NIH 3T3 cells . We developed estimates of the parameters of bifurcating autoregression model with lognormal distributions given observed cell-cycle phase durations , and fitted single-cell protein trajectories to the ODE model and found correlations among parameters between single cells ( sib-sib , parent-progeny , and other ) . Parent-progeny and sib-sib correlations from the experimental data were well-reproduced by our modelling , as demonstrated by comprehensive comparisons . Results showed stronger inheritance of the S/G2/M duration compared to G0/G1 . Using the model developed , we simulated its transient and long-term behavior and interpreted it in the terms of population genetics . Long-term simulations demonstrated the model had homeostatic properties . However , progeny of a single cell with a short interdivision time had a high probability of eventually dominating the population , due to heritability of cell-cycle phases . Analysis of model simulations showed that an effect similar to genetic drift was present in the model; however , it was amplified by family relationships among cells . This was manifested by reduction of the effective population size compared to the standard Wright-Fisher model of drift . Such patterns of epigenetic heritability in proliferating cells are important for understanding long-term trends of cell populations which are either required to provide influx of maturing cells ( such as hematopoietic stem cells ) , or which relaxed controls and started proliferating uncontrollably ( such as cancer cells ) . Specifically , we investigated adherence of our simulations to the Wright-Fisher model . We found that after 300 h , population started from N ancestral cells consists of their descendants in random proportions similar as in the Wright-Fisher model with effective population size K much smaller than the census size ( straight count of individuals ) . This is different from the previous studies fitting cell population drift using a Moran model ( which may be considered as a version of Wright-Fisher ) , in which the effective population size was equal to the census size; cf . [47] and references therein . In addition , we investigated the dependence of K on the correlations existing in our model . As depicted in Fig 5 , K is largest for the case in which both parent-progeny and sib-sib correlations are close to 0 . Also in this case , K is almost identical to that obtained from the harmonic mean of population sizes at different times , which is the textbook method for computing the effective population size for expanding populations ( [29] , Equ . ( 2 . 13 ) ; also see Methods , Wright-Fisher model and the cell cycle model ) . For numerical comparisons , see Table 2 . These findings illustrate the importance of including the correlations in the model . The most important conclusion is that in the presence of family relations , the estimated effective population size K is smaller than that obtained by using the harmonic-mean law . Using parent-progeny and sibling correlations estimated from data , we obtain K that is about 45% lower . As a consequence , drift acts in cell populations stronger than under the strict Wright-Fisher model with population growth , which increases the impact of random fluctuations in such populations ( see the last section of Results ) . This seems to be of importance in two contexts , in which continued cell proliferation takes place . One of these is cancer growth , in which an initially small population expands and diversifies by somatic mutations but also epigenetic changes [48–50] . Genetic or epigenetic drift acts at the stage when the tumor is very small , but also in isolated secondary foci in some cancers . A very well-documented study of neutral evolution of this kind has been carried out for hepatocellular carcinoma [51] . The other context is healthy human hematopoiesis , in which a relatively small population ( ca . 10 , 000 cells ) of hematopoietic stem cells ( HSC ) proliferates throughout the lifetime , diversifying into a number of descendant lineages and producing about 109 mature blood cells per day [52] . In the course of infections , the HSC become activated and if the incidents recur , their number and heterogeneity may permanently decrease [20] , which also makes the healthy HSC less competitive if a malignant clone arises . Since HSC population is distributed among smaller bone-marrow neighborhoods called niches , drift is likely to act strongly in this population . Reduced K amplifies these effects . This may also be the case in development of other stem-cell types , such as in hippocampal neurogenesis . The role of heterogeneity in this system is becoming an intensive research focus [53] . One of the interesting phenomena is heterogeneity reduction with age , which hypothetically might be due to stem-cell population bottlenecks , which have been demonstrated using the branching process model by Li and co-workers [54] .
Replication-defective , self-inactivating retroviral constructs were used for establishing NIH3T3 FUCCI-2A cell line as described in ref . [55] . These cells stably express the Cdt1 and Geminin sequences coding for G1 and S/G2/M probes , fused to mKO2 and E2-Crimson fluorescent reporters , respectively . They are separated by a 2A sequence to allow post-translational cleavage and followed by a puromycin-resistance cassette for subsequent selection . Before recording , cells were seeded at 7–10% confluence ( 105 cells per well ) in a 6-wells plate ( Falcon ) , with white DMEM medium ( high glucose ) containing 1% Penicillin/Streptomycin , 10 mM HEPES and either 10% or 15% FBS . Cells were left undisturbed for 48 hours . For recording , cells were placed in a Zeiss Axiovert 200M microscope ( Zeiss ) with a 20X Ph objective . A culture chamber , temperature and CO2 controller ( Pecon ) were used to ensure constant suitable conditions for long-term recording of the cells . Images were recorded every 15 minutes for 72 hours , using a Coolsnap HQ/Andor Neo sCMOS camera . Cells were briefly illuminated with a FluoArc HBO lamp ( Zeiss ) at reduced intensity . Epifluorescence signals were recorded as follows: mKO2: 300 ms ( filter cube: Ex 534/20 –Di 552 –Em 572/38 ) , E2Crimson: 800 ms ( filter cube: Ex 600/37—Di 650—LP 664 ) . The modelling paradigm we employ is based on the hypothesis that the timing of major events in the cell cycle is heritable in proliferating cells . This timing and its heritability are controlled by an intricate mechanism , which has been partly elucidated [56] , but its details require more resolution than we can build into our model . Other processes , such as synthesis of FUCCI proteins , occur within this time . Another driving factor is unequal division of proteins among progeny cells . It has been demonstrated theoretically [12 , 57] and confirmed by fitting models to data [4 , 11] , that models based on similar hypotheses exhibit homeostatic properties . This amounts to regulatory feedbacks acting in the model . However , none of these models has been based on data of such resolution as the present one .
|
All cells in multicellular organisms obey orchestrated sequences of signals to ensure developmental and homeostatic fitness under a variety of external stimuli . However , there also exist self-perpetuating stem-cell populations , the function of which is to provide a steady supply of differentiated progenitors that in turn ensure persistence of organism functions . This “cell production engine” is an important element of biological homeostasis . A similar process , albeit distorted in many respects , plays a major role in cancer development; here the robustness of homeostasis contributes to difficulty in eradication of malignancy . An important role in homeostasis seems to be played by generation of heterogeneity among cell phenotypes , which then can be shaped by selection and other genetic forces . In the present paper , we present a model of a cultured cell population , which factors in relationships among related cells and the dynamics of cell growth and important proteins regulating cell division . We find that the model not only maintains homeostasis , but that it also responds to perturbations in a manner that is similar to that exhibited by the Wright-Fisher model of population genetics . The model-cell population can become dominated by the progeny of the fittest individuals , without invoking advantageous mutations . If confirmed , this may provide an alternative mode of evolution of cell populations .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"g1",
"phase",
"cell",
"division",
"analysis",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"population",
"genetics",
"simulation",
"and",
"modeling",
"physiological",
"processes",
"homeostasis",
"effective",
"population",
"size",
"bioassays",
"and",
"physiological",
"analysis",
"population",
"biology",
"research",
"and",
"analysis",
"methods",
"cell",
"analysis",
"biochemistry",
"biochemical",
"simulations",
"population",
"metrics",
"population",
"size",
"cell",
"biology",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"evolutionary",
"biology"
] |
2019
|
Mathematical modelling reveals unexpected inheritance and variability patterns of cell cycle parameters in mammalian cells
|
Multidrug resistant leprosy , defined as resistance to rifampin , dapsone and fluoroquinolones ( FQ ) , has been described in Mycobacterium leprae . However , the in vivo impact of fluoroquinolone resistance , mainly mediated by mutations in DNA gyrase ( GyrA2GyrB2 ) , has not been precisely assessed . Our objective was to measure the impact of a DNA gyrase mutation whose implication in fluoroquinolone resistance has been previously demonstrated through biochemical studies , on the in vivo activity of 3 fluoroquinolones: ofloxacin , moxifloxacin and garenoxacin . We used the proportional bactericidal method . 210 four-week-old immunodeficient female Nude mice ( NMRI-Foxn1nu/Foxn1nu ) were inoculated in the left hind footpad with 0 . 03 ml of bacterial suspension containing 5×103 , 5×102 , 5×101 , and 5×100 M . leprae AFB organisms of strain Hoshizuka-4 which is a multidrug resistant strain harboring a GyrA A91V substitution . An additional subgroup of 10 mice was inoculated with 5×10−1 bacilli in the untreated control group . The day after inoculation , subgroups of mice were treated with a single dose of ofloxacin , moxifloxacin , garenoxacin or clarithromycin at 150 mg/kg dosing . 12 months later mice were sacrificed and M . leprae bacilli were numbered in the footpad . The results from the untreated control group indicated that the infective inoculum contained 23% of viable M . leprae . The results from the moxifloxacin and garenoxacin groups indicated that a single dose of these drugs reduced the percentage of viable M . leprae by 90% , similarly to the reduction observed after a single dose of the positive control drug clarithromycin . Conversely , ofloxacin was less active than clarithromycin . DNA gyrase mutation is not always synonymous of lack of in vivo fluoroquinolone activity in M . leprae . As for M . tuberculosis , in vivo studies allow to measure residual antibiotic activity in case of target mutations in M . leprae .
Mycobacterium leprae is responsible for leprosy , that the World Health Assembly decided , in 1991 , to “eliminate as a public health problem” by the year 2000 . But , though a decreasing number of new cases registered each year ( ∼219 , 000 ) during the recent years [1] , [2] , it is generally admitted that the goal of leprosy elimination is far from being reached [3] . Future projections of the global leprosy burden indicates that 5 million new cases would arise between 2000 and 2020 , and that in 2020 there would be still 1 million people with WHO grade 2 disability due to leprosy . Reports from Asian countries with a high leprosy prevalence estimated rates of resistance at 15–20% to dapsone ( DDS ) and 3–8% to rifampin [4] , [5] . Some studies estimated around 50% of resistance to DDS in relapsing cases [6] , [7] . Although the exact magnitude of resistance to these drugs is difficult to assess , resistance in M . leprae is a concern particularly in relapsing multi-bacillary leprosy patients , by strongly reducing the possibilities of an effective treatment [1] , [8]–[10] . Quinolones are good candidates for the development of more powerful treatments of leprosy , as demonstrated for moxifloxacin which is the only drug other than rifampin to be consistently bactericidal against M . leprae in clinical trials [11] . Fluoroquinolones play a crucial role in the treatment for drug-resistant leprosy and single-lesion new cases [11] , but some quinolone-resistant M . leprae strains have been described [12] , [13] . The mode of action of quinolones against M . leprae has been clearly identified and mechanisms of resistance have been investigated [4] , [5] , [12]–[17] . Substitutions within a highly conserved region of GyrA and possibly GyrB , which are subunits of the tetrameric DNA gyrase ( A2B2 ) , the so-called quinolone resistance-determining region ( QRDR ) , are associated with the development of ofloxacin resistance in M . leprae , as demonstrated before in M . tuberculosis . Due to the lack of M . leprae growth in vitro , the exact impact of DNA gyrase mutations on fluoroquinolone susceptibility remains largely unknown . We previously demonstrated , using an enzymatic assay , that GyrA substitutions do not have the same impact on all the fluoroquinolones [15] . As an example , garenoxacin had the same inhibitory activity against M . leprae DNA gyrase carrying mutation implicated in resistance to ofloxacin as against wild-type enzyme , underscoring the potential advantage of this compound in leprosy [15] . The aim of the present study was to evaluate the bactericidal activities of several quinolones ( i . e . ofloxacin , moxifloxacin and garenoxacin ) against a M . leprae strain carrying an A91V GyrA substitution known to be involved in quinolone resistance [18] . Relation between GyrA A91V mutation and ofloxacin resistance has been extensively proven in the literature both in patients experiencing relapse after ofloxacin treatment ( notably for strain for the strain used in the present work ) [12] , [18] , [19] and through DNA gyrase inhibition in vitro [15] . We demonstrated that despite GyrA A91V mutation , garenoxacin and moxifloxacin maintained some in vivo activity .
The laboratory has been approved on April , 24th 2012 to carry out animal experiments . Nicolas Veziris who carried the animal experiments has the following license number: 75-1531 . Aurélie Chauffour who performed the animal experiments has the following license number: B-75-1214 . We followed the animal experiment guidelines of the Faculté de Médecine Pierre-et-Marie Curie . Animal experiments were performed in accordance with prevailing regulations regarding the care and use of laboratory animals by the European Commission . The experimental protocol was approved by the Departmental Direction of Veterinary services in Paris , France . Hoshizuka-4 strain is a multidrug resistant strain with mutation in the three main genes involved in M . leprae resistance to antibiotics: folP gene ( P55S ) , involved in dapsone resistance; rpoB gene ( S456L ) , involved in rifampicin resistance; and gyrA gene ( A91V ) involved in quinolones resistance ( using numbering system of the M . leprae genome TN , GenBank nuNC002677 ) . Briefly , Hoshizuka-4 strain was isolated from a patient who developed a lepromatous leprosy after repeated clinical relapses [18] . He received dapsone , streptomycin , rifampin , clofazimine , isoniazid , ofloxacin and prothionamide to treat subsequent relapses . These drugs were administrated irregularly as monotherapy or in combinations , often at doses below recommended levels and standard multidrug therapy was never applied . The drug resistant profile of the isolated strain was confirmed by the mouse footpad method ( in nude mice ) for fluoroquinolones ( two fluoroquinolones , ofloxacin and sparfloxacin , were tested at 2 concentrations , 0 . 0001% and 0 . 001% ) [18] . The GyrA A91V substitution corresponds to amino acid 90 and 83 in M . tuberculosis and E . coli numbering system , respectively . Four week-old Nude mice ( NMRI-Foxn1nu/Foxn1nu ) were purchased from JANVIER breeding center , Le Genest Saint-Isle , France . Ofloxacin was purchased from Sanofi-Aventis , France; moxifloxacin from Bayer Santé , France; garenoxacin from EasyBuyer LTD , China , and clarithromycin from Abbot France , France . Animal experiments were performed in accordance with prevailing regulations regarding the care and use of laboratory animals by the European Commission . The experimental protocol was approved by the Departmental Direction of Veterinary services in Paris , France . The ‘proportional bactericidal’ technique , described by Colston , allows measuring the bactericidal activity of a compound [20] . Mice are inoculated with serial 10-fold dilutions of the suspension of M . leprae . A group of mice were left untreated; the other mice are treated for a period of time that varies from a single dose to 60 days ( usually 10 mice per dilution of inoculum for each treatment-group ) . After treatment , the mice are held for 12 months , to permit a single surviving organism to multiply to a readily detectable level ( M . leprae divides every 14 days ) . Harvests of M . leprae are then performed from individual feet; the organisms are considered to have multiplied in those feet found to contain ≥105 AFB . The proportion of viable M . leprae surviving treatment may then be calculated from the number of organisms that infects 50% of the mice . The proportion of viable M . leprae killed by the treatment is calculated by comparing the proportion of viable organisms in the treated mice to that in the control mice . Two hundred and ten 4 week-old immunodeficient female Nude were divided among 5 groups , each containing four subgroups with 10 mice each . The mice of each subgroup were inoculated in the left hind footpad with 0 . 03 ml of bacterial suspension containing 5×103 , 5×102 , 5×101 , and 5×100 M . leprae organisms of strain Hoshizuka-4 [18] . The suspension needed to inoculate mice was prepared from one footpad harvested from a nude mouse ( according to the Shepard and Mac Rae method [21] ) , that had been inoculated one year earlier in the lab with 6 . 104 AFB/footpad . Ten µl of the suspension were taken to create slides and AFB/footpad were counted after Ziehl-Neelsen coloration and ten-fold dilutions were made if it was necessary . All further ten-fold dilutions were made in Hank's balanced salt . A fifth subgroup of the untreated control group was inoculated with 5×10−1 Acid Fast Bacilli ( AFB ) per footpad . The day after inoculation , a table of randomization was created on website randomization . com in order to randomly allocate mice in different groups: untreated control , 10 mice per inoculum concentration from 5 . 103 to 5 . 10−1 AFB ( n = 50 ) ; treated mice , 10 mice for each inoculum ranging from 5 . 103 to 5 . 100 AFB ( n = 40 ) and for each antibiotic: ofloxacin 150 mg/kg; moxifloxacin 150 mg/kg; garenoxacin 150 mg/kg; and clarithromycin 150 mg/kg included as a positive control . Single dose was given the day after inoculation and randomization , and all drugs were administrated by gavage in 0 . 2 mL sterile water as a single dose of antibiotic . Mice were held for 12 months , a period of time sufficient to permit multiplication of a single surviving organism to multiply to a readily countable level . At the end of this period , tissues from the footpad were removed aseptically and homogenized in a final 2 ml volume of Hank's solution as described by Shepard and McRae method [20]–[22] . M . leprae was considered to have multiplied ( i . e . , viable organisms survived the treatment ) in those footpads found to contain ≥105 bacilli . Total DNA was extracted from footpad of all mice inoculated with 5 . 103 AFB , following the heat-shock procedure [23] . DNA was subjected to PCR amplifying the QRDRs in gyrA as previously described [10] and in gyrB using the following primers: GyrBlepS: 5′ ACG AGA GTT AGT GCG TCG AAA 3′ and GyrBlepAS: 5′ GCT GCG CTA AAA ACA CGT AC 3′ . Typical reaction mixtures ( 50 µl ) contained 0 . 5× reaction buffer , 2 , 5 mM of MgCl2 , 0 , 25 mM of dNTPs , 0 , 4 µM of each primer ( Eurofins MWG operon ) , 0 , 01 U of Taq polymerase ( BIO X ACT SHORT TAQ POL , BIOLINE , France ) and 5 ml of DNA extract . PCR-amplified fragments were purified by using QIAGEN DNA purification kit ( QIAGEN , France ) and sequenced by the dideoxy-chain termination method with the ABI PRISM BigDye Terminator v3 . 1 Cycle Sequencing Kit ( Applied Biosystems , Courtaboeuf , France ) . The oligonucleotide primers used for DNA sequencing were those used for PCR . The nucleotide and deduced amino acid sequences were analyzed with the Seqscape v2 . 0 software ( Applied Biosystems ) . The proportion of viable M . leprae organisms remaining after treatment was determined as the 50% infectious dose , and the significance of the differences between the groups was calculated by the method of Spearman and Kärber [22] . For multiple comparisons between the groups , Bonferroni's correction was applied , i . e . , the difference would be significant at the 0 . 05 level only if the P value adjusted to the number of groups: 0 . 05/n in which n was defined as the number of primary comparisons . Thus the corrected P was 0 . 05/5 = 0 . 01 .
Fifty five mice died during the study . Twenty seven mice died due to their advanced age . Twenty eight died due to an accidental problem of water supply during the experiment: 4 mice in ofloxacin 5 . 103 group , 7 mice in untreated control group 5 . 102 , 1 mouse in garenoxacin 5 . 101 group , 1 mouse in garenoxacin 5 . 100 group , 3 mice in clarithromycin 5 . 100 group and 1 mouse in untreated control 5 . 10−1 group . Results are presented in table 1 . In the untreated control group there were 22 . 6% viable bacilli at the end of the 12 months . Clarithromycin killed 90% of viable bacilli , ofloxacin 73% , moxifloxacin 90% and garenoxacin 88% . Compared to untreated control group , the percentage of viable bacilli was significantly smaller in the following treated groups: p = 0 . 0005 for clarithromycin , p = 0 . 0005 for moxifloxacin , p = 0 . 0009 for garenoxacin . For ofloxacin the percentage of viable bacilli was smaller than that of control group but not after Bonferroni correction ( p = 0 . 014 ) . On the other hand , the percentage of viable bacilli was similar in the group treated by ofloxacin compared to groups treated by moxifloxacin and garenoxacin ( p = 0 . 276 and 0 . 334 ) . Clarithromycin was as active as garenoxacin ( p = 0 . 723 ) and moxifloxacin ( p = 0 . 757 ) . Clarithromycin was more active than ofloxacin but not after Bonferroni correction ( p = 0 . 034 ) . No mutation in gyrA or gyrB was found in mice footpads demonstrating the absence of second-step mutant selection in our experiment . This result is not surprising since a single pulse drug is unlikely to result in selection of second step mutations .
Phenotypic assessment of M . leprae drug resistance is usually done using the continuous method in the mouse footpad model [10] , [18] , [24] . This method does not allow assessing bactericidal activity . This study is the first , to the best of our knowledge , assessing in vivo the bactericidal activity of various antibiotics of the same family against a M . leprae strain carrying mutation involved in drug resistance . Although multiple doses of treatment have been used by others , we chose to treat mice with a single dose of fluoroquinolones in order to be able to compare our present results to our previously published results [25] , [26] . We demonstrated that despite the presence of a GyrA substitution well known to confer FQ resistance , i . e . A91V [10] , [13] , [15] , the 3 fluoroquinolones tested had still some in vivo activity ( Table 1 ) . Ofloxacin activity was marginally significant , but garenoxacin and moxifloxacin remained active . We don't believe that mortality due to the water supply problem biased significantly the ofloxacin results because , for all groups of mice , including the untreated control group , there was no difference between 5 . 102 and 5 . 103 inocula , and all of mice inoculated with 5 . 102 and 5 . 101 and receiving ofloxacin were positive . Therefore it's highly probable that all mice would have been positive in 5 . 103 ofloxacin . Also the rank of activity between ofloxacin and moxifloxacin seen in wild-type strains in a previous study was maintained [27] . Despite the general rule of cross resistance between quinolones , garenoxacin , a new non-fluorinated quinolone , retains most of its activity against strains harboring QRDR mutations , in species such as Streptococcus pneumoniae and Helicobacter pylori [28]–[31] . This characteristic , combined with a lower rate of resistant mutant and favorable PK-PD parameters contribute to its higher activity against strains harboring DNA gyrase mutations compared to other quinolones . We previously demonstrated that garenoxacin has the same inhibitory activity against purified M . leprae DNA gyrase carrying mutation implicated in quinolone resistance than against the wild-type enzyme [15] . Although less active than moxifloxacin against wild-type strains [32] , garenoxacin is as active as moxifloxacin against the GyrA A91V mutant . Garenoxacin is currently under development in several countries [33] . Surprisingly , moxifloxacin also retained most of its activity against the strain harboring the GyrA A91V substitution , despite this substitution reduces moxifloxacin inhibition activity against purified gyrase [15] . Two parameters could explain the maintained activity of moxifloxacin against the strain harbouring the GyrA A91V substitution . First , it should be kept in mind that moxifloxacin is more active than ofloxacin [25] and garenoxacin [32] against wild-type M . leprae [27] . More importantly , it is likely that moxifloxacin MIC against the mutant strain remains lower than moxifloxacin peak serum level , thus allowing some in vivo activity as already shown for M . tuberculosis [34] . Ofloxacin is naturally less active than moxifloxacin and garenoxacin against M . leprae [25] , [27] , [32] . In other bacterial species like M . tuberculosis , ofloxacin is also less active than mo , xifloxacin against susceptible strains and this difference remains also against strains harboring DNA gyrase mutations [35] . In other words , in case of DNA gyrase mutation , susceptibility decreases for both ofloxacin and moxifloxacin but the difference of activity between these 2 antibiotics remains the same and explains why despite maintained moxifloxacin activity , ofloxacin was only marginally active against M . leprae . How these activities observed in a murine model can be translated in humans ? We used Nude mice ( NMRI-Foxn1nu/Foxn1nu ) rather than Swiss or conventional BALB/c because this species is more sensitive for the detection of antibiotic activity [36] . As the strain used was multidrug resistant we could not use dapsone nor rifampin as controls and consequently chose clarithromycin . In a previous study conducted in Swiss mice inoculated with a wild-type M . leprae strain , the killing rate was lower for clarithromycin ( 75% ) than for moxifloxacin ( 92% ) [25] . In the present study , the killing rates were equivalent between these two antibiotics ( Table 1 ) . Thus GyrA A91V substitution reduced the activity of moxifloxacin in vivo that became equivalent to that of clarithromycin . In human , clarithromycin although less active than rifampin , has shown bactericidal activity [37] . Thus in case of multidrug resistance , moxifloxacin could still be used in combination with clarithromycin , in a second-line drug scheme against GyrA A91V mutants . We showed a similar phenomenon in M . tuberculosis in which the substitution A90V ( equivalent to A91V in M . leprae ) downgrades the bactericidal moxifloxacin into a bacteriostatic drug in immunocompetent Swiss mice [34] . An important point regarding translation of mouse results to human is the dosing of antibiotics used . The 150 mg/kg dosing used generates , in the mouse , an AUC ( Area Under the Curve ) equivalent to the 400 mg dosing in human for the 3 fluoroquinolones [34] , [38]–[42] . For clarithromycin , the 150 mg/kg dosing generates an AUC equivalent to 500 mg twice-daily human dosing [43] , [44] . Thus , for all tested drugs , the AUC , which is the main PK parameter predicting efficacy , was equivalent to AUC in human at the usual dosing of the antibiotic . Taken together these data indicate that the GyrA A91V substitution confers low-level fluoroquinolone resistance in M . leprae and that moxifloxacin can be used in humans against such mutant strains . Molecular tools are more and more described and used for the diagnosis of drug resistance in leprosy [10] , [13] , [19] , [45] . We demonstrated in the present study that detection of a mutation is not sufficient to exclude a drug from therapeutic regimen , especially when there are a few or no other alternatives . Regarding fluoroquinolones , the present study is relevant for leprosy since , in M . leprae , the GyrA A91V substitution is the most prominent substitution described in the literature [13] , [19] , [45] , [46] , while the other substitution found in M . leprae GyrA ( G89C substitution , corresponding to G88C in the M . tuberculosis numbering system ) has been described only once [13] . Based on results from enzymatic studies performed on M . leprae and M . tuberculosis , the latter substitution should decrease susceptibility to fluoroquinolones at least at the level obtained with GyrA A91V substitution and could lead to a high-level resistance phenotype [35] . Thus the drug-resistance level generated by DNA gyrase mutation probably differs depending on the mutations and the therapeutic consequences should also differ . Equivalent in vivo experiments with rpoB and folP mutants would be important in order to measure the in vivo impact of drug resistance on rifampin and dapsone activity . Regarding rifampin , low-level resistance has been extensively studied in M . tuberculosis . For example , M . tuberculosis strains harboring rpoB L533P mutation display low-level rifampin resistance and the same mutation described in M . leprae [13] would deserve further evaluation .
|
Although there is efficient multidrug therapy to cure leprosy , the transmission of M . leprae is still active , leading to 219 , 000 new cases in 2011 . Drug resistant leprosy has been described and may prevent eradication of the disease , notably multidrug resistant defined as resistance to rifampin , dapsone and fluoroquinolones ( FQ ) . Resistance to FQ is due to mutations in DNA gyrase . We used a mouse model to measure the impact of DNA gyrase mutations on in vivo FQ activity . All the FQ tested showed in vivo activity against the mutant tested ( A91V mutant in subunit A of DNA gyrase ) . However , whereas ofloxacin was less active than the control treatment clarithromycin , it appeared that latter generation fluoroquinolones moxifloxacin and garenoxacin were as active as clarithromycin . Our results demonstrate that DNA gyrase mutation is not synonymous of total lack of in vivo FQ activity against M . leprae . Therefore , as for M . tuberculosis , in vivo studies are mandatory in order to measure the impact of DNA gyrase mutations on treatment efficacy against M . leprae .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2013
|
Resistance of M. leprae to Quinolones: A Question of Relativity?
|
Triatomine bugs are the vectors of Trypanosoma cruzi , the agent of Chagas disease . Vector control has for decades relied upon insecticide spraying , but insecticide resistance has recently emerged in several triatomine populations . One alternative strategy to reduce T . cruzi transmission is paratransgenesis , whereby symbiotic bacteria are genetically engineered to produce T . cruzi-killing proteins in the vector’s gut . This approach requires in-depth knowledge of the vectors’ natural gut microbiota . Here , we use metagenomics ( 16S rRNA 454 pyrosequencing ) to describe the gut microbiota of field-caught Triatoma sordida–likely the most common peridomestic triatomine in Brazil . For large nymphs ( 4th and 5th stage ) and adults , we also studied separately the three main digestive-tract segments–anterior midgut , posterior midgut , and hindgut . Bacteria of four phyla ( 12 genera ) were present in both nymphs ( all five stages ) and adults , thus defining T . sordida’s ‘bacterial core’: Actinobacteria ( Brevibacterium , Corynebacterium , Dietzia , Gordonia , Nitriliruptor , Nocardia , Nocardiopsis , Rhodococcus , and Williamsia ) , Proteobacteria ( Pseudomonas and Sphingobium ) , and Firmicutes ( Staphylococcus ) . We found some clear differences in bacterial composition and relative abundance among development stages; overall , Firmicutes and Proteobacteria increased , but Actinobacteria decreased , through development . Finally , the bacterial microbiotas of the bugs’ anterior midgut , posterior midgut , and hindgut were sharply distinct . Our results identify the ‘bacterial core set’ of T . sordida and reveal important gut microbiota differences among development stages–particularly between 1st–3rd stage nymphs and adults . Further , we show that , within any given development stage , the vectors’ gut cannot be regarded as a single homogeneous environment . Cultivable , non-pathogenic ‘core’ bacterial species may now be tested as candidates for paratransgenic control of T . cruzi transmission by T . sordida .
Chagas disease is a potentially life-threatening illness caused by the protozoan Trypanosoma cruzi . T . cruzi is a parasite of mammals primarily transmitted through the feces of infected vectors–blood-sucking bugs of the subfamily Triatominae . Chagas disease is the fourth most important infectious disease in the Americas , with about 8 million people infected and at least 120 million people living at risk of contagion [1] . Triatomines are hemimetabolous insects with five immature nymphal stages between egg and adulthood . Adults are sexually mature and have wings , but both nymphs and adults display similar feeding behavior and occupy the same habitats . As all stages feed on vertebrate blood , they are all prone to acquiring and transmitting T . cruzi [2] . Once the parasite arrives at the triatomine’s midgut with a blood meal , it comes into contact with the local microbiota . To survive and develop inside the insect’s gut , the parasite must evade the immune system and avoid detrimental interactions with the microbiota [3] . In the anterior midgut the parasite differentiates from the blood-borne trypomastigote to a spheromastigote , and then to the epimastigote replicative form . Elongated epimastigotes attach to the waxy cuticle of the hindgut wall , multiply by binary fission , and change into the infective metacyclic trypomastigote form , which is excreted with the feces from the rectum , ready to begin a new infective cycle [4] . Of the 152 formally described triatomine bug species , 67 are known to occur in Brazil [5]; the four species most frequently caught in and around houses in the country are Panstrongylus megistus , Triatoma brasiliensis , T . pseudomaculata , and T . sordida [6 , 7] . T . sordida is native to the Cerrado savannahs , although ecological niche modeling suggests that it may also occur in the semiarid Caatinga and the Pantanal floodplains [8 , 9] . To the south , it has also been recorded in the Chaco of Argentina , Bolivia , and Paraguay , suggesting that the taxon might in fact be a species complex [10–12] . Chagas disease control has for decades relied on the reduction of domestic and peridomestic vector populations through pyrethroid insecticide spraying [13] . In recent years , however , insecticide resistance has been detected in several triatomine populations [14] . This has brought renewed thrust to research aimed at the development of alternative control approaches . For example , both in vitro [15] and in vivo experiments [16–18] have shown that the gut microbiota of Rhodnius prolixus can modulate T . cruzi survival and development . A more direct attempt at disease control has been paratransgenesis–the use of transgenic gut bacteria that secrete T . cruzi-killing proteins [19] or express dsRNAs that reduce survival or hinder reproduction of the vectors [20] . The use of the insect microbiota to combat infection and transmission thus represents an interesting alternative to traditional control methods . Elucidating the role played by the gut microbiota in vector survival and T . cruzi infection and transmission may thus help devise novel disease-control strategies [21–23] . One key limitation of our current knowledge about the microbiota of triatomine bugs , however , is that little is known regarding field-collected specimens . Additionally , although enzymatic activities and nutrient absorption differ across digestive-tract segments , the gut of triatomines has hitherto been studied whole , as if it were a single homogeneous environment . Until recently , the investigation of bacterial diversity in insect guts rested upon the isolation and identification of cultivable bacteria–a method that inevitably misses many taxa . DNA sequencing , and in particular high-throughput technologies and metagenomics , now allow fast and accurate detection and determination of bacterial diversity ( including non-cultivable species ) virtually anywhere–for example , inside animal hosts [24] . In this work , we combined a metagenomics approach with bacterial-community analyses to investigate the gut microbiota of field-collected T . sordida . We asked whether and how the microbiota changes through bug development , and , for a subset of bugs , determined and compared the segment-specific microbiotas of the anterior midgut , the posterior midgut , and the hindgut .
Triatomine bugs were manually captured from chicken coops of six dwellings in a rural area of Posse ( 14°05’19”S; 46°21’18”W ) , state of Goiás , Brazil . Property owners provided oral informed consent to have their chicken coops surveyed for triatomines . The region is within the Cerrado biome and has a dry tropical climate with a dry season from May to September , a rainy season from December to March , and two shorter , transitional seasons . Fieldwork took place in three ( five-day ) trips in December 2013 ( rainy season; 201mm total rainfall , 27 . 8°C mean temperature ) , May 2014 ( dry season; 107mm , 25 . 2°C ) , and November 2014 ( transitional season; 170mm; 26 . 9°C ) . The bugs were transported alive to the laboratory , where they were morphologically identified based on Lent and Wygodzinsky’s keys [8] . We randomly selected five apparently fully blood-engorged bugs of each development stage ( 1st to 5th instar nymphs plus male and female adults ) for dissection . Each stage-specific pool ( pyrosequencing sample ) included bugs caught in the dry ( two bugs ) , rainy ( two bugs ) , and transitional ( one bug ) seasons . Prior to dissection , we sterilized each bug’s external cuticle by immersion in 70% ethanol ( 2 min ) followed by five rinses in phosphate-buffered saline ( PBS ) [25] . Bugs were individually dissected on sterile glass slides with sterilized forceps and disposable needles . After dissection , the guts of larger nymphs ( 4th and 5th stage ) and male and female adults ( five specimens each ) were cut into three segments corresponding to the major anatomic sections of the bugs’ digestive tract–the anterior midgut ( AM ) , the posterior midgut ( PM ) and the hindgut ( H ) . Due to their small size , the guts of 1st , 2nd , and 3rd stage nymphs were left whole . For comparisons of whole guts among development stages , the three segments of 4th-5th stage nymphs and adults were analyzed jointly ( i . e . , grouping the sequences of the three libraries together ) . Although this grouping might introduce taxonomic biases and thus brings limitations to the analyses , it allows for a more comprehensive view of bacterial-community changes along the entire development process . Comparisons involving whole guts from 1st , 2nd , and 3rd stages and each of the three intestinal segments from 4th-5th stage nymphs and adults are presented as supporting information ( S1 Table , S1 and S2 Appendices , all in S1 Text ) . Dissected material was isolated in 1 . 5 ml tubes , aseptically macerated in 300 μl of PBS solution with 50% glycerol , and stored at −80°C until DNA extraction . To identify the presence of T . cruzi in field-collected bugs , DNA was extracted from individual triatomines with the Qiamp blood mini kit ( Qiagen ) to PCR- amplify the kinetoplast DNA of T . cruzi as described by Cummings et al . [26] . The reaction mix was prepared using a Taq PCR Master Mix Kit ( Qiagen; as recommended by the manufacturer ) , 10 pmol of each primer ( TCZ-F* 5'-GCTCTTGCCCACAMGGGTGC-3' and TCZ-R 5'-CCAAGCAGCGGATAGTTCAGG-3' [26] , and 10ng of DNA in a final volume of 20 μl . Three μl of the PCR products were run in a 2% agarose-TBE gel stained with ethidium bromide ( 10 μg/ml ) ; samples yielding a 182-bp band were considered positive for T . cruzi DNA . DNA extracted from two T . cruzi strains ( CL Brener and Y ) was used as a positive control . The Y strain belongs to the major lineage circulating in the study area , T . cruzi II , whereas CL Brener is a T . cruzi I/T . cruzi II hybrid . The following assemblages were considered as individual samples for DNA library construction: the whole gut of 1st stage nymphs ( 1I ) ; the whole gut of 2nd stage nymphs ( 2I ) ; the whole gut of 3rd stage nymphs ( 3I ) ; the anterior midgut of 4th stage nymphs ( 4AM ) ; the posterior midgut of 4th stage nymphs ( 4PM ) ; the hindgut of 4th stage nymphs ( 4H ) ; the anterior midgut of 5th stage nymphs ( 5AM ) ; the posterior midgut of 5th stage nymphs ( 5PM ) ; the hindgut of 5th stage nymphs ( 5H ) ; the anterior midgut of adult females ( FAM ) ; the posterior midgut of adult females ( FPM ) ; the hindgut of adult females ( FH ) ; the anterior midgut of adult males ( MAM ) ; the posterior midgut of adult males ( MPM ) ; and the hindgut of adult males ( MH ) . Sample codes are composed of a first character that identifies the bugs’ development stage ( 4 , 4th stage nymphs; 5 , 5th stage nymphs; F , adult female; M , adult male ) , followed by letters that identify intestinal segments ( AM , anterior midgut; PM , posterior midgut; H , hindgut ) We extracted DNA with the DNeasy Blood & Tissue Kit ( Qiagen ) according to the manufacturer’s instructions . The hypervariable regions ( V3 to V5 ) of the bacterial 16S rRNA gene were amplified with primers 357F ( 5’-CCTACGGGAGGCAGCAG-3’ ) and 926R ( 5’-CCGTCAATTCMTTTRAGT-3’ ) containing 454 sequencing adapters and Multiplex Identifier ( MID ) tags [27] . PCR was performed with High Fidelity Platinum Taq DNA Polymerase ( Invitrogen ) , with initial denaturation at 95°C for 2 min and 30 cycles of denaturation at 95°C for 20 sec , annealing at 50°C for 30 sec , and extension at 72°C for 5 min . Each 16S rRNA amplicon library was constructed from five independent PCRs pooled in equimolar concentration . PCR products were purified with the Agencourt AMPure XP kit ( Beckman Coulter ) . Pyrosequencing was performed using a 454 Genome Sequencer Junior System ( Roche ) . We removed low-quality sequences shorter than 250 nucleotides or containing more than one ambiguous base , as well as sequences of the 16S rRNA primers and MID tags , using the trim . seqs script of Mothur v . 1 . 30 . 2 [28] . The remaining sequences were aligned against the SILVA alignment database ( http://www . mothur . org/w/images/9/98/Silva . bacteria . zip ) . We used Mothur pre . cluster scripts denoise sequences , and the screen . seq , filter . seq , and chimera . slayer scripts to screen for high-quality sequences . Then we used Mothur sub . sample scripts to ( i ) assemble three normalized subsets of sequences from 4th-5th nymphal stages and adults and ( ii ) merging their respective anterior midgut , posterior midgut and hindgut libraries . Operational taxonomic units ( OTUs ) were determined using the cluster script with the nearest-neighbor algorithm and a 3% distance level cutoff ( see [28 , 29] ) . We classified bacteria based on each sequence’s best match in the SILVA database . Sequences identified as DNA from mitochondria , Archaea , and Eukarya , as well as singletons , were removed from the bacterial community analysis . We used Good’s coverage index ( the number of OTUs sampled more than once divided by the total number of OTUs ) , as implemented in Mothur , to estimate sequencing depths [25] . Rarefaction curves were produced by plotting the number of unique sequence tags as a function of the number of randomly sampled tags with the vegan package in the R computing environment [30 , 31] . We computed OTU richness as the number of observed OTUs; however , to ensure that our richness estimate was reliable we used the bias-corrected Chao1 estimator . We also computed Shannon’s diversity index , which takes into account both the abundance and the evenness of species in a community [32] . These indices were calculated with the Mothur software [28] . In this paper , we define T . sordida’s ‘bacterial core’ as the set of bacterial OTUs that are present in all of the bug’s development stages–that is , the intersection of all development stage-specific OTU sets . Nonmetric multidimensional scaling ( NMDS ) represents the pairwise dissimilarity between samples in a low-dimensional space [33] . We used the ‘ordinate’ function of the R Phyloseq package [34] to simultaneously perform weighted UniFrac and a Principal Coordinates Analysis ( PCoA ) using differences in OTU relative abundances within each sample . We conducted exploratory analyses of similarities ( ANOSIM ) , with Bonferroni-adjusted p-values , to assess and compare the differences between the groups identified through PCoA [33] .
We collected 304 T . sordida specimens; our kDNA PCR did not detect T . cruzi DNA in any of the samples ( see S2 Table in S1 Text , for details ) . Pyrosequencing of 15 T . sordida samples ( i . e . , 15 pools of five whole guts or five gut segments ) generated a total of 98 , 872 good-quality sequences ( overall abundance ≥ 1%; mean±SE 6591 . 5±1072 . 1 sequences per sample ) . These sequences were taxonomically identified to the genus level based on a 97% sequence similarity cutoff . Sequences were clustered into 52 bacterial OTUs representing 49 genera in 38 families and four phyla . Rarefaction curves , supported by Good’s coverage index , showed that sampling depth was sufficient ( >0 . 93 mean±SE 0 . 959±0 . 003 ) to accurately characterize T . sordida’s bacterial communities ( S3 Table and S3 Appendix in S1 Text ) . This subsection addresses the question , “does T . sordida’s gut bacterial community change through bug development ? ” The Chao1 index of OTU richness increased from 25 . 0±2 . 30 SE OTUs in 1st stage nymphs to 50 . 0±2 . 10 SE OTUs in both 5th stage nymphs and adult males . Similarly , Shannon’s diversity index rose from 1 . 44±1 . 05 SE bits in 1st stage nymphs to 3 . 37±1 . 20 SE bits in adult males ( Table 1 ) . A representative sequence of each OTU present in each sample was used in a principal coordinate analysis ( PCoA ) based on weighted UniFrac distances as well as in ANOSIM . Principal Coordinate Analysis ( PCoA ) revealed a trend towards separation of 1st–3rd stage nymphs from older nymph stages and adults along Axis 2 ( Fig 1 ) . ANOSIM results confirmed OTU divergence between adults and the first three nymphal stages; they suggested , in addition , that the microbiota of 4th stage nymphs differed from that of adult males but not from that of adult females ( Table 2 ) . Taxonomic classification of T . sordida’s gut microbiota revealed the steady presence of four bacterial phyla ( Actinobacteria , Bacteroidetes , Firmicutes , and Proteobacteria ) in all of the bugs’ development stages . Actinobacteria was the predominant phylum ( Fig 2A ) , particularly in the first three nymph stages . The second most abundant bacterial phylum was Firmicutes , which was present in very low numbers in 1st stage nymphs ( 0 . 34% ) but increased sharply in abundance in 2nd stage nymphs and remained high until adulthood . There was an apparent increase of Firmicutes and Proteobacteria , at the expense of Actinobacteria , with bug development ( Fig 2A ) . We found no signs of genus-level bacterial dominance or abrupt change in relative abundance with bug development ( Fig 2B ) . Twelve bacterial genera were common to all development stages ( Brevibacterium , Corynebacterium , Dietzia , Gordonia , Nitriliruptor , Nocardia , Nocardiopsis , Rhodococcus , Pseudomonas , Sphingobium , Staphylococcus , and Williamsia ) , with no obvious differences in abundance . No single development stage had unique bacterium genera ( Fig 2B ) . The most abundant genera in the 1st and 2nd stages were Williamsia ( 22 . 37% and 17 . 69% , respectively ) and Rhodococcus ( 16 . 85% and 14 . 91% , respectively ) ( Fig 2B ) . Dietzia and Enterococcus were the most abundant genera in 3rd stage nymphs ( 12 . 98% and 12 . 16% , respectively ) . Clostridioides , Enterococcus , Kocuria , and Serratia were present from the 3rd stage onwards ( Fig 2B ) . Species of Bacillus , Streptococcus , and Lactobacillus appeared in 4th stage nymphs and remained until adulthood ( Fig 2B ) . This subsection addresses the question , “does T . sordida’s gut microbiota composition differ across the three intestinal segments for each development stage ? ” Estimates of bacterial OTU richness and diversity in each intestinal segment and are shown in Table 3 . Overall , we found no clear differences in bacterial OTU diversity across intestinal segments in the development stages we studied ( Table 3 ) . Bacterial communities present in the intestinal segments were also compared using PCoA based on pairwise weighted UniFrac distances and ANOSIM . PCoA-based comparisons revealed differences in the intestinal segment-specific gut microbiota of each development stage ( 4th and 5th stages nymphs and adults ) ( Fig 3; S4 and S5 Appendices and S5 Table , all in S1 Text ) . ANOSIM results suggested that ( i ) all three intestinal segments of 4th stage nymphs have distinct microbiotas; ( ii ) in 5th stage nymphs , the hindgut microbiota is different from that in the other two segments; and ( iii ) in adults , the posterior midgut microbiota may be slightly different from that in the other two segments ( Table 4 ) . We found a consistent increase in abundance of Firmicutes , at the expense of Actinobacteria , from anterior midgut to hindgut in all development stages ( Fig 4A ) . This was also the case when samples were clustered together by intestinal segment ( S6A Appendix in S1 Text ) . The relative abundance of some bacterial genera tended to change across intestinal segments ( Fig 4B ) . For example , Enterococcus was rare in the anterior midgut ( 4AM 0 . 71%; 5AM 1 . 69%; FAM 0 . 41%; and MAM 1 . 49% ) but became more common in the posterior midgut ( 4PM 6 . 85%; 5PM 4 . 31%; FPM 7 . 70%; and MPM 5 . 48% ) and , except for 4th stage nymphs , in the hindgut ( 4H 4 . 58%; 5H 12 . 87%; FH 10 . 36%; and MH 10 . 37% ) ( Fig 4B ) . Kocuria is another example–it was nearly absent from the anterior and posterior midguts of 4th ( 0 . 04% and 0 . 25% , respectively ) and 5th stage nymphs ( 0 . 07% and 0 . 39% , respectively ) , but rose to 4 . 38% and 2 . 13% in the hindguts ( Fig 4B ) . Kocuria was likewise nearly absent from the adult bugs’ anterior midgut ( females 0 . 01% , males 0 . 02% ) , but its abundance was much higher , particularly in females , in the posterior midgut ( 4 . 47% and 2 . 48% ) and the hindgut ( 4 . 40% and 2 . 08% ) ( Fig 4B ) . To further investigate these differences in the microbiota of the three intestinal segments , we pooled the sequences from each stage-specific sample by intestinal segment ( S4 and S5 Tables and S4 Appendix , all in S1 Text ) . PCoA analysis of these intestinal segment-specific sequence pools confirmed the distinctness of the three intestinal segments , and ANOSIM suggested that the hindgut was significantly different from the two other segments ( S5 Table and S4 Appendix in S1 Text ) . This subsection addresses the question , “does T . sordida’s gut microbiota composition in each intestinal segment change with development ( i . e . from one stage to the next ) ? ” Estimates of bacterial OTU richness and diversity in each intestinal segment through development are shown in Table 3 . PCoA and ANOSIM revealed some indication of within-segment homogeneity ( Fig 3 ) . The only differences suggested by ANOSIM were ( i ) between the anterior midgut microbiota of males and that of 5th stage nymphs and females; and ( ii ) between the hindguts of 4th stage nymphs and females ( Table 4 ) . We found no differences in bacterial phylum composition in the anterior midgut across development stages . The posterior midgut microbiota was similar in phylum composition both in 4th stage nymphs and females and in 5th stage nymphs and males . A decrease in the abundance of Actinobacteria was observed in the hindgut of 4th to 5th stage nymphs and adults ( Fig 4B ) . We identified only small differences in bacterial dominance or abundance when we compared intestinal segment-specific microbiotas through development . Thus , in the posterior midgut , Gordonia decreased in abundance from the 4th nymphal stage ( 18 . 68% ) to females ( 4 . 15% ) ; Nitriliruptor , on the other hand , increased in abundance from the 5th nymphal stage ( 1 . 74% ) to adults ( 4 . 07% for females and 8 . 27% for males ) . In the hindgut , Dietzia decreased in abundance from the 4th nymphal stage ( 12 . 4% ) to males ( 1 . 14% ) , and Williamsia also decreases in abundance from the 4th nymphal stage ( 11 . 18% ) to the 5th stage ( 1 . 79% ) and adults ( 2 . 7% ) ( Fig 4B ) . The bugs’ anterior midguts had a higher abundance of Williamsia , Rhodococcus , Lactobacillus , and Nocardia ( 16 . 41% , 11 . 81% , 10 . 78% , and 4 . 52% , respectively ) . Bacillus ( 17 . 10% ) was the most abundant genus in posterior midgut samples , followed by Williamsia and Gordonia ( 15 . 13% and 10 . 86% ) . Bacillus was also dominant in hindgut samples ( 13 . 65% ) , followed by Gordonia and Enterococcus ( 12 . 75% and 9 . 34% , respectively ) ( Fig 4B , S6B Appendix in S1 Text ) . Some bacterial genera , although present in all intestinal segments , seemed to favor one specific segment ( e . g . Enterococcus , Clostridioides and Dermacoccus are more abundant in the hindgut ) , whereas other genera were absent from certain segments ( e . g . Bacillus from the anterior midgut; Corynebacterium and Serratia from the posterior midgut , and Rhodococcus from hindgut samples ) ( Fig 4B , S6B Appendix in S1 Text ) .
We identified 52 bacterial OTUs in the 15 T . sordida samples we analyzed . This observed OTU richness is higher than reported for field-caught T . pseudomaculata ( 23 OTUs ) or T . brasiliensis ( 35 OTUs ) [46] . The numbers of OTUs found in the guts of these Triatoma species are , however , much smaller than those reported for other insects–e . g . , 300 OTUs in the Asian longhorn beetle Anoplophora glabripennis [51] or 417 OTUs in the tiger mosquito Aedes albopictus [52] . This suggests that the digestive tract microbiota of triatomines may encompass much fewer phylotypes than the gut microbiotas of other insects . T . sordida gut-associated OTUs belonged in four main phyla: Actinobacteria , Proteobacteria , Firmicutes , and Bacteroidetes . Species of 12 genera were present in all development stages and both sexes , and can therefore be collectively regarded as T . sordida’s gut ‘bacterial core set’: Brevibacterium , Corynebacterium , Dietzia , Gordonia , Nitriliruptor , Nocardia , Nocardiopsis , Rhodococcus , Williamsia ( Actinobacteria ) , Pseudomonas , Sphingobium ( Proteobacteria ) , and Staphylococcus ( Firmicutes ) . Given their persistence across all development stages , one should expect that at least some species of these 12 genera play important roles in T . sordida physiology , so that bugs carrying them in their guts have improved chances of survival and reproduction . Actinobacteria species of the genus Rhodococcus play important roles in the life cycle of triatomine bugs , providing their hosts with essential B-complex vitamins [37] . Dietzia , Gordonia , and Williamsia species , among others , produce secondary metabolites with antibacterial and antifungal properties and that contribute to gut microbiota homeostasis [53 , 54] . Proteobacteria can modulate the interaction between insect vectors and the pathogens they transmit . Thus , Pseudomonas putida decreases Plasmodium falciparum levels in Anopheles gambiae mosquitoes [23] , and Pseudomonas fluorescens causes T . cruzi lysis in vitro [55] . It is possible that some of the Pseudomonas species present in all T . sordida development stages and intestinal segments confer some degree of refractoriness to infection by T . cruzi . The number of phylotypes in the gut bacterial community of T . sordida increased through bug development , with clear differences in richness between the first nymph stages and later stages ( Table 1 ) . We however found no differences in Shannon’s diversity index values among the bugs’ development stages ( Table 1 ) . PCoA analyses revealed a tendency for separation of the first three nymph-stage sequences from those of the later stages ( Fig 1 ) , and ANOSIM suggested significant differences in gut bacterial composition between the first three stages and adults , and between 4th stage nymphs and males . These observations strongly suggest that the composition of T . sordida gut bacterial community changes through bug development . Gut bacterial communities also change through development in other insect groups , such as mosquitoes ( e . g . , [41] ) . Mosquitoes , however , undergo complete metamorphosis , with larvae developing in aquatic environments and only adult females feeding on vertebrate blood . In T . sordida , as in all Hemiptera , metamorphosis is incomplete and morphology , behavior , habitats , and habits including blood-feeding are all similar through development–yet the bugs’ gut microbiota still changes from early to late development stages . We note , in addition , that such changes were observed in bugs caught in chicken coops and , hence , most likely fed on chickens . These observations might indicate that environmental factors including bloodmeal sources have relatively little effect on the bugs’ gut bacterial communities , which may instead be to some degree directly linked to development–through , e . g . , maturation of the immune system or selection of particular bacterial species that increase host fitness . Bacteria of four phyla were present in all development stages: Actinobacteria ( the most abundant ) , Bacteroidetes , Firmicutes , and Proteobacteria ( Fig 2A ) . The abundance of Actinobacteria decreased , whereas Firmicutes and Proteobacteria species became more abundant , through the bugs’ development . These results are consistent with findings of Actinobacteria as the main phylum in field-caught T . brasiliensis , T . pseudomaculata , and R . pallescens; Proteobacteria species , however , predominated in peridomestic T . maculata [46 , 48] . Recent studies spanning a more comprehensive taxonomic sample of insects reported a clear dominance of Proteobacteria ( 60% ) and Firmicutes ( 20% ) among 218 and 62 insect species , respectively [56 , 57] . Proteobacteria species also seem to predominate in mosquitoes [52] . Proteobacteria are commonly found in the gut communities of a wide range of animals , including humans and insects , and are involved in vitamin biosynthesis [54] , protection against pathogens [17] , and degradation of plant compounds [58] . Proteobacteria species also increased in abundance with the reduction of Actinobacteria species , except for 4th stage nymphs . No single developmental stage presented unique genera . Although nearly absent from the guts of 1st stage nymphs , Firmicutes rose sharply from the 2nd stage onwards; overall , this was the second most abundant bacterial phylum in our samples . It is worth noting that some genera , such as Streptococcus and Bacillus , were absent from the first three stages of development , but appeared and became abundant from the 4th stage on; Lactobacillus species also appeared only on 4th stage nymphs and remained until adulthood ( Fig 2B ) . No overall differences were detected between males and females , but we found some evidence of intestinal segment-specific sex differences ( see below ) . Our results suggest that the digestive tract of T . sordida should not be regarded as a single homogeneous environment . Pairwise comparisons of the three intestinal segments of 4th and 5th stage nymphs and adults show at least two significant differences in bacterial community OTU richness per stage ( Table 4 ) . One such difference is between the anterior midgut of males and females ( note that no differentiation was detected when pooling the segment-specific data ) . This indicates that unless intestinal segments are analyzed individually , some differences might go undetected . We observed no differences in OTU diversity among intestinal segments . OTU relative abundance , however , varied significantly among intestinal segments . We found an increase of Firmicutes abundance at the expense of Actinobacteria along the T . sordida intestinal tract . This was also the case when samples were clustered together by intestinal segment . Certain genera appeared to be better represented in specific segments; for example , Enterococcus , Clostridioides and Dermacoccus were more abundant in the hindgut . On the other hand , some genera were absent from certain segments; for example , Bacillus was absent from the anterior midgut , Corynebacterium and Serratia were absent from the posterior midgut , and Rhodococcus was absent from hindgut samples ( Fig 4B , S6B Appendix in S1 Text ) . Enterococcus bacteria produce cytolysin , a lytic molecule with activity against diverse prokaryotic and eukaryotic cells such as Gram-positive bacteria , erythrocytes , leucocytes , and epithelial cells [59] . Serratia marcescens also secretes cytolysin [60] and is capable of inhibiting T . cruzi development inside the triatomine gut by attaching itself to the parasite surface [21] . The presence of Serratia sp . in the anterior midgut of T . sordida can contribute to the observed reduction of T . cruzi numbers in this compartment in the first days of infection [59 , 60] . Conversely , the absence of Serratia sp . in the posterior midgut may facilitate the replication and establishment of T . cruzi in the bugs’ gut , although the presence of Enterococcus may inhibit parasite differentiation in the hindgut . When each intestinal segment is compared across development stages , there is an indication of within-segment homogeneity in OTU richness . Only minor differences were observed between the anterior midgut of males and 5th stage nymphs and females , and hindguts of 4th stage nymphs and females . No single segment had a notably higher richness among the stages we analyzed . We found no differences in bacterial phylum composition in the anterior midgut across development stages ( Fig 4A ) . In the posterior midgut , phylum composition was very similar in 4th stage nymphs and females , as well as in 5th stage nymphs and males ( Fig 4A ) . The abundance of hindgut Actinobacteria decreased from 4th to 5th stage nymphs to adults ( Fig 4A ) . These differences in bacterial composition along the three major anatomical segments of T . sordida’s gut may be explained by the specific function of each segment during blood meal digestion . The anterior midgut of triatomines has a neutral-basic pH near 7 . 2 and functions as a reservoir of the ingested blood , which remains essentially undigested [61] . Only water elimination , erythrocyte lysis [59] , and inhibition of blood clotting by anticoagulants [62] take place . The anterior midgut harbors several symbiotic bacteria [37] , which ( especially actinomycetes ) may reach densities of up to 109 colony-forming units per insect after a blood meal [63] . Actinobacteria species are also predominant in T . sordida’s anterior midgut ( Fig 4A ) . This dense bacterial population may result from high nutrient contents; in addition , triatomines decrease reactive oxygen species ( ROS ) levels immediately after blood meal ingestion by reducing the production of mitochondrial superoxide [64] . In contrast , the proliferation of bacteria in the anterior midgut activates the bug’s immune response , as evidenced by the high antibacterial activity seen in this segment when compared to the posterior midgut in R . prolixus [65] . This immune activation , coupled with the presence of Serratia sp . ( mentioned above ) , can be important for the reduction of trypomastigote populations observed in this segment [62 , 63] . On the other hand , the presence of T . cruzi also decreases bacterial abundance in the anterior midgut of R . prolixus in the first days of infection [18] . The parasite can induce a Kazal-type protease inhibitor during the first hours of infection , which allows microbiota modulation and thus its successful maintenance in the host [66] . The posterior midgut is where complete blood digestion and nutrient absorption takes place , with participation of cathepsin L , carboxypeptidases , and aminopeptidases [67] . Symbiont population density is strongly reduced in this intestinal segment after a blood meal [63] . This suggests that proteases involved in blood digestion may also participate in microbiota control . In R . prolixus , digestion seems to have unequal lytic effects on different T . cruzi strains [68] . The hindgut receives and stocks blood remains until defecation . We observed a tendency of increasing abundance of Firmicutes species in the hindgut and in the posterior midgut compared with the anterior midgut in all development stages ( with the exception of adult males ) . Firmicutes bacteria also produce antimicrobial molecules such as polyketides and lipopeptides [69] . The balance between ROS production , immune activation , microbiota proliferation and bacterial profile changes along the digestive tract must be critical for proper establishment ( replication and differentiation ) of the T . cruzi parasite in the bug , with obvious consequences in terms of triatomine vectorial competence . Early studies on triatomine endosymbionts described R . rhodnii ( Actinobacteria ) as responsible for providing nutrients ( e . g . vitamins ) that enable the successful growth of R . prolixus [37 , 70] . Several genes for the biosynthesis of natural products have been identified in the genome of R . rhodnii such as polyketide and fatty acid synthases , nonribosomal peptide synthases , phytoene , carotenoid and vitamin B synthases [71] . The genome of the Actinobacteria Wigglesworthia , a tsetse fly obligate symbiont , has genes related to the biosynthesis of chorismic and folic acids and phenylalanine [72] , which may affect host physiology and vector competence to trypanosomes [73] . The Actinobacteria also produce a wide variety of secondary metabolites and antimicrobial compounds ( antibacterial and antifungal ) that may protect hosts against pathogens [53] . This might explain why Actinobacteria are dominant in the T . sordida’s anterior midgut .
We have described the gut microbiota of field-collected T . sordida through all the bugs’ development stages and across the three major intestinal segments . Species in 12 genera were consistently found in all development stages and can be regarded as T . sordida’s ‘bacterial core set’ . Some of these bacteria species , if proven cultivable , and non-pathogenic for humans or domestic animals , could be tested further for genetic tractability , stability after insertion , and fitness compared with wild type populations . They would hence become good candidates to be used in novel control strategies that make use of the vectors’ own microbiota to reduce pathogen transmission . For example , some bacteria can naturally control parasite loads through superactivation of the insect immune system , secretion of anti-pathogenic molecules , or by physically inhibiting their development inside the vector [23] . A second strategy is paratransgenesis , whereby specific bacteria are genetically transformed so that they secrete pathogen-killing molecules inside the vector [74 , 75] or synthesize double-stranded RNA molecules that interfere with the vectors’ development , survival , or reproduction [20] . The development of insecticide resistance has brought to our attention the immediate need we have to diversify our tools to control vectors and vector-borne diseases .
|
Triatomines are blood-sucking bugs that transmit Trypanosoma cruzi , the agent of Chagas disease . Insecticide spraying has been very successful at controlling house-infesting bugs , but some triatomines have developed insecticide resistance . One alternative disease-control strategy involves modifying the genomes of bacteria living inside the bugs’ gut so that they produce T . cruzi-killing substances . An obvious requirement of this strategy is in-depth knowledge of the natural bacterial community ( the ‘microbiota’ ) of the vectors’ gut . In this study , we evaluated bacterial diversity inside the guts of field-collected Triatoma sordida–a common pest in parts of Brazil , Argentina , Paraguay , and Bolivia . We found that a ‘core’ set of 12 bacterium genera occur in both immature ( five stages ) and adult bugs ( male and female ) , but also noticed changes in the gut’s microbiota through development . We finally investigated whether and to what degree the microbiota differed across the bugs’ three intestinal segments , and found clear variation . Our results will help pinpoint suitable candidates for genetic modification aimed at controlling T . cruzi inside T . sordida–that is , non-pathogenic bacteria that belong to the ‘core’ set and are easy to rear and maintain in the lab .
|
[
"Abstract",
"Introduction",
"Material",
"and",
"methods",
"Results",
"Discussion",
"Conclusions"
] |
[
"medicine",
"and",
"health",
"sciences",
"gut",
"bacteria",
"microbiome",
"body",
"fluids",
"microbiology",
"parasitic",
"protozoans",
"developmental",
"biology",
"nymphs",
"protozoans",
"bacteria",
"microbial",
"genomics",
"digestive",
"system",
"medical",
"microbiology",
"life",
"cycles",
"actinobacteria",
"gastrointestinal",
"tract",
"trypanosoma",
"cruzi",
"trypanosoma",
"eukaryota",
"blood",
"anatomy",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"organisms"
] |
2018
|
Field-collected Triatoma sordida from central Brazil display high microbiota diversity that varies with regard to developmental stage and intestinal segmentation
|
Post-kala-azar dermal leishmaniasis ( PKDL ) is a cutaneous complication appearing after treatment of visceral leishmaniasis , and PKDL patients are considered infectious to sand flies and may therefore play a role in the transmission of VL . We estimated the risk and risk factors of PKDL in patients with past VL treatment in south-eastern Nepal . Between February and May 2010 we traced all patients who had received VL treatment during 2000–2009 in five high-endemic districts and screened them for PKDL-like skin lesions . Suspected cases were referred to a tertiary care hospital for confirmation by parasitology ( slit skin smear ( SSS ) ) and/or histopathology . We calculated the risk of PKDL using Kaplan-Meier survival curves and exact logistic regression for risk factors . Out of 680 past-treated VL patients , 37 ( 5 . 4% ) presented active skin lesions suspect of PKDL during the survey . Thirty-three of them underwent dermatological assessment , and 16 ( 2 . 4% ) were ascertained as probable ( 2 ) or confirmed ( 14 ) PKDL . Survival analysis showed a 1 . 4% risk of PKDL within 2 years of VL treatment . All 16 had been previously treated with sodium stibogluconate ( SSG ) for their VL . In 5 , treatment had not been completed ( ≤21 injections ) . Skin lesions developed after a median time interval of 23 months [interquartile range ( IQR ) 16–40] . We found a higher PKDL rate ( 29 . 4% ) in those inadequately treated compared to those who received a full SSG course ( 2 . 0% ) . In the logistic regression model , unsupervised treatment [odds ratio ( OR ) = 8 . 58 , 95% CI 1 . 21–374 . 77] , and inadequate SSG treatment for VL in the past ( OR = 11 . 68 , 95% CI 2 . 71–45 . 47 ) were significantly associated with PKDL . The occurrence of PKDL after VL treatment in Nepal is low compared to neighboring countries . Supervised and adequate treatment of VL seems essential to reduce the risk of PKDL development and active surveillance for PKDL is needed .
Post-kala-azar dermal leishmaniasis ( PKDL ) is a late complication of visceral leishmaniasis ( VL ) , which usually appears several months after treatment of a VL episode . PKDL is seen in areas where L . donovani is endemic i . e . in Asia ( India , Nepal and Bangladesh ) and in east Africa ( Ethiopia , Kenya and Sudan [1] . In the Indian subcontinent , L . donovani is transmitted by the bite of a female sand fly of the Phlebotomus argentipes species , and the transmission cycle is considered to be anthroponotic with humans as the only known infection reservoir [2] . In Nepal , the standard treatment for VL with SSG was 20 mg/kg/day for 30 days without any upper limit recommended by WHO and drug was provided by the program to all government hospital in the endemic area . Due to associated toxicity and emerging drug resistance , SSG has been replaced in 2007 by Miltefosine 50 mg BID . PKDL is characterized by a spectrum of skin lesions ranging from hypo-pigmented macules , papules to nodules or combinations over the trunk and face that can be easily confused with other skin conditions such as vitiligo or leprosy [1] , [3] , [4] . So far , no convincing clinical predictors for PKDL have been identified [1] and its origin is believed to be multi-factorial and complex [5] . In Sudan , PKDL is more commonly reported in inadequately or irregularly treated VL cases [6] . PKDL is also sporadically reported in individuals without past history of VL [1] , [7] . The incidence of PKDL varies from country to country for reasons that are not entirely clear . In Sudan , PKDL was described in 50–60% of cured VL patients within weeks to a few months after treatment [1] . In Bangladesh , a cross-sectional survey carried out in 2009 of patients who suffered from VL in 2002–2007 found 10% of them with active or past PKDL usually occurring within 36 months after VL treatment [8] , [9] . In India , PKDL is reported in 5–10% of patients treated for VL usually after an interval of 2 to 4 years [3] and in 15–20% of PKDL cases there is no previous history of VL [3] . In Nepal VL is endemic in the south eastern Terai plains bordering the highly endemic districts of Bihar state of India , but systematic epidemiological data on PKDL are still lacking . PKDL patients have probably epidemiological importance in VL transmission as the lesions can harbour a large amount of Leishmania parasites , and as such could constitute a reservoir in the community capable of triggering a new epidemic [9] . As PKDL causes little or no clinical discomfort , and PKDL treatment with intramuscular SSG injections is long ( 3–4 months ) , painful and cumbersome , few patients seek treatment [1] , [5] , [10]–[13] . Since 2005 , the government of Nepal is involved in a collaborative effort with India and Bangladesh to reduce VL incidence to less than 1 per 10 000 population by 2015 [14] . PKDL is not addressed so far in this elimination initiative , which poses a threat to its success [15] . Better information on the epidemiology and burden of PKDL might help the national policy makers and health authorities to develop regional or national guidelines for its surveillance , control and treatment . We therefore conducted a retrospective cohort study and studied the probability and risk factors for PKDL development in past treated VL cases in five districts of south-eastern Nepal .
The study was conducted from February to May 2010 , in the districts of Jhapa , Morang , Sunsari , Saptari , and Siraha , known to be highly endemic for VL , with incidence rates of 2 . 0–4 . 0 per 10 , 000 person-years ( pyr ) in 2006/2007 . All VL cases that were notified by these 5 districts during the period 2000–2009 were taken as the study population . Information to trace the past treated VL cases ( pVL ) at household level was obtained through the district public health office ( DPHO ) of the districts and the VL patient database ( 2000–2009 ) of B . P . Koirala Institute of Health Sciences ( BPKIHS ) . BPKIHS is a university hospital located in Sunsari district that serves as the referral hospital for the region . The VL treatment centre maintains a patient register with clinical and epidemiological information . All pVL patients were approached at their residence by our field workers along with the vector control officer working at the DPHO . Written informed consent was obtained from pVL patients before including them into the study . The study was set up as a retrospective cohort study . Power calculations were as follows . We defined a priori 2 groups of patients according to their VL treatment experience: ( i ) VL patients treated in the government/private health facilities where most of the cases were treated on an ambulatory basis ( unsupervised ) by local medical assistants , and ( ii ) VL patients treated at BPKIHS where VL cases mostly were hospitalized for the full duration of treatment ( supervised ) . To detect a risk ratio of 5 , a sample of 332 was required in each group ( treatment in government/private health facilities vs at BPKIHS ) with the two-sided confidence level of 5% , a power of 80% , and a 1% expected frequency of PKDL in unexposed ( i . e . supervised treatment ) . A dermatologist examined all hypo-pigmented skin lesions cases referred to BPKIHS clinically and took samples for slit skin examination ( SSE ) and punch biopsy for histopathology for L . donovani . Differential diagnosis including leprosy , skin tuberculosis and other fungal infections was done in parallel . Every suspect was also tested by the rK39 immunochromatographic test for VL ( see below ) . Based on this assessment , a probable PKDL case was defined as a person having a past history of VL and multiple hypo-pigmented skin lesions ( macules , papules , plaques or nodules ) with a positive rK39 test but negative for L . donovani in SSE or histopathological examination . A confirmed PKDL was defined as a patient with multiple hypo pigmented macules , papules , plaques or nodules , who was parasite positive in SSE or biopsy . The Institutional Ethical Review Board of the BPKIHS Dharan , Nepal , and the corresponding bodies at the University Antwerp ( UZA ) , Antwerp , Belgium reviewed and approved the study protocol . Informed consent forms were developed in the national language and informed consent was obtained from individuals and from parents for children and adolescents . Written approval also was obtained from the authority of District Public Health Offices ( DPHO ) when information about past VL was collected from respective districts .
From the records , a total of 742 past VL patients ( pVLs ) were identified clustered in 17 highly endemic villages . Field workers succeeded in tracing 680 ( 91 . 6% ) . All subjects consented to the study and were interviewed and screened for PKDL-like skin lesions at their residence [Table 1] . 560 ( 82 . 4% ) had been treated with SSG ( 20 mg/kg/d for 30 consecutive days ) , 66 ( 9 . 7% ) with Amphotericin B , and 54 ( 7 . 9% ) with Miltefosine . In the SSG group , 17 ( 3% ) had not completed the treatment ( <21 injections ) . Half of the VL cases ( 347 or 51 . 0% ) had been treated at BPKIHS hospital under supervision during the period of treatment and all had received complete treatment; 333 ( 49 . 0% ) were treated at government/private hospitals on ambulatory basis , including 17 who had been treated in India . In our study , 370 ( 54 . 4% ) were male , and the median age of the study population was 28; inter quartile range ( IQR , 15–40 ) year and the large majority ( 75 . 4% ) was aged ≥15 years . Among the 680 pVL cases screened 37 ( 5 . 4% ) individuals showed hypo pigmented skin lesions . . No pVL patient in the study reported any skin disorder that since had disappeared spontaneously . All the 37 individuals with current skin lesions were referred to BPKIHS , as well as 2 others with suspect skin lesions who presented themselves spontaneously to the surveyors: one without history of clinical VL and a second with VL prior to 2000 . Thirty-three of the 37 referred pVLs reached the BPKIHS hospital and 16 ( 2 . 4% of 680 ) were diagnosed as PKDL ( probable: 2 and confirmed: 14 ) [Figure 1] . The other 17 were diagnosed as Pityriasis versicolor ( 12 ) , other fungal infection ( 2 ) , and vitiligo ( 3 ) . PKDL was also confirmed in the person without VL history , but not in the person who spontaneously reported with VL before 2000 . All 37 suspects tested positive in the rK39 rapid diagnostic test . HIV status was not tested . The median age of the 16 PKDL patients was 23 . 5 years ( IQR , 16–40 ) , and majority ( 14/16 ) was aged ≥15 years . All PKDL cases had previously been treated for VL with SSG and no PKDL was found in the treatment with Amphotericin B and Miltefosine . Five had not completed treatment ( less than 21 injections instead of the required 30 ) . Most of the patients ( 15 ) had been treated at government/private hospitals on ambulatory basis . The overall prevalence of PKDL in SSG treatment was 2 . 9% , 0 . 3% in supervised and 4 . 5% in unsupervised treatment . The median duration between treatment and onset of skin lesions was 23 months ( IQR , 15–41 months ) . The majority of patients ( 9/16 ) reported that the skin lesions occurred within 24 months after VL treatment . Hypo-pigmented macules/plaques were the most common hypo pigmented lesions and were present first on the face ( 11/16 ) , and upper and lower limbs were also affected . In most ( 13/16 ) cases reported , skin lesions appeared gradually from face to lower extremities and was associated with itching on erythema when exposed in sun light . Only 4 patients with skin lesions had sought medical treatment mainly for cosmetic purposes . We didn't find any PKDL cases with hepato-and/or splenomegaly and other associated complications such as post-kala-azar conjunctivitis or uveitis . Overall , the risk to develop PKDL was 1 . 4% within two years after VL treatment , 2 . 5% within 4 years and 3 . 6% within 8 years ( see Table 1 ) . The risk of PKDL by treatment ( adequate SSG , inadequate SSG and other treatments is shown in Figure 2 . In the SSG treated group alone ( 560 patients ) the prevalence rate of PKDL was 2 . 9% . In the univariate analysis , PKDL was significantly associated with unsupervised treatment at government/private hospitals ( OR = 16 . 28; 95% CI 2 . 48–689 . 00 ) with inadequate SSG treatment in the past ( OR = 19 . 77; 95% CI 4 . 66–75 . 00 ) . Both findings remained independently significant in the multiple logistic regression model . In the univariate analysis , the risk of PKDL appeared higher in private hospitals ( OR = 13 . 4; 95% CI 0 . 7–802 . 7 ) , but this finding was not significant and was not supported when corrected for treatment . None of the other assessed risk factors ( age , sex , hospitalization during treatment with SSG ) were significantly associated with PKDL ( Table 2 ) .
There have been few reports and studies on PKDL from Nepal [7] , [16]–[18] , and the frequency and risk factors of PKDL have not been studied previously . When re-examining a group of patients who were treated for VL in the previous ten years , we have found 2 . 4% having PKDL , and 2 . 9% in the sub-group of those treated with SSG in particular . The risk estimate for PKDL after VL was 1 . 4% within 2 years , and 3 . 6% within 8 years based on Kaplan–Meier analysis . This risk is lower than that reported in other VL-endemic areas in the Indian subcontinent [3] , [8]–[9] , [17] . Still , risk estimates reported are hard to compare due to the unequal follow up times . The median time from VL treatment to PKDL onset was 23 months ( IQR , 15–41 months ) . PKDL was more common in those with incomplete VL treatment and in settings with little treatment supervision . A limitation of our study is that enrolment in the cohort was based on data obtained from one tertiary hospital and governmental health facilities . It is generally assumed that VL is underreported [19] , as patients may be seeking treatment in the private sector . However , in our region in Nepal only few VL patients attend private clinics for VL treatment as anti-VL drugs are provided free-of-charge in the public health structures and are not available in private pharmacies . No cases of former VL treatment through private practitioners or pharmacies have ever been reported to the staff in the VL treatment centre at BPKIHS ( personal communication ) . This could possibly help explaining why the frequency of PKDL in Nepal is lower than in the other reports from the Indian subcontinent . Secondly , we assessed the presence of PKDL in 2010 in a cohort of patients diagnosed with VL between 2000 up to 2009; time of follow-up is variable , which makes the analysis more complex . The national VL control program initiated miltefosine-based treatment protocols in 2007 , and the 54 pVL cases treated with miltefosine have been regrouped for the purpose of this analysis with those who received Amphotericin . Though none of the VL patients in this group ( i . e . Miltefosine or Amphotericin B ) developed PKDL , follow-up time for miltefosine was definitely shorter than for the other drugs and therefore no final conclusions should be drawn yet about this drug [8] . Cases of PKDL in patients treated with Miltefosine have already been reported from India [20] . In another PKDL study conducted in the Fulbaria sub-district of Mymensingh district in Bangladesh , a cross-sectional survey was used to detect all past or active VL cases and active PKDL cases in a time period ( 2002–2007 ) and calculated incidences [8] . In this study clinical signs of PKDL were found in 9 . 8% of the 813 identified pVLs , almost 2 times higher than the proportion of suspects found in our study ( 5 . 4% ) while the time period studied was shorter . The authors mention however that parasitological confirmation of PKDL was only done in 10 suspected PKDL cases and confirmation was only obtained in 4 of these , a confirmation rate that was similar to ours . In another study in Trishal subdistrict of Mymensingh in Bangladesh without restriction in time of onset of VL [9] 52 PKDL suspects were identified for 235 pVLs , of which 18 ( 7 . 6% ) were as probable cases and 9 ( 3 . 8% ) were ultimately confirmed . Neither of both studies estimated the risk for PKDL over time using survival analysis . The single cross-sectional dermatological assessment may have made us underestimate the true incidence of PKDL , though no self-healing has been described for this pathology . No pVL patient in the study reported any symptoms of PKDL that since had disappeared - spontaneously or after treatment . One case of PKDL without antecedents of VL treatment was identified during the survey and confirmed at BPKIHS . In the cross-sectional surveys in Bangladesh , PKDL without previous VL accounted for 10% of all PKDL cases [8]–[9] . In Nepal an earlier study at BPKIHS reviewing the 22 cases of PKDL that were diagnosed between 1998 and 2000 , only 1 case had no clinical VL history [7] . It is thus unlikely that a high number of PKDL cases without previous VL have been missed in our survey . Median time of onset and clinical features of PKDL patients were all consistent with the findings from India and Bangladesh [7] , [12] , [21] . All 16 PKDL cases had lesions in the face , but only four had sought treatment for PKDL . All four were female and 3 were unmarried . In our study , the risk analysis only included data on VL history and treatment , and did not look into clinical markers such as HIV- and nutritional status , or immunological markers such as cytokines [22]–[23] . HIV prevalence is low in Nepal and even more so in the rural population affected by VL . Inadequate treatment received by pVLs in the past represented the most significant risk factor for PKDL ( OR 11 . 68 , 95% CI 2 . 71–45 . 47 ) . This is in line with earlier findings from Sudan where inadequate dosage and duration , and irregular treatment were important predictors for PKDL [6] , [23] . Without supervision of treatment and adherence counselling , patients may indeed abandon treatment earlier than prescribed , as clinical improvement usually appears within the first week of VL treatment , and there is little incentive to continue the painful intramuscular injections with SSG . In Nepal , VL treatment is provided for free to overcome financial barriers , but in some cases , treatment interruption was reportedly due to stock shortages of SSG at the hospital level . Treatment compliance should therefore be correctly monitored by clinicians and programmes , and all patients should be counselled about the importance of treatment adherence . This is important not only to avoid development of PKDL but also to reduce risks of treatment failure and development of drug resistance . It should be clear that PKDL is a multi-factorial phenomenon of complex origin [5] whereby drug related factors are not the only reasons for PKDL development . Host and parasite factors need to be further elucidated [24] . In conclusion , the occurrence of PKDL after VL treatment is relatively rare in Nepal compared to the two neighbouring countries involved in the VL Elimination Initiative . SSG , ambulatory treatment at government health facilities and inadequate treatment for VL in the past were significantly associated with PKDL . Counselling and supervision of treatment adherence in VL seems therefore essential to reduce PKDL incidence in the future , even if SSG is no longer used in Nepal . Reporting of cases of PKDL should be an integral part of the surveillance and monitoring system . Early identification can be improved by counselling VL patients on the risks and the signs of PKDL during their treatment Ultimately , the burden of PKDL can only be efficiently tackled if a more effective , affordable and short treatment can be offered to the patients .
|
Post-kala-azar dermal leishmaniasis ( PKDL ) is a skin disorder seen in patients treated for Leishmania donovani visceral leishmaniasis ( VL ) , a neglected tropical disease that is fatal if left untreated . In the Indian subcontinent , PKDL is seen in 5–10% of all past VL cases and is also reported in some without history of VL . As persons with PKDL do not feel sick , the disease has only cosmetic significance for the individual and treatment is rarely sought . However , PKDL lesions harbour parasites and therefore could represent a source of transmission , through the bite of female sand flies . Our study shows that the occurrence of PKDL in patients with past treated VL is low in Nepal compared to neighboring countries . Treatment of the original VL episode with SSG ( sodium stibogluconate ) , inadequate treatment and treatment on ambulatory basis were significantly associated with PKDL . Though SSG has since been replaced by other drugs , counseling and supervision of adherence to the prescribed VL treatment is of vital importance to reduce risk of treatment failure and relapse as well as later development of PKDL . Policy makers should include surveillance and case management of PKDL in the VL elimination program .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"leishmaniasis",
"neglected",
"tropical",
"diseases"
] |
2011
|
Post-Kala-azar Dermal Leishmaniasis in Nepal: A Retrospective Cohort Study (2000–2010)
|
To control and prevent rabies in Latin America , mass dog vaccination campaigns ( MDVC ) are implemented mainly through fixed-location vaccination points: owners have to bring their dogs to the vaccination points where they receive the vaccination free of charge . Dog rabies is still endemic in some Latin-American countries and high overall dog vaccination coverage and even distribution of vaccinated dogs are desired attributes of MDVC to halt rabies virus transmission . In Arequipa , Peru , we conducted a door-to-door post-campaign survey on >6 , 000 houses to assess the placement of vaccination points on these two attributes . We found that the odds of participating in the campaign decreased by 16% for every 100 m from the owner’s house to the nearest vaccination point ( p = 0 . 041 ) after controlling for potential covariates . We found social determinants associated with participating in the MDVC: for each child under 5 in the household , the odds of participating in the MDVC decreased by 13% ( p = 0 . 032 ) , and for each decade less lived in the area , the odds of participating in the MDVC decreased by 8% ( p<0 . 001 ) , after controlling for distance and other covariates . We also found significant spatial clustering of unvaccinated dogs over 500 m from the vaccination points , which created pockets of unvaccinated dogs that may sustain rabies virus transmission . Understanding the barriers to dog owners’ participation in community-based dog-vaccination programs will be crucial to implementing effective zoonotic disease preventive activities . Spatial and social elements of urbanization play an important role in coverage of MDVC and should be considered during their planning and evaluation .
The city of Arequipa is in the midst of a sustained dog rabies outbreak . The introduction of rabies virus into the city has been ascribed to the unintentional transport of rabid dogs from the rabies-endemic state of Puno during human migration [1–3] , and the persistence of transmission is likely due to low coverage in the annual city-wide dog vaccination campaigns [3] . Following the detection of the outbreak in Arequipa city in 2015 , the Ministry of Health of Peru ( MOH ) initiated additional vaccination campaigns in the city with varying intensity [4] . These additional efforts have not quelled the epidemic: more than 160 rabid dogs have been detected as of 2019 . Epidemics of dog rabies are ongoing in major urban centers across Latin America and worldwide [1 , 5–8] . Since bites from rabies-infected dogs cause 99% of human rabies deaths [9] , the control and elimination of dog-mediated human rabies relies on a One Health strategy: mass dog rabies vaccination [8 , 10 , 11] . Dog vaccination has dramatically decreased the global burden of human rabies since 1955 [5 , 12 , 13]; in the Americas , national programs centered around mass dog vaccination have achieved enormous advances [8 , 14 , 15] , reducing the incidence of dog rabies by 98% since 1983 [14] . In most rabies-affected countries , government health entities ( e . g . MOH ) organize annual mass dog vaccination campaigns ( MDVC ) that are held in outdoor settings . These campaigns are usually free of charge and voluntary [8 , 16] and campaign promotion varies greatly in format , content and intensity [1 , 17 , 18] . There are three non-mutually exclusive strategies used in Peru to implement MDVC: fixed vaccination posts , mobile teams setting up a temporary mobile post or conducting ‘street vaccination’ , and door-to-door vaccination [19] . For the fixed-point strategy , the vaccinators wait for the dog owners to bring the dogs to a unique place . For the mobile team strategy , the vaccinators move from one location to another during the day , vaccinating dogs along their way and spending short periods ( i . e . under an hour ) in a location before moving on , waiting for the dog owners to bring the dogs to these moving locations . For the door-to-door strategy , vaccinators knock on doors asking to vaccinate dogs in the household . Locations of the fixed vaccination sites are typically determined by a combination of convenience and prominence of the location ( e . g . the entrance to a health post , a well-known park ) [11] . In Peru , routes for door-to-door and mobile team approaches may or may not be decided in advance , and teams may move during the course of the day looking for dogs along their routes . The fixed-point strategy has been extensively used in Latin America and Africa , even though it has frequently failed to attain coverage targets [1 , 8 , 20–24] . The main reasons for its extensive application are that fixed-point vaccination is easier to implement and less costly than other strategies [18] . In many cases fixed-point is combined with other strategies , particularly when initial activities are unsuccessful [19 , 23 , 25] . However , high dog owner participation in MDVC and other dog-centered health campaigns ( e . g . de-worming dogs to prevent human hydatid disease ) has proven difficult to achieve in many areas [24 , 25] . It is necessary to understand barriers to community-based control strategies targeting dog populations in dog rabies-affected countries where coverage does not reach the minimum 70% recommended by the World Health Organization to attain herd immunity [11] , much less the 80% recommended by the Pan-American Health Organization for the region [26] . In cities , the social and spatial aspects of urbanization can facilitate the emergence of dog rabies and complicate its control [27 , 28] . The city of Arequipa comprises communities spanning different stages of urbanization and different migration histories , from old established neighborhoods , to young neighborhoods , to recent invasions [29] . Within this gradient of development , young neighborhoods and recent invasions are often located on the periphery of the city ( peri-urban area ) and the older localities are nearer to the center ( urban area ) [29] . Compared to the urban area , peri-urban areas generally have lower socioeconomic status , fewer community resources , more security problems , and often more rugged and uneven terrain ( Fig 1 ) . In the city of Arequipa , the locations of rabid dogs have been associated with urban structures [27] , and dog owners from areas with different levels of urbanization have reported distinct correlates of vaccinating their dogs against rabies [1] . The changing urban landscape and social processes in rapidly-growing cities have been associated with uptake of health-related services [30–33] and may be related to the low dog vaccination coverage in Arequipa . The objectives of the present study were to quantitatively assess the association between distance to a vaccination point and dog owner participation in mass dog vaccination campaigns in an urban setting , and to evaluate the effect of such distance on overall vaccination coverage and spatial distribution of participation .
Ethical approval was obtained from Universidad Peruana Cayetano Heredia ( approval number: 65369 ) , Tulane University ( approval number: 14–606720 ) , and University of Pennsylvania ( approval number: 823736 ) . All human subjects in this study were adults . The study was conducted in Alto Selva Alegre ( ASA ) ( human population for 2015: 82 , 412; density: 11 , 902 people/km2 ) , one of the 14 districts of the city of Arequipa . Arequipa , Peru’s second largest city , is home to 969 , 000 people and is situated at ~2 , 300 meters above sea level . The first detection of a rabid dog in the city of Arequipa occurred in March 2015 in ASA . By June 2016 , when our data were collected , 43 rabid dogs had been detected in 8 districts , but it is assumed that number represents a small fraction of the total number of cases [27] . The 14 districts in Arequipa vary in human population size , house density , and socioeconomic status , among other variables related to canine rabies ecology . These districts are formed by contiguous neighborhoods that also vary in those characteristics , and this variation is usually associated to the level of urbanization of those neighborhoods . As new neighborhoods mature into established neighborhoods with wealthier residents , homes are improved with better quality construction material and permanent utility connections , and connectivity with the rest of the city increases with better sidewalks , roads , and transportation access . ASA spans the gradient of urban development , running from the center to the periphery , and the district continues to grow towards the outskirts of the city . In our study , participants represented either the urban or peri-urban residential areas of the city of Arequipa ( Fig 1 ) . We included 21 urban neighborhoods founded many decades ago , and 21 peri-urban neighborhoods that originated around 2000 or later . In ASA , the MOH conducted a mass dog vaccination campaign in June 2016 . The human-to-dog ratio used to estimate the dog population was 10:1 in 2015 and it was reduced to 6:1 for 2016 , the year we conducted our survey . A detailed description of the mass dog vaccination campaigns can be found elsewhere [1] . In collaboration with the MOH , we georeferenced every stationary and mobile vaccination team during the three weekends when the campaign was implemented in ASA . Due to volunteer assistance , some dogs were vaccinated in weekdays during the campaign; we did not georeference the location of those volunteering vaccinators , mainly because their schedule was haphazard and unpredictable . We started door-to-door surveys immediately after the vaccination campaign , visiting every household in the study area and consenting and surveying only one adult ( ≥18 ) per household . The door-to-door survey was designed based on the rabies literature and based on our qualitative studies of local communities , in which we found specific household-level barriers to vaccination [1] . We collected household variables ( e . g . number of household members; number of children under 5 years old ) , dog owner or interviewee variables ( e . g . gender; educational attainment ) , and dog variables ( e . g . vaccination status; age ) . All houses in the study localities were geocoded and the survey data were linked to the household coordinates . We estimated the Euclidean distance between households and the closest vaccination point ( fixed , mobile , or either ) . We estimated the total vaccination coverage in the study area and compared the coverage in urban vs . peri-urban localities with a chi-squared test . We estimated the human to dog ratio and bootstrapped it 10 , 000 times to estimate its confidence interval . To evaluate the baseline characteristics of households and dog owners by participation in the MDVC , we defined an ordinal outcome for houses with dogs: no participation ( no dog vaccinated in the house ) , partial participation ( some , but not all dogs in the house , vaccinated ) , and full participation in the MDVC ( all dogs in the house vaccinated ) . We used a chi-square test to compare categorical variables with 10 or more observations per group , Fisher's exact test for categorical variables with fewer than 10 observations in any subgroup , and Mann–Whitney U test for age of the dog owner or interviewee , which did not follow a normal distribution and was truncated at 18 years . We compared the individual characteristics of vaccinated and unvaccinated dogs with chi-square for categorical variables and with Mann–Whitney U test for dog's age . Our main objective was to assess the association between distance to the vaccination point and participation in the MDVC . For distance to the vaccination point , we used the Euclidean distance from the dog’s house to the closest vaccination point , either fixed or mobile . For participation in the MDVC we used the ordinal values described above: no participation , partial participation , and full participation . We compared proportional odds logistic regression ( POLR ) , non-proportional odds logistic regression , and multinomial regression . The ordinal models were superior to the multinomial regression , and the proportional odds assumption holds for most of the covariates . Given that the categories of participation are inherently ordinal and that providing a single point estimate per covariate is more interpretable , the POLR model was favored . Based on the recent literature and our local studies [1 , 34 , 35] , the following covariates were used for model building: having a dog leash at home , number of children under 5 years old at home , time living in the area , rabies status of the last place they lived in before living in the study area , number of dogs at the house , age and gender of the dog owner , and educational attainment . For rabies status we used three categories: ‘endemic’ for departments that have reported dog rabies transmission during the last decades , ‘epidemic’ for Arequipa , and ‘free of dog rabies’ for the rest of the departments . We considered transformed distance to capture non-linear effects and interactions between distance and having a leash . The fit of the alternative models to the data was compared with Akaike’s Information Criteria ( AIC ) . We also attempted to build a hierarchical model to take into account the spatial autocorrelation within locality . However , given that within each locality there was at most one vaccination point , the variable distance from the house to the vaccination point would be unidentifiable under such hierarchical model . The final POLR model fitted with the R package MASS [36] was: ln{Odds ( Y≤k ) Odds ( Y>k ) }=b0+b1 . distance+b2 . leash+b3 . children+b4 . recencyinthearea+b5 . rabiesstatusofpreviousresidence+ε where k takes the values 0 ( no participation ) , 1 ( partial participation ) , and 2 ( full participation ) . All statistical tests were 2-sided , and significance level was 0 . 05 . We tested the spatial pattern of vaccinated and unvaccinated dogs in relation to vaccination tents for clustering using the bivariate cross K-function . This function estimates spatial dependence between two types of points ( i . e . unvaccinated dogs and vaccination points ) by measuring the expected number of points of type i within a given distance to a point of type j divided by the overall density of the points of type i . We used the Kcross function in the R package spatstat [37] to estimate deviations between the K function estimated for our data and the theoretical K function corresponding to a completely random Poisson point process for vaccinated dogs to tents and unvaccinated dogs to tents . Deviations greater than the theoretical K function indicate that the mean point count is higher than expected under complete spatial randomness ( CSR ) and thus some degree of clustering is present between the two event types at the indicated distance . Similarly , deviations less than the theoretical K function indicate that the mean point count is lower than expected under CSR which therefore indicates that some degree of dispersion is present between the two event types at the indicated distance . In order to investigate the association between geolocation and the odds of canine vaccination we fitted Generalized Additive Models ( GAMs ) to our data using the R package MapGAM [38] . GAMs are an extension of linear regression models in which both parametric and non-parametric terms are used to estimate the outcome of interest . We used a two-dimensional locally weighted smooth ( LOESS ) of latitude and longitude for our non-parametric term . The LOESS smoother fits each data point by weighting it towards nearby points , where weighting is based on the distance to the point being fitted . The percentage of data points in the region that will be used to predict a particular point is referred to as the span . The optimal span size used for smoothing was determined by minimizing the AIC . We mapped the odds ratio for each point on pre-specified grids of each locality ( from polygon data ) and next tested the null hypothesis that the odds of each points’ vaccination status did not depend on geolocation using permutation tests . For each test , the paired latitude and longitude coordinates were randomly permuted but vaccination status was held fixed . 1000 permutations were run for each locality and contour lines encircle areas with significantly increased or decreased vaccination odds as indicated by point wise p-values computed from the permutation ranks . All models and figures were created with R [39] .
Based on our survey , the estimated vaccination coverage of the MOH MDVC was only 58 . 1% , and it was low in both urban and peri-urban localities ( 58 . 0% vs . 58 . 6% respectively , chi2 = 0 . 086 , p = 0 . 769 ) . Only 3 . 4% of dogs were ( reportedly ) vaccinated in private clinics , bringing our estimated total coverage to 61 . 5% . Participation in our survey was higher in the urban area ( 88 . 8% ) compared to the peri-urban area ( 61 . 6% ) ( mean = 82 . 0% , chi2 = 6458 . 5 , p<0 . 001 ) . The total number of dogs in the surveyed houses was 5 , 292 and the human to dog ratio was 3 . 78:1 ( 95% CI: 3 . 69:1–3 . 89:1 ) , much lower than the 6:1 ratio used by the MOH for planning and evaluating the MDVC . In total , 65 . 3% of surveyed houses had dogs , but this number was higher in peri-urban areas ( 70 . 0% compared to 64 . 6% in urban areas , chi2 = 6 . 529 , p = 0 . 011 ) . For 76 . 9% of vaccinated dogs in the area , the person who took them to the MDVC is the person who responded to the survey . In our study area the urbanization process involves new localities being founded and settlers moving in . Accordingly , we found that the time of residency ( or the year people moved into this area ) was clustered at the locality level . However , we found that people living in or founding a locality do not necessarily share the same place of origin or previous residence . When we compared the distance from each house to the closest fixed vaccination point , to the closest mobile vaccination point , and to either the closest fixed or mobile vaccination point , we found a clear gradient in distance from vaccination point from non-participant houses ( farther ) to houses that partially participated ( closer ) to houses that fully participated ( closest ) ( Table 1 ) . The proportion of households with children under 5 was higher in households that did not participate in the campaign ( 36% ) compared to houses that participated fully or partially in the campaign ( 31% and 29% respectively; p-value = 0 . 019 ) . The proportion of houses with a dog leash increases from those who did not participate , to those who participated partially , to those that fully participated in the MDVC . There were some differences in MDVC participation by migration history: people who have lived longer in ASA tend to report higher participation in the MDVC compared to those who have lived fewer years in ASA . Also , there is a slight difference in participation in the MDVC depending on rabies status of previous residence , with more people participating in the campaign , partially or fully , if they were migrants from a rabies endemic area ( Table 1 ) . Other variables , such as educational attainment , the proportion of female dog owners or interviewees , and the proportion of households in urban localities were all similar in those households that participated fully or partially in the campaign compared to those households that did not participate of the campaign ( Table 1 ) . Having multiple dogs is a prerequisite to be in the partial participation group; therefore , houses with partial participation had on average more dogs , but the number of dogs per house was very similar in houses that participated fully compared to those that did not participate in the campaign ( Table 1 ) . Those who did not participate in the MDVC reported more frequently not knowing about the campaign before it happened ( Table 1 ) , but many of them reported learning about the campaign the same day it occurred . In S1 Table , we report the media channels through which they learned about the campaign either before it occurred or the same day it occurred . Compared to vaccinated dogs , unvaccinated dogs were older , were more likely female , had more free access to the street , had owners with no leash for them at home , and were less likely to be walked . Multipurpose dogs ( dogs reported as guard and company dogs ) were more likely to be vaccinated ( Table 2 ) . Based on the stated source of the dog , those received as gifts are more likely to be vaccinated and dogs born at home or adopted/picked on the street are less likely to be vaccinated ( Table 2 ) . Being considered purebred or being spayed/neutered was not associated with dog vaccination status ( Table 2 ) . In the multivariable regression analysis , we found that distance to the vaccination site is strongly associated with participation of the MDVC . The odds of participating in the MDVC were 16% lower for someone who lived 100 meters farther from the vaccination point after adjusting for other covariates and this difference was statistically significant ( Table 3 ) . Those having a dog leash at home had 35% higher odds of participation in the MDVC , either fully or partially , compared to those who did not have a dog leash after adjusting for other covariates . The odds of participating in the MDVC , either fully or partially , were 13% lower for each additional child under 5 years at home , after adjusting for other covariates . Migration history was associated with participating in the MDVC; participation was lower in those who migrated more recently to the study area ( 8% lower odds of participation for each decade less lived in the area , after adjusting for other covariates ) . Another component of migration history , the previous residence region , was also associated with participating in the MDVC: those whose previous residence was a rabies-free region or was any district within Arequipa were 23% to 32% less likely to participate in the campaign compared to those whose previous residence was a rabies-endemic region , after adjusting for other covariates ( Table 3 ) . Demographic variables such as owner’s/interviewee’s age , gender and educational attainment , and other household-level variables , such as number of dogs at the house , were dropped during model selection because they neither improved the fit of the model nor were statistically associated with participating in the MDVC . The significant association between distance to the vaccination point and odds of participating in the MDVC has consequences for the distribution of unvaccinated dogs in the area . We observed spatial clustering of unvaccinated owned dogs as a function of the distance from the house to the vaccination point . These pockets of unvaccinated dogs closer to each other than expected by chance occur at 500 meters from the vaccination point or further ( Fig 2 ) . We also analyzed the spatial odds of participating in the MDVC , that is , the association between their specific geolocation and the vaccination point . For areas served by fixed-point vaccination , there was a clear smooth spatial effect with higher odds of participating for houses closer to the vaccination point and a decreasing gradient farther away from the vaccination point . The spatial effect of the fixed-point vaccination strategy creates two clearly defined zones: a large zone with statistically significantly high odds of participation in the MDVC and another large zone with statistically significantly low odds of participation in the MDVC ( Fig 3 ) . For areas served by mobile teams , there were more spots of significant low and high odds of participation in the MDVC and these spots were spread in the study area without a clear association between the spots and the locations where the vaccination teams stopped to wait for dogs or to vaccinate dogs ( Fig 3 ) .
In Arequipa Peru , the social and spatial aspects of urbanization facilitate the emergence of dog rabies and complicate its control . In 21 urban and 21 peri-urban localities in Arequipa , Peru , we found low vaccination coverage and coverage that was spatially uneven . We found a strong effect of a potential proximal determinant of participation in the MDVC: distance to the vaccination point . The unadjusted data show a clear negative gradient with higher levels of participation in the MDVC at shorter distances to the vaccination point . After accounting for other important individual- and household-level variables , distance to the vaccination point remains an important factor associated with participating in the campaign . The association between distance to and participation in the MDVC also impacts the spatial distribution of vaccination . We found areas with statistically significant lower odds of dogs being vaccinated , and the LOESS smoother maps correlated well with maps of vaccination coverage . Therefore , it seems that both fixed point and mobile team canine vaccination approaches produced spatially heterogeneous vaccination coverage . However , we found that vaccination coverage was more “patchy” in localities served by mobile vaccination teams . This combination of mobile and fixed points was used also in 2015 , but the same localities are not always served with the same approach ( e . g . a locality that was served with mobile teams in 2015 could be served with fixed point vaccination in 2016 ) . As others have reported [40] , there is potential that in some localities owners in 2016 expected that the MDVC would be brought to their doors and did not plan or intend to bring the dogs to the fixed points in their areas . Spatially heterogeneous vaccination coverage is undesirable for dog rabies control and elimination . Townsend et al . [41] found that such patchy coverage can "profoundly damage prospects of elimination […] by creating pockets where rabies could persist" , and modeled patchy coverage within 1 km2 cell grids . The low coverage ‘patches’ in our study had smaller areas than 1 km2 , thus the potential for a threat to elimination efforts may be different or non-existent . However , it is unknown if these ‘patches’ are large enough to sustain rabies transmission in these densely populated areas or if there are much larger ‘patches’ in other parts of the city out of our study areas . Many studies have explored logistical , informational , social and structural barriers for dog rabies vaccination experienced by owners in rabies-affected areas [1 , 17 , 18 , 23 , 24 , 40 , 42–52] . Two of the most common reasons identified are difficulty handling the dogs [1 , 23 , 24 , 48 , 49 , 51 , 52] and lack of time [1 , 24 , 48 , 49 , 51 , 52] . These two logistical barriers are correlated with a less studied element: the distance to the vaccination campaigns . Distance to health services has been fairly well studied in terms of availability of and access to health care and impact on health , especially for maternal health , treatment and prevention of chronic diseases and treatment adherence for infectious diseases [53–55] , but not as much for preventing infectious diseases . Some rural studies mention distance as a potential factor for low dog vaccination coverage [23 , 56] and two studies directly evaluated the association between distance and overall rural villages vaccination coverage [49] and the association between distance and attendance at the MDVC in Sub-Saharan Africa [18] . In the sub-Saharan Africa study , researchers found that distance in dispersed communities have an impact on MDVC attendance [18] . Given the hilly landscape with rare direct paths between houses and vaccination points , they estimated the shortest-path distance for their analysis to take into consideration the long and tortuous routes dog owners had to follow to visit the vaccination sites . In our study area , with non-dispersed urbanized highly-populated localities and high density of street intersections that increase walkability , distance is still an important proximal explanation for low participation in the vaccination campaigns . Our study area consisted of urban and peri-urban localities . Surprisingly , there were no clear differences in participation in the MDVC between these two groups . However , the distribution of other proximal rabies-related characteristics is different between them ( e . g . more free-roaming dogs in peri-urban areas , more neutered/spayed dogs in urban areas ) . There are other social determinants that provide more distal explanations for participation in the MDVC . In previous focus groups conducted by our team , young females reported that having a baby at home could prevent them from participating in the campaign [1] . Similarly , we found that households with children under 5 years old were less likely to vaccinate their dogs compared to houses without children , and each additional child under 5 reduced the odds of vaccinating the dogs in the house . This insight suggests an opportunity to increase participation in the MDCV by framing the decision to vaccinate as an action taken to protect children in the household from rabies . There was also a clear difference in participation between those who lived in a dog rabies-endemic area before living in Arequipa . A possible explanation for that difference is higher awareness among that group . In our focus groups , we found low awareness and low perception of severity among residents of Arequipa [1] . Interestingly , another component of migration history was also associated with participation in the MDVC: time living in the area . This phenomenon has been observed for the utilization of other health services in different settings and populations [53–55] . Migration , settlement and adaptation are processes that take time and are necessary for the uptake of health services [56–59] , and could be influencing the participation in the MDVC . Importantly , in the peri-urban localities there are more households with children under 5 , more recent migrants , and more people whose previous residence was in a rabies-affected region . Our study has a number of limitations . Other studies have focused attention on dog-level variables ( e . g . sex , age , function ) that might be associated to participation in MDVC [18] . We did not analyze these; rather we focus on owner and community characteristics that can be utilized by the health authorities ( who rarely have the opportunity to collect detailed house-by-house information ) to increase participation in MDVC . Vaccination status and access to the street were reported by the owner/interviewee and we did not request proof of vaccination ( e . g . vaccination certificate ) . Given the bad publicity in the local media about owned free roaming unvaccinated dogs and the authorities’ threats to fine ‘irresponsible’ dog owners , there is potential for social desirability bias to inflate our estimate of vaccination coverage and deflate that of the proportion of dogs that have access to the street . We did not ask the interviewees to show the vaccination certificate they receive at the MDVC since many of them do not save these certificates . We used Euclidean distances , which are only a proxy for the real distance traveled by individuals to a vaccination post , and we did not include terrain elevation or slope which might affect walkability and the effect of distance on participation in the MDVC . Additionally , it is important to note that the MDVC in Peru only targets owned dogs . Our objective was to evaluate barriers to participation in the MDVC , so we only estimated the vaccination coverage in the owned dog population . The MDVC ignores important subpopulations for which size estimates are lacking: community dogs , strays , and feral dogs . The minimum vaccination coverage established by WHO ( 70% ) first with empirical data [57 , 58] and afterwards validated with rigorous modeling [59] was directed to dogs at risk of rabies so the low coverage we found must be even lower when unowned dogs are considered . Dog rabies virus transmission continues in Arequipa , putting at risk millions of people in the city and the surrounding departments . Dog-focused public health strategies are not limited to rabies: deworming dogs to prevent human echinococcosis [60] , use of insecticide treatment [61] or vaccines [62] on dogs to prevent human Leishmaniasis or Chagas disease , are just a few examples . These programs , if they are to be successful , require high coverage and even spatial distribution in their implementation . The same approaches to reach the appropriate levels of community participation that might have worked in the 1980s are not working today . Understanding the barriers for dog owners’ participation in community-based programs will be crucial to implement effective zoonotic disease preventive activities . Distance to health services and the heterogeneous social composition of growing cities have to be examined when designing field programs to protect against zoonotic diseases .
|
In Peru and other dog rabies-affected countries , mass dog vaccination campaigns ( MDVC ) are implemented primarily through fixed-location vaccination points: owners have to bring their dogs to the vaccination points where they receive the vaccination . To stop rabies virus transmission , a high and even dog vaccination coverage is desired . In Arequipa , Peru , following a MDVC , we conducted a door-to-door survey of >6 , 000 houses to assess how the placement of vaccination points affected coverage of the campaign . When comparing dog owners with similar characteristics , we found that the odds of participating in the MDVC was reduced by 16% for every 100 m distance from the nearest vaccination point . Some social conditions were also associated with participating in the MDVC: for each child under 5 in the household , odds of participating in the MDVC decreased by 13% , and for each decade less lived in the area , the odds of participating in the MDVC decreased by 8% . Distance to the vaccination point and variation in social conditions across the city play important roles in achieving coverage of MDVC and should be considered during campaign planning and evaluation .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"animal",
"types",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"neighborhoods",
"immunology",
"tropical",
"diseases",
"geographical",
"locations",
"vertebrates",
"microbiology",
"social",
"sciences",
"pets",
"and",
"companion",
"animals",
"dogs",
"animals",
"mammals",
"peru",
"viruses",
"preventive",
"medicine",
"rabies",
"rna",
"viruses",
"neglected",
"tropical",
"diseases",
"vaccination",
"and",
"immunization",
"zoology",
"human",
"geography",
"rabies",
"virus",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"geography",
"zoonoses",
"south",
"america",
"medical",
"microbiology",
"microbial",
"pathogens",
"people",
"and",
"places",
"lyssavirus",
"eukaryota",
"urban",
"areas",
"viral",
"pathogens",
"earth",
"sciences",
"geographic",
"areas",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"amniotes",
"organisms"
] |
2019
|
Socio-spatial heterogeneity in participation in mass dog rabies vaccination campaigns, Arequipa, Peru
|
The small GTPase Rac is known to be an important regulator of cell polarization , cytoskeletal reorganization , and motility of mammalian cells . In recent microfluidic experiments , HeLa cells endowed with appropriate constructs were subjected to gradients of the small molecule rapamycin leading to synthetic membrane recruitment of a Rac activator and direct graded activation of membrane-associated Rac . Rac activation could thus be triggered independent of upstream signaling mechanisms otherwise responsible for transducing activating gradient signals . The response of the cells to such stimulation depended on exceeding a threshold of activated Rac . Here we develop a minimal reaction-diffusion model for the GTPase network alone and for GTPase-phosphoinositide crosstalk that is consistent with experimental observations for the polarization of the cells . The modeling suggests that mutual inhibition is a more likely mode of cell polarization than positive feedback of Rac onto its own activation . We use a new analytical tool , Local Perturbation Analysis , to approximate the partial differential equations by ordinary differential equations for local and global variables . This method helps to analyze the parameter space and behaviour of the proposed models . The models and experiments suggest that ( 1 ) spatially uniform stimulation serves to sensitize a cell to applied gradients . ( 2 ) Feedback between phosphoinositides and Rho GTPases sensitizes a cell . ( 3 ) Cell lengthening/flattening accompanying polarization can increase the sensitivity of a cell and stabilize an otherwise unstable polarization .
Many types of eukaryotic cells undergo directed motion in response to external spatial signals in a process known as chemotaxis . Before starting to move , a given cell polarizes according to directional cues in the environment , forming nascent “front” and “back” regions . At the front , actin cytoskeleton assembly powers protrusion , whereas at the back , actomyosin contracts and pulls up the rear . Orchestrating the localization of actin network regulators and myosin activators are signalling molecules such as Rho-GTPases and phosphoinositides ( PIs ) . The spatio-temporal distribution of such regulatory molecules is thus critical to the correct polarization , motility , and chemotactic response of such cells . Proteins of the family of Rho-GTPases ( Rac , Rho , Cdc42 ) and the lipid PIs ( PIP , , ) , evolutionarily conserved across a wide range of eukaryotic cells , are implicated in cell polarization . These have garnered substantial interest as they are among the first elements in the chemotactic pathway to respond to a stimulus . Zones rich in Rac , Cdc42 , are associated with actin branching and growth , and zones rich in Rho are associated with myosin induced contraction . In many cell types , these zones are complementary , defining a “front” and “back” of the cell . Depending on cell type , the internal graded distribution of the GTPases and PIs amplifies shallow external gradients ( of as little as 1–2% across the cell ) into robust internal gradients [1]–[4] . The question of how such polarized distributions self-organize has attracted attention in both experimental and theoretical studies . Motivating the theoretical development to be described in this paper , is a collection of microfluidic experiments outlined in [5] . In these experiments , mammalian ( HeLa ) cells were placed in narrow channels that constrain lateral movement and restricts them to a single dimension . The cells were modified so that diffusion-driven linear gradients [6] of a small molecule would induce translocation of the Rac activator Tiam1 to the plasma membrane; this resulted in graded Rac activation across the cell length independent of upstream effectors . Polarization and protrusion were observed in these experiments with variations depending on the slope and intercept of the applied stimulus and the strength of PI feedback . Such experiments provide ideal testing ground for model development , refinement and validation . Our approach is to first consider the simplest hypotheses , reject those that are not supported by experiment , and successively build up the proposed network . Here we report in detail how models were constructed in a step-wise process to complement and crosscheck against these experimental observations . As the experiments also probed the effect of PI feedback on polarization , we are able to show agreement between theory and experiment linking GTPase and PI dynamics . To our knowledge , this is one of the first examples of such a match between GTPase-PI model predictions and observations . Numerous models of GTPases and PIs have been proposed , but few have been developed in tandem with experiments . ( See [7] for a recent review of qualitative models . ) Models of the PI pathway are provided in [8] , [9] . A model of Cdc42 in yeast cells is given by [10] , [11] . Models of polarization via three interacting Rho GTPases include [12]–[14] . Some of these are based on a Turing mechanism [15] for spontaneous pattern formation . It was shown by Mori et al . [16] , [17] in a reduced model with a single GTPase that rapid polarization can be achieved by “wave-pinning” as in [13] , [14] . In this phenomenon , bistability drives the formation of a wave of activity that stalls due to substrate depletion . Models of this type are attractive since they can capture both sub-threshold ( bistable ) dynamics observed in [5] and noise sensitive ( Turing ) dynamics . Dawes et al . [18] connected the GTPase model [13] with a model of PI kinetics and explored the role of PI feedback . Marée et al . [19] have refined and studied this in depth in a 2D model of a motile cell . Numerous other models such as [8] , [20] , [21] consider fundamental aspects of polarization without identifying specific regulatory proteins . Some , such as [9] proposed a local excitation global inhibition ( LEGI ) model for the dynamics of PI3K , PTEN , and and found experimental agreement in amoeboid cells . The availability of microfluidic cell polarity data provides a new opportunity to reconsider a variety of hypotheses in light of real cell behaviour . The essentially one dimensional geometry of the apparatus and direct activation of Rac , independent of upstream components , makes these particular experiments [5] amenable to model comparison . With this data , it is possible to revisit models that were purely theoretical so far , test their validity , and revise their structure . As explained below , this data quickly pointed to flaws in network connectivity that had been assumed in previous theoretical models , motivating the stepwise reconstruction of this connectivity . Here we develop a sequence of polarity models , starting with the simplest Rac-based polarization mechanism , and proceeding to include other GTPases that are widely known to be implicated in cell polarization and motility . For the simplest Rac-based model , Figure 1a , we rely on our previous theoretical work on “wave pinning” ( WP ) [16] , [17] . This choice is motivated by observations [5] that cells display clearly distinct behaviours below versus above a threshold stimulus strength , a feature that the WP model recapitulates . Extending that work , we include interactions with phosphoinositides . In subsequent iterations , we incorporate the remaining GTPases Cdc42 and Rho . We focus on three particular experimental results: 1 ) the presence of a temporal bifurcation in motility response , 2 ) the apparent distinct functional effects of the input signal attributes ( mean vs . gradient of Rac activation ) , and 3 ) the loss of response in some cases upon removal of PI feedback . We also explore the previously neglected effect of cell geometry , specifically cell aspect ratio , on polarization behaviour .
Rho-GTPases are molecular switches that exist in both membrane-bound and cytosolic states . The membrane bound forms are activated by GEFs and inactivated by GAPs . Inactive GTPases are extracted from the membrane by GDIs and distribute in the cytosol ( Figure 2a ) . In [5] , endogenous Rac was activated by applying a gradient of rapamycin to HeLa cells that had two constructs . One of these was a fluorescently labelled Rac-GEF , and a second was a cell membrane anchor . Rapamycin acts to dimerize these two constructs and localize the GEF at the cell membrane where it can activate Rac . Our model will be formulated to take this Rac-GEF activation stimulus into account . As the membrane-cytosol exchange of small GTPases plays an important role in the dynamics of these proteins , we first review aspects of the models that account for this cycling . This development follows [14] , but emphasizes the effect of cell elongation that was not previously considered therein . We denote the concentration of a given GTPase by in its active membrane form and , in the inactive membrane bound and cytosolic forms respectively . We make the biologically reasonable assumptions that each Rho protein has a constant total amount , , over the timescale of the experiments and that membrane cycling dynamics are much faster than activation/inactivation dynamics [22] . The latter hypothesis is a convenient simplification , that is not critical for model dynamics . As in [14] , we write down a set of three balance equations for each GTPase , one PDE for each of the states defined above . ( See the Methods for details , and Table 1 for meanings and values of all parameters . ) Briefly , are membrane and cytosolic rates of diffusion of the GTPase , is GAP-mediated inactivation rate , is the membrane dissociation rate , and the membrane association rate . is a GEF-mediated activation rate that , we assume , depends on crosstalk . In each of the models we discuss , we provide the explicit assumption about the form of that captures the assumed crosstalk . In view of the small thickness of the cell , we neglect gradients in the depth direction and integrate in both depth ( ) and width ( ) directions to arrive at a 1D spatial model . Cell length is retained as a parameter as discussed in the Methods . Adopting a quasi steady state ( QSS ) assumption that cycling between membrane and cytosol is very fast , we arrive at a model where each GTPase is assumed to have two forms , active ( ) and composite inactive ( ) . The latter is a sum of the inactive forms and ( projected from a 3D cell volume into the 1D spatial domain of the model , described in more detail in Methods ) . The GEF mediated reaction rates and “effective rates of diffusion” are modulated by cell geometry/length in the equations so obtained: ( 1 ) whereThe parameter is a composite that weights the respective rates of diffusion of inactive GTPase forms by the average time spent on the membrane versus the cytosol . In [14] , it was assumed that the GEF activation reaction could access the entire composite inactive pool . In reality , this reaction can only access the membrane bound proportion . The incorporation of this feature into the model equations ( 8 ) will have a dramatic effect as discussed in ‘Hysteresis and the role of cell length’ . Derivation of these model equations is found in Methods . While Turing instability is often invoked to account for spontaneous polarization [12] , [20] , this mechanism is not well suited for describing polarization of HeLa cells [5] or fibroblasts [23] , which have a stable rest state and are polarizable by a sufficiently strong graded stimulus , but not by weak signals of small amplitude noise . In contrast , mechanisms based on Turing instabilities are sensitive to noise of arbitrarily small amplitude . We will refer to such cells having that property as ‘hypersensitive’ . As HeLa cells are not hypersensitive , we here investigate only models where a threshold must be breached for a symmetry breaking event to occur . Mathematically , this threshold type response results from bistability . In a spatial setting , models with bistable kinetics and diffusion can spawn waves of activity that initiate polarization . Typically , waves propagate into the domain from one or several initial foci . Halting the wave is essential to lead to a polarized domain , and this requires that the wave slows down and stops . This has been shown [16] to occur in conservative systems exhibiting a form of bistability , referred to as “wave pinning” . In this setting , a threshold based response initiates a wave and conservation leads to the depletion of an inactive substrate , stalling the wave and leaving regions of high and low activity separated by a narrow interface . This is the mechanism for polarization underlying the sequence of models discussed below . In addition to GTPases , PIs are known to play an integral role in symmetry breaking that was investigated experimentally in [5] . Here we describe the sequence of model explorations that led to the model adopted for the GTPase-PI signalling layers . We briefly describe the attributes of each model variant , but only the final version of the model is analyzed in full detail . Phosphoinositides are membrane lipids that play well-known regulatory roles for the actin cytoskeleton . Both and become highly expressed at the nascent front of a polarizing cell , and they interact with small GTPases and with actin-associated proteins . PIs are successively phosphorylated by kinases such as PI5K , PI3K and dephosphorylated by phosphatases such as PTEN ( bottom layer of panels in Figure 1 ) . The reaction-diffusion equations for the PIs are similar to those in [18] , [19] , and given in detail in the Methods . Proposed interactions between GTPases and PIs for all models are drawn from literature [18] , [24]–[28] . The functions represent rates of phosphorylation by PI5K , and PI3K and dephosphorylation by PTEN and are assumed to depend on crosstalk from GTPases , as shown schematically in Figure 1 . In testing the suitability of models described below , we studied properties both with and without feedback to/from the PIs . While there are many hypotheses for the crosstalk and interactions between GTPases and PIs , the actual network at play in any given cell type , subject to various stimulus types and conditions is generally unknown . We first considered a theoretical model proposed by Dawes et al . [18] ( not shown ) and its modification , studied in detail by Marée et al . [19] . This pre-existing model couples Cdc42-Rac-Rho GTPase dynamics with PI exchange and bidirectional feedback . Mutual inhibitory feedback between Cdc42 and Rho is assumed , as well as positive feedback from Cdc42 to Rac and from Rac to Rho . This was a reasonable first candidate for a model of HeLa cell polarization and motility . In both [18] , [19] , stimulus was assumed to flow via Cdc42 activation to other parts of the signaling pathways . However , the experiments reported by [5] shortcut the natural signal flow by directly activating a Rac GEF . Incorporating this simple change in these previous models led to a surprising prediction that cells so stimulated should polarize in the wrong direction ( opposite to the stimulus gradient ) . Thus , experimental data allowed us to reject this candidate pre-existing model . In hindsight , the reason for this is clear . In the original models , traversing the circuit from Rac to Rho to Cdc42 back to Rac encounters only a single negative feedback . Thus a mild stimulus-induced asymmetry in the Rac profile feeds back negatively onto itself . As this feedback is the source of amplification , it overpowers the original signal and leads to polarization in a direction opposing the initial stimulus . In view of the observation that HeLa cells polarize in the correct direction ( up the gradient of Rac activator ) , we discarded these previous full models and decided to reconstruct a new model from the ground up . We use evidence from [5] and the broader literature on polarization as a basis to support or discard each model . In the following sequence of models , we first asked whether a single GTPase coupled to PIs ( Model 1 ) could account for major features of the data . We find that a single GTPase module can account for threshold based polarization with and without PI feedback . If PI feedback is added , the polarization capability is enhanced . However it is known that Cdc42 and Rho also participate in polarization . We next discuss Model 2 where Rho is passively coupled to the single GTPase polarization model ( Model 1 ) . For reasons discussed previously , we reject Models 1 , 2 as incomplete . In Model 3 , a modification of an existing model in [13] , Rac and Rho are assumed to inhibit each other . This model has the desired polarization properties , but it omits Cdc42 , widely believed to be one of the master regulators of cell polarity/motility [29] , [30] . Model 4 , which is the subject of the remainder of the paper , maintains the structure of Model 3 while incorporating Cdc42 based on extensive background cell biology literature . To assemble a new model , we started with the most basic relevant single-GTPase model due to [16] , to which we added the appropriate feedback . We here identify the single GTPase with Rac , the target of chemotactic stimuli in the experiments of interest . The model for a single GTPase is a well studied cooperative feedback model whose mathematical workings ( “wave pinning” ) were described in [16] , [17] . Adopting the same assumptions , we take Eqs . ( 1 ) with representing Rac , and Rac feedback onto its own GEF-induced activation as ( 2 ) ( are constants representing basal activation , feedback-induced activation , and level of Rac for a half-saturated feedback activation via GEF . ) In this model , a slow active form and fast inactive form interconvert . The active form feeds back onto its own production through cooperative binding . Inactivation is a first-order process . As discussed in [16] , this system exhibits threshold behaviour , i . e . is consistent with a polarizable ( rather than hypersensitive ) cell in the appropriate parameter regime . We connected the basic GTPase model to the model for PIs as in Figure 1a . We used Eqn . ( 16 ) with feedback terms ( 3 ) and take . ( Here are phosphorylation rates , and is total level of Rac in the cell . ) We also assumed that PIs affect Rac dynamics by modifying ( 2 ) to ( 4 ) where is some constant reference level of PIP and represents the strength of feedback to Rac activation . With parameters for the GTPase equations taken from [16] ( , , , , ) and PI-related parameters in Table 1 , this model exhibits wave pinning based polarization for a range of feedback values as required based on [5] . However , it is widely recognized that Cdc42 and Rho also participate in cell polarization , so we also consider a variety of possible connectivities that include these components along with Rac . To avoid introducing too many features at once , we first consider a situation where Rac is a primary regulator that directs Cdc42 and Rho . Figure 1b illustrates Model 2 , given by ( 1 ) , ( 2 ) , ( 6 ) , ( 7 ) , ( 16 ) , where Rac directs Rho , both of which affect the PIs . To understand how this model behaves , first consider what happens in the absence of PIs . In that case , the Rac module is identical to Model 1 and the Rho module is “enslaved” to it . Rac polarizes and Rho sets up a complementary profile due to the negative feedback link . Now including PIs merely introduces a secondary positive feedback . An important flaw in this model is that in the absence of PI feedback , Rho cannot influence Rac . While they are not specifically probed in the experiments that motivate these investigations , Rho and Cdc42 are observed to be more than passive regulators enslaved to Rac [29]–[31] . In this model , PI feedback between Rac , , and Rho does form a complete circuit where Rho can influence Rac through . However it is observed in [5] that inhibition of PI3K , which reduces levels , does not destroy polarization . This suggests that a secondary feedback mediated by PIs is not the primary circuit linking the three GTPases . Thus , we do not consider Model 2 or any similar models where Rac unilaterally polarizes and directs the remaining GTPases as realistic . Additional experiments where Cdc42/Rac are experimentally inhibited or knocked out would provide a test of the hypothesis that they are members of a complete GTPase circuit as opposed to passive regulators driven by Rac . In the following iterations we consider minimal models that contain complete circuits . We adapted Model 2 by revising the number and types of feedback arrows to incorporate mutual Rac-Rho inhibition , as shown in Figure 1c . Without the PIs , this model recapitulates a first case studied in [13] . Here the model equations are ( 1 ) with and ( 5 ) ( 6 ) ( The constants are typical values of Rho , Rac , associated with a significant feedback on activation . ) The second term in represents PI feedback to Rac . In this case , we use Eqs . ( 16 ) , ( 3 ) and define ( 7 ) to describe PI kinetics outlined in Figure 1c . While this model appears to be schematically similar to Model 2 , it incorporates an important structural difference . The bistability necessary for wave pinning to occur results from mutual inhibition ( two negative feedbacks ) as opposed to cooperative positive feedback . This is a natural next step in light of a result long posited by Thomas and recently proved [32] that bistability can result from networks with an even number of negative feedbacks while an odd number tends to yield limit cycles and other non-equilibrium dynamics . Reviewing Models 1–3 , note that Model 1 had no negative feedbacks . The GTPase portion of the Model 2 effectively consisted of feedback loops as well . Since Rho was slaved to Rac the inhibitory link does not act as a feedback , and the circuit involving is a positive feedback loop . In model 3 , the presence of negative feedback loops led to the required bistable behaviour as discussed in [13] . Model 3 exhibits the following minimal required features to account for basic experimental observations on HeLa cells . ( 1 ) It has regimes with bistable kinetics needed for polarizability ( as well as additional regimes of hypersensitivity in the Turing-instability sense ) . ( 2 ) It exhibits complementary localization of Rac and Rho , known to be related to protrusion and retraction respectively . This allows us to account for both “frontness” and “backness” cell attributes . ( 3 ) These behaviours occur both in presence and absence of PI feedback with all other system parameters held fixed , but can be “tuned” by the magnitude of that feedback . In principle , Model 3 would comprise the minimal required model . For completeness , we added Cdc42 , as shown in Figure 1d , given its importance as a master regulator [29] . However , results of Model 4 ( described further on ) also hold for Model 3 . We introduce Cdc42 with four criteria in mind . First , we sought interactions that lead to co-localization of Cdc42 and Rac that are complementary to the Rho profile . Second , we preserved the essential construction of two inhibitory connections of Model 3 to retain its bistable character . Third , we added a minimal number of overall GTPase interactions consistent with biological literature . Fourth , none of the GTPases is enslaved to the others . The model depicted in Figure 1d is the minimal possible model that satisfies these criteria . Removal of any connection or reversal of any feedback from positive to negative ( or vice versa ) destroys one or another of the required features , or requires additional compensating loops to avoid doing so . ( Although a reversal of all three GTPase connections restores the required behaviour , it is contrary to biological literature showing positive feedback from Cdc42 to Rac . ) We coupled the GTPase equations to the PI equations ( 16 ) with Eqs . ( 3 ) , ( 7 ) . The resulting model is described by ( 1 ) with and ( 8 ) The parameter in ( 8 ) represents the strength of the feedback from to Rac as shown in Figure 1d . The Rac-GEF parameters , , , along with signal will be the target of further analysis with all other parameters left fixed . A more complete discussion of the forms of the GEF kinetic terms is given in [13] but it is important to note that is required for bistability . Unless otherwise stated , this is the model we refer to from here on . The GTPase part of this model consisting of Eqs . ( 1 ) , ( 8 ) exhibits the bistability necessary for wave pinning to occur . To see this , consider the case of no PI feedback ( ) and no signal ( ) with , . Set at its resting steady state value and define . Now solve Eqs . ( 1 ) , ( 8 ) for with fixed as a parameter . Then it is straightforward to show thatand ( 9 ) Define as the Rho-dependent expressions on the left and right hand sides of Eqn . ( 9 ) , respectively . Then by plotting both together ( with parameters in Table 1 ) in the plane it can be shown that two stable steady states separated by an unstable repeller can exist for . Furthermore , for suitable parameters , this can be made true for a range of values of . Thus the necessary conditions for wave pinning [16] are satisfied . The complete model contains numerous parameter values ( Table 1 ) . Many of their values are based on previous literature . We summarize the default values of basic rates and diffusion coefficients below , and then explain the procedure used to find interesting ranges of behaviour of the model when other key parameters were varied . Consistent with [13] , [14] , [18] , we take , . With assumed values of the necessary parameters , can be computed using ( 10 ) completing the parameter set associated with membrane cycling . Given the undetermined nature of many of these parameters , we instead vary the composite parameter described in Methods , which represents the bulk effect of length variation in the model cell as it polarizes . GTPase crosstalk parameters are modifications of [14] to fit our system . PI parameters are a modification [18] by Marée et al . [19] . To gain insight into how parameter variations affect model behaviour , we utilized the ‘Local Perturbation Method’ described briefly in the Methods . This considers the stability of a homogeneous steady states against localized delta-function-like perturbations . The idea of the method is to replace the system of PDEs by approximating ordinary differential equations ( ODEs ) for local versus global variables ( according to slow versus fast-diffusing intermediates ) . Then we can use bifurcation diagrams to explore the transitions between different regimes of behaviour . The LPA method allows us to detect both ultrasensitive and polarizable behaviour , a distinction of particular interest here . With the above preparation , we now explore how specific aspects of the stimulus , the assumed feedback structure , and cell geometry affect the dynamics of the model 1D cell behaviour . Figure 3 maps out a typical parameter space structure for the discussed models . In the coming sections we discuss the relevance of each of these parameter regions and the bifurcations that occur between them . Consider a cell , initially at rest , characterized by a low homogeneous steady state ( HSS ) of GTPase activity in Region II of Figure 3 . Let the applied stimulus gradient be represented by . Recall that such gradients could be formed and maintained in experiments described in [5] . As in the experimental stimulus , we assume that this produces an internal Rac-GEF gradient . ( A similar analysis can be performed with a Cdc42-GEF signal . ) To polarize the cell , at least part of the cell domain must be elevated to Rac activity level above the threshold shown ( dotted ) in Region II of Figure 3 . When this happens , that part of the cell evolves to a high Rac activity level ( highest solid line , Region II ) , and , by virtue of diffusive coupling , creates a wave of activity that invades nearby portions of the cell . The wave stalls and leads to a polarized cell for parameter values in Region II ( Figure 4 , left ) . Both the signal strength ( ) and gradient ( ) contribute to the ultimate response , but each plays a slightly distinct role . serves to produce an internal asymmetry in the GTPase profile and augments the size of the gap that has to be breached to induce polarization . In ( 8 ) the parameter is directly added to , the bifurcation parameter in Figure 3 . Thus , increasing is equivalent to moving the state of the model cell to the right on that bifurcation diagram . This reduces the gap between the stable and unstable states and consequently the size of the perturbation required to induce polarization . Thus , effectively controls the sensitivity of the cell to heterogeneous stimuli . , in contrast , produces the actual asymmetry necessary for the system to polarize . Numerical simulations of the full PDE system confirm this prediction of the reduced system . This sensitivity relationship and the functionally distinct roles of and recapitulate the experimental observations in [5] . In the graded-stimulus experiments , a bifurcation occurred after some time . Stimulated cells had a long nascent period followed by an abrupt change to a much more active state . This suggests a temporal build up of Rac-GEF which sensitizes the cell . The resulting bifurcation would then lead to polarization . Other experiments and irreversibility of the stimulus-induced GEF activation [5] support this hypothesis . We asked next how the positive feedback from Cdc42 to the Rac-GEF pathway affects model cell dynamics . This feedback is controlled by the parameter . Figure 5 summarizes changes in the bifurcation structure of the reduced ( LPA ) model as this parameter is varied . We first decreased below 0 . 55 and noted that pattern forming capabilities of the system are completely lost . Next , we increased this parameter . As expected , an increase in the strength of this positive feedback serves to sensitize the cell , i . e . , increases the extent of the ultrasensitive Region III . For example , while for this region spans roughly ( bounded by intersections of thinnest monotonic curve with smallest ellipse in Figure 5 ) , when , Region III has expanded to . As is further increased , Region II is squeezed into the negative half plane , where it is no longer biologically feasible . Thus , Turing instability characterized by Region III takes over larger portions of the parameter plane . For ( for example when in Figure 5 ) , a new regime forms between the original Regions III and IV of Figure 3 . Here we find a new bistable region , with lower steady state ( shown in red ) , higher one ( blue ) and unstable repeller ( dotted red ) in the approximate range . The size of this range grows in size as is increased . Unlike Region II of Figure 4 where a pulse of activation is needed to polarize , this new bistable region requires a pulse of inactivation ( reducing the HSS below the dotted elliptical arc ) to obtain polarization . ( This prediction was verified with the full PDE system . ) Experimental manipulations in [5] addressed the effect of a PI3K inhibition on the cells' response to graded stimuli . We used the full ( 9 PDE ) model to address these observations . Having understood the behaviour of GTPase layer of signaling using the above analysis and simulations , we now turn to the full GTPase-PI feedback model . The parameter is used to tune the level of that feedback as shown in Figure 1 . Recall that PIs are membrane-bound lipids . Their rates of diffusion are neither as fast as cytosolic GTPases , nor as slow as the membrane-bound GTPase forms . To gain some intuition using the LPA method , we therefore conducted two separate tests . We first treated the PI variables as fast ( global ) variables . The left panel of Figure 6 shows the effect of increasing the PI feedback parameter in this case . As seen , this produces a direct linear shift of the entire bifurcation plot to the left . This can be explained by the fact that in the infinitely fast diffusion limit for PIs , the feedback term is spatially homogeneous , and therefore simply increments the bifurcation parameter . This can be interpreted as sensitizing the cell: for a given set of parameters , as increases , the critical asymmetry required to produce polarization is reduced . We next investigated the approximation that PIs are slow ( local ) variables , as shown in the right panel of Figure 6 . While features of the two panels ( global vs local ) are not identical , qualitative aspects and , surprisingly , the dominant feature of leftward linear shift is preserved . This model prediction suggests that the primary role of PIs is to act as a global mechanism for increasing sensitivity . To check this prediction , we carried out simulations of the full 9 PDEs under systematic variation of the two parameters and . Results , shown in Figure 7 reveal a linear boundary separating bistable behaviour ( “Region II” , shaded grey ) from ultrasensitive behaviour ( “Region III” , white ) . The linearity of this two-parameter bifurcation plot is consistent with the observed linear shift in Figures 6 . Further , the total rate of shift in Figures 6 with respect to and the slope of the bifurcation line in Figure 7 are close to , the parameter that controls the relative strength of this feedback . The combination of these three facts strongly suggests that the primary role of PI feedback is to provide global sensitivity . This feature is consistent with recent experiments in [5] , and provides one of the strongest predictions of the model . Experimental observations in [5] reveal that as a cell polarizes and elongates in the confined channels , its overall height changes inversely to its length . This feature was introduced into our models through volume conservation . Recall that the composite inactive form was introduced under a QSS assumption as a weighted sum of membrane bound and cytosolic inactive forms . These weights are explicitly linked to the geometry of the cell ( details in the Methods ) and can be explored consequently . As the model cell lengthens and flattens , the surface area to volume ratio increases . Given the form of in our equations , this leads to a larger proportion of the inactive GTPase in the membrane bound form , resulting in two changes: ( i ) the composite form diffuses more slowly , and ( ii ) the GEF activation reaction can access a greater portion of inactive GTPase . However , whereas ( i ) has little effect , due to the relative insensitivity of the bistable and Turing unstable regimes to diffusion in the PDE system , ( ii ) has a substantial effect . As shown in Figure 8 , increasing cell length tends to sensitize the model cell . This effect is similar to the effect of increasing either Cdc42 or PI feedback to the Rac-GEF pathway ( , or ) . Meyers et al . [33] similarly considered the role of cell depth/length in polarization with a similar result that larger surface area to volume ratios lead to larger proportions of GTPases being in the phosphorylated active form . However they considered GEF's to be membrane bound and GAP's to be cytosolic where we consider both to be membrane bound . In either case , the end result is the sensitization of a cell as it flattens . An additional feature seen here is the reduction and subsequent elimination of hysteresis as is increased . This hysteresis is present when the loop of steady states is entirely contained in the right half plane and a stable region for low values of is present as with . We refer to this as hysteresis for the following reason . Consider , first , the following experiment with a resting cell of some fixed length . If , then the cell state is in the stable region ( e . g . , point on Figure 8 ) , where no heterogeneous signal can lead to polarization . Apply a signal of the form where is an increasing function of time . This would be the case for a signal that cumulatively builds up over time . The buildup will cause the model to become increasingly sensitive to the applied asymmetry until it becomes sufficiently sensitive to respond/polarize . Graphically , the cell moves to point and subsequently of Figure 8 upon polarization . Once polarized , the asymmetric component of the signal can be turned off ( ) and the cell will stay polarized . As the background signal is washed out ( reduced and the state shifts leftwards on Figure 8 ) , polarization will be maintained until the cell falls off the ellipse at point and again takes on a stable HSS at point . The state trajectory would follow the path on Figure 8 . Interestingly , in addition to sensitizing the model , increasing length also removes hysteresis by pushing part of the loop into the left half plane and removing the stable region . A similar feature is seen in Figure 5 for ( and for other parameters we explored , not shown ) . However geometry/length is inherently a dynamic quantity whereas other parameters could be considered static on the time scales considered . So while genetic diversity in a host of parameters could play a role in the variability of behaviours among a population of cells or across cell lines , this length dependent removal of hysteresis can temporally stabilize an otherwise unstable polarization in a single cell . Now consider the same experiment but with a cell capable of length change . Begin with a stable cell in a resting state ( point ) with the same cumulative stimulus . Again , after some point the cell will polarize ( moving through point to point as before ) . As discussed previously , a static cell will lose polarization upon removal of this stimulus . In a dynamic cell however , such length change effectively shrinks or eliminates the stable region and associated hysteresis . When the cell lengthens , its state moves from to and , upon the removal of the stimulus , to . Thus , if the onset of polarization causes cell lengthening , the geometric effect described here affects internal signaling to stabilize the polarization , as indicated by the path on Figure 8 .
While the ultimate model we considered is a modification , extension , and rederivation of previously published models , it brings several new ideas and new results: first , all previous papers were theoretical , whereas here we were able to reassess details of the models in direct comparison with experimental data . Second , while previous models could account for polarization via Cdc42 stimuli , they produced incorrect predictions - thus invalidated - in view of that data , mandating a revision of the previously proposed GTPase connectivity . Third , using the novel LPA analysis , we have shown how the parametrization and analysis of behaviour could be accomplished with a novel analytic tool . Fourth , we provided here a hypothesis for how environmental factors can influence the response threshold through the GEF pathway . Fifth , and finally , we showed how the ratio of surface area to volume of the cell can influence the signalling . We found that Model 4 is capable of qualitatively capturing many aspects of symmetry breaking and polarization in HeLa cells observed in microfluidic gradient generation experiments . We have included only features necessary to describe such observations . To aid the process of model development , model analysis , and parametrization , a novel analytic approximation technique , Local Perturbation Analysis , was introduced and applied . This proved to be fruitful as the model helped interpret experimental results and provided non-trivial insights into the behaviour of the experimental system . The experiments were designed so as to allow convenient simplifications in modelling . The geometry of channels makes a 1D spatial representation both relevant and accurate . The tightly controlled gradient stimulus makes the assumption of ( linear ) signal shape appropriate . Finally , the stimulus bypasses a number of upstream signalling components and directly targets the Rac-GEF , making the input to the model clear and direct . Through these simplifications , we have produced a model that is both a reasonable representation of the system , and numerically and analytically tractable . This allowed for qualitative comparisons between model and experiment . Because of the unique form of stimulation ( via Rac , not Cdc42 activation ) , we could not directly use previously developed GTPase-PI model that had been tuned to stimulus inputs via Cdc42 GEFs . Rather than tinkering with that model we developed the new version from the ground up , proceeding from the simplest bistable GTPase module . A sequence of models involving one , two , or three GTPases with and without PI feedback were developed , allowing us to identify models with the minimal required capabilities . We showed that although the simplest model ( with a single GTPase coupled to PIs driving polarization through positive feedback ) does reproduce polarization ( via “wave-pinning” ) it is less suitable than models based on mutual inhibition since it does not incorporate the remaining GTPases , Cdc42 and Rho in the polarization process . In terms of complexity , the final variant ( Model 4 ) consisting of three GTPases is a minimal mutual-inhibition model that mimics the typically observed GTPase localization behaviour , and accounts for the observed response to PI feedback tuning . We investigated the roles of stimulus mean and gradient , feedback , as well as cell geometry using Model 4 . Both full simulations of the model PDEs as well as bifurcation analysis of the LPA reduction provided insights . We found that signal mean could affect overall cell sensitivity while signal gradient drives the asymmetries needed to overcome a threshold for polarization . Further temporal buildup of Rac-GEF that results from a prolonged exposure to stimulus can account for bifurcations observed experimentally . This leads to the idea that that cells become increasingly sensitive with sustained stimulus , and is consistent with experiments . As far as the role of feedback between PIs and GTPases , we found that removal or reduction of PI feedback reduces sensitivity of the model cell to applied stimulus gradients . This , along with matching experimental results , supports the idea of feedback between PIs and GTPases ( as opposed to PIs acting upstream of GTPases ) . Finally , we also found a role for changing cell geometry . When the cell lengthens , an increase in its surface area to volume ratio can remove hysteresis . This suggests that such purely geometric effects could stabilize otherwise unstable polarizations . Limitations of the model include the absence of the cytoskeletal network , and possible feedback to and from that layer . In [19] , we have shown that dynamic cell shape in 2D ( top-down view of the cell ) can feed back onto the internal biochemistry . Probing the multiple feedbacks and interactions in a similar 2D computational platform could provide new insights . In order to extend this work to other settings , it is important to similarly probe the Cdc42-GEF and/or Rho-GEF pathways ( both experimentally and with a similar model ) to more fully understand feedback to and from other GTPases . While the model was developed in the context of a specific cell type , many of its characteristics are observed in other cell lines . The model reductions and LPA approximation are also applicable to other settings . As more data regarding these types of signalling networks becomes available , these approaches will speed model development and aid in understanding the structure and dynamics of such networks .
Experiments were performed by methods described in [5] . Briefly , constructs were introduced into HeLa cells , ( a cytoplasmic YFP labeled TIAM1 , a Rac GEF , conjugated to FKBP ( YF-TIAM1 ) and Lyn11-FRB ( LDR ) that acts as membrane anchor ) to directly activate Rac independent of upstream effectors [34] . HeLa cells were introduced into microfluidic chambers and allowed to settle ( 3–4 h ) . Linear gradients of rapamycin were created and maintained by actuation of flow in the microfluidic system . ( The rapamycin dimerizes the constructs and leads to membrane-associated Rac activation . ) Cells were imaged and observed over several hours , and classified according to initial and final polarization states . The PI3K inhibitor LY294002 was used to determine the effect of reducing feedback from PIPs to the GTPases . Models were formulated to describe the dynamic behaviour of these cells in several stages , as described in the main text . Detailed equations are provided in the following sections . Bifurcation diagrams were produced using MatCont [35] , a numerical continuation package designed in MatLab ( MathWorks ) . The full set of partial differential equations ( PDEs ) for each model were simulated using an implicit-diffusion explicit-reaction scheme with grid values . Example PDE simulations are seen in Figure 4 . In general , we consider up to three GTPases: Cdc42 , Rac , and Rho , each of which is assumed to have three forms . Rather than writing all 9 PDEs , we here provide the form for a given GTPase , using the notation to represent any one of Cdc42 , Rac , and Rho . Let ( ) denote the level of active ( inactive ) membrane bound GTPase and let denote its cytosolic form . The total amount of GTPase in all these forms , , is assumed to be constant over the domain on the timescale of the experiments . Based on the schematic shown in Figure 2a ) , we write the set of equations ( 11 ) where are membrane and cyosolic rates of diffusion , is GAP-mediated inactivation rate , is the membrane dissociation rate , and the membrane association rate . is a GEF-mediated activation rate and depends on crosstalk assumed in the specific models discussed . Based on the 1D experimental geometry and controlled stimulus , it is reasonable to neglect gradients in all but the length direction . Define a 1D projection of the variable as ( 12 ) where is approximated as nearly uniform across the width and depth directions . Physically , represents the number of molecules in a slice of width within the cell . It follows that ( 13 ) As ( but not ) is directly observable experimentally , we rewrite to eliminate the less readily measurable cell depth . We now invoke the assumption that cycling between membrane and cytosol is very fast to make a quasi steady state ( QSS ) assumption . Then the fractions of the inactive form on the membrane and in the cytosol are , respectively , and . We now define a composite inactive form by ( 14 ) and an “effective diffusion constant” ( 15 ) The parameter is a composite that weights the respective rates of diffusion of and by the average time spent on the membrane versus the cytosol . With this reduction , we reduce the system of three equations ( 11 ) to a system of two equations in one space dimension and obtain Eqs ( 1 ) . The normalization factor has been introduced to simplify parameter identification . We henceforth use the notation . Let represent the phosphoinositides PIP , and . The interconversions of these are shown in bottom layer of each panel in Figure 1 . We incorporate the feedbacks to phosphorylation by PI5K , and PI3K , and dephosphorylation by PTEN in the functions . The set of equations adopted for the PIs are similar to those in [18] , [19] , ( 16 ) is a constant source of , and a constant rate of decay . All PIs are assigned the same rate of diffusion , . We briefly outline the LPA method first introduced in [36] . The method simplifies the system of PDEs by considering the limit of infinitely fast diffusion of inactive GTPases and infinitely slow diffusion of the active GTPases . Under this limit , the full system of PDEs can be reduced to a system of ODEs that provide information about the initial growth of perturbations . This diffusion limit is particularly relevant to small GTPases where rates of diffusion of cytosolic and membrane bound forms vary by orders of magnitude . Now consider a small perturbation that leads to localized high activation of the GTPase ( square pulse in Figure 9 ) . In the given diffusion limit , the active form will take on a local behaviour near the pulse , and some uniform global behaviour far away . We denote those levels by , respectively , ( local ) and ( global ) as indicated in Figure 9 . In the limit these two hardly interact . In contrast , in the limit , the inactive form will take on a purely global behaviour , distributing the effect of the perturbation instantly . The PDE system ( 1 ) can then be approximated by the set of ODE's ( 17 ) for some initial time period until the perturbation is no longer localized . Applying the conservation of each GTPase and assuming the perturbation to be small in size yields . In this case can be eliminated , leading to ( 18 ) A bifurcation analysis of the reduced ODE system provides clues as to how a localized perturbation will evolve over time in the PDE system . Even though the two mathematical structures are distinct , the large disparity in the true diffusion rates makes the LPA reduction a good approximation . The bifurcation diagram in Figure 3 shows the results of this method applied to the GTPase system with no PI feedback ( ) . In this case the system of 6 PDE's reduces to 9 ODE's ( 3 for each GTPase ) and , using conservation , further reduces to 6 ODE's ( 2 for each GTPase ) . The blue curve represents the steady states of the reduced system where . This is a solution of the well mixed system . It is also a homogeneous steady state ( HSS ) of the original PDEs that corresponds to a spatially uniform “rest state” of a cell before a stimulus ( local pulse ) is applied . Red curves represent additional states that can be reached by a highly localized patch while the bulk of the cell remains at its HSS . Dashed ( solid ) lines indicate that the state is unstable ( stable ) to arbitrarily small localized perturbations . While the details of the patterned states are not depicted in this type of bifurcation plot , the qualitative behaviour of the rest state and its response to a pulse can be seen . Four distinct parameter regions are found: insensitive ( I ) , polarizable ( bistable ) ( II ) , ultrasensitive ( Turing unstable ) ( III ) , and overstimulated ( IV ) . Cells with states represented by Region ( I ) do not respond to a pulse stimulus , and return to the rest state rather than polarizing . Cells with state in Region ( IV ) have a uniformly high level of active GTPase throughout , and cannot polarize - they might typically flatten and protrude in all directions , but retain their uniform GTPase distribution . Region III represents cell states wherein polarization can occur spontaneously , or in response to noise of arbitrarily small magnitude . Finally , Region II represents cells that require a heterogeneous stimulus with a sufficiently asymetric profile in order to polarize . That is , the stimulus must be sufficient for part of the cell to breach some threshold ( depicted by the dotted red elliptical arc ) . Mathematically , these observations can be inferred from Figure 3 as follows . In the insensitive regions , there is a single HSS ( single solid blue curve in Regions I , IV of Figure 3 ) ; this means that local perturbations or arbitrarily large amplitude decay back to that HSS and no spatial patterning can form . In the ultrasensitive region , the HSS ( dotted blue line in Region III ) is unstable to arbitrarily small heterogeneous perturbations , so that any noise will lead to new attractor states ( represented by two solid red elliptical arcs in region III ) . In the polarizable region , the HSS is locally stable: both homogeneous and small heterogeneous perturbations decay back to this HSS . However , a sufficiently large local perturbation that increases the local level of one of the active GTPases beyond the threshold ( dotted red elliptical arc in Region II , representing a repeller state ) can induce patterning . The vertical distance between the HSS and repeller represents the magnitude of perturbation required to produce the spatially heterogeneous polarized state . This analysis of the local-global LPA reduction provides insights , but is not fully predictive of the behaviour of the PDE system with finite rates of diffusion . The related collection of ODEs provides an approximation of the PDEs only as long as the perturbation is spatially localized . Once it spreads and a pattern begins to emerge , an asymptotic assumption that the integrated size of the perturbation be small fails and the approximation breaks down . Further , the bifurcation points present in the related ODEs are an approximation , rather than exact match , to full PDE bifurcation points . Thus , numerical simulations are necessary to provide a more complete understanding of the system . Figure 4 shows numerical solutions of the PDE system in the ( “kymograph” ) plane . Two pattern-forming regimes predicted in Figure 3 are ilustrated . In the bistable case ( left ) , an initial local perturbation induces a wave that propagates into the domain and finally stalls , indicative of wave pinning . In the ultrasensitive regime , which is representative of noise sensitive cells , standard Turing patterning occurs where a wave with some dominant wave-number destabilizes the HSS and grows . Note that alternative techniques such as Turing stability analysis could be used to detect this regime . However , for our simplest model of 6 nonlinear PDEs , such analysis is challenging , and less revealing . LPA is a simpler alternative that provides an excellent numerical approximation for the Turing regime as well as its relationship to the WP regime . Finally , because our experimental cells have a stable rest state , the Turing regime is a less suitable regime to explore .
|
Cell polarization is associated with intracellular gradients of signaling proteins such as Rho GTPases that organize the cytoskeleton in cell motility . We previously observed cells in microfluidic channels and studied their polarization and motility in a simplified ( nearly 1 dimensional ) geometry . There , precise gradients of chemically-inducible molecular probes were presented to elicit gradients of active Rac , independent of the upstream signaling . Here we develop a set of spatio-temporal mathematical models to account for the observed polarization behaviour of those cells , and their threshold response to induced Rac activity . These reaction-diffusion models for the interactions of signaling proteins ( GTPases Rac , Rho , and Cdc42 ) and membrane lipids ( phosphoinositides PIP , , ) are analyzed by a new method ( ‘Local Perturbation Analysis’ ) that explores the effect that pulses of stimuli have on local ( global ) variables , i . e . those intermediates that have slow ( fast ) rates of diffusion . Together , the models and experiments suggest that ( 1 ) spatially uniform stimulation makes the cells more sensitive to applied gradients . ( 2 ) Feedback between phosphoinositides and Rho GTPases sensitizes a cell . ( 3 ) Cell lengthening/flattening accompanying polarization can increase the sensitivity of a cell and stabilize an otherwise unstable polarization .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"systems",
"biology",
"mathematics",
"applied",
"mathematics",
"regulatory",
"networks",
"biology",
"computational",
"biology",
"signaling",
"networks"
] |
2012
|
Modelling Cell Polarization Driven by Synthetic Spatially Graded Rac Activation
|
Women comprise a minority of the Science , Technology , Engineering , Mathematics , and Medicine ( STEMM ) workforce . Quantifying the gender gap may identify fields that will not reach parity without intervention , reveal underappreciated biases , and inform benchmarks for gender balance among conference speakers , editors , and hiring committees . Using the PubMed and arXiv databases , we estimated the gender of 36 million authors from >100 countries publishing in >6000 journals , covering most STEMM disciplines over the last 15 years , and made a web app allowing easy access to the data ( https://lukeholman . github . io/genderGap/ ) . Despite recent progress , the gender gap appears likely to persist for generations , particularly in surgery , computer science , physics , and maths . The gap is especially large in authorship positions associated with seniority , and prestigious journals have fewer women authors . Additionally , we estimate that men are invited by journals to submit papers at approximately double the rate of women . Wealthy countries , notably Japan , Germany , and Switzerland , had fewer women authors than poorer ones . We conclude that the STEMM gender gap will not close without further reforms in education , mentoring , and academic publishing .
Although women are increasingly studying Science , Technology , Engineering , Mathematics , and Medicine ( STEMM ) subjects at university , women comprise a minority of senior staff , are less often trained in elite research groups , are promoted more slowly , and are more likely to leave STEMM careers [1–3] . Academic publications are the primary means of disseminating scientific knowledge and the principal measure of research productivity [4] and thus influence the career prospects and visibility of women in STEMM . Author lists of these publications also provide information on the gender ratio of people working in a given field . For these reasons , at least 61 studies have estimated the gender ratio of authors on academic publications ( S1 Table ) . Of these , 52 used manual data collection ( e . g . , reading author lists ) , limiting their scope , while 9 used computational approaches , producing enough data to address more complex questions . For example , computational studies have so far mapped differences in gender ratio across research disciplines [5 , 6] , compared geographic regions [6 , 7] , revealed biases in citation rate [7–9] , and shown that women tend to do a greater share of experimental work [10] . Although there is a consensus that the STEMM gender gap is shrinking ( S1 Table ) , to our knowledge , no study has used formal modelling to predict when the gap will close , so it remains unclear when parity will be reached given present rates of change . Performing this analysis is necessary to identify disciplines that will retain an imbalanced gender ratio without additional interventions and may help to uncover previously unrecognised biases . We aimed to determine the gender and country of affiliation for each author on every publication listed in the PubMed database and the arXiv preprint server ( see Methods ) . Together , PubMed and arXiv index around 30 million articles from the medical and life sciences , chemistry , physics , mathematics , computer science , and certain branches of engineering . We managed to assign gender with ≥95% confidence to 35 . 5 million authors from 9 . 15 million articles indexed on PubMed ( 2002–present ) and to 1 . 1 million authors from 0 . 5 million arXiv preprints ( 1991–present ) . We obtained sufficient data to measure the author gender ratio , its rate of change , and the expected number of years to reach gender parity for 4 , 720 journals and 119 arXiv sub-categories ( S1 , S2 and S3 Data ) . Our web app ( https://lukeholman . github . io/genderGap/ ) allows one to explore the PubMed data and view the past , present , and projected future gender ratio for approximately 25 , 000 combinations of research discipline , journal , authorship position , and country . We also confirmed that the number of women authoring research papers is a reliable predictor of the number of women working in each discipline and that the gender assigned by our computational methods to author names was correct roughly 99 . 7% of the time ( see Methods ) .
Figs 1 and 2 reveal that 87 of the 115 disciplines examined have significantly fewer than 45% women authors , 5 have significantly more than 55% , and the remaining 23 are within 5% of gender parity . Topics such as physics , computer science , mathematics , surgery , and chemistry had the fewest women authors , while health-related disciplines like nursing , midwifery , and palliative care had the most . Of the gender-biased disciplines , almost all are moving towards parity , though some are predicted to take decades or even centuries to reach it . Nursing , midwifery , and critical care were the only disciplines in which men authors are becoming significantly more common . The PubMed categories Social Sciences ( which contains predominantly Social Work journals ) and Speech-Language Pathology currently have >50% women authors and are becoming significantly more female-biased . Comparable information for dozens of arXiv subcategories is shown in S1–S7 Figs: gender ratio varies by up to 20% between subfields of physics , mathematics , and computer science . In almost all disciplines examined , women were substantially underrepresented as the last-named author in the author list and as single authors and overrepresented as first authors ( contradicting prior studies [5–7] ) relative to the overall author gender ratio . A small minority of journals bucked the overall trend and had fewer women first authors than expected , rather than more; these journals were predominantly well-known , prestigious titles such as Nature , Lancet , New England Journal of Medicine , and BMJ ( S8 Fig ) . In most disciplines represented in our dataset , prevailing conventions regarding authorship order mean that first authors are usually early career researchers , while last authors tend to be comparatively senior [11] . Thus , these results suggest that early career researchers are more likely to be women and senior researchers more likely to be men , relative to the overall gender ratio of the discipline in question , consistent with United States-specific survey data showing that the underrepresentation of women is highest among senior academics [2] . In some fields , the convention is for the authors to be listed alphabetically by surname . Publications in our dataset using the alphabetical surname convention would tend to dilute the overall difference in gender ratio between first and last authors , meaning that the true difference in gender ratio between early career and senior researchers might be greater than our results suggest . The underrepresentation of women as last authors and single authors probably has multiple , complex causes . Firstly , the number of women graduates was lower in the past , when today’s senior researchers were training ( termed ‘demographic inertia’ [2] ) . However , Shaw and Stanton [2] used demographic data to show that demographic inertia can only partly explain the present shortage of senior women researchers in the US . Shaw and Stanton’s results ( among others , e . g . , [12] ) point to a second reason for the dearth of senior women in STEMM: that women are more likely than men to leave STEMM careers before progressing to senior positions . A common metaphor for this issue is the ‘leaky pipeline’ , which likens a STEMM career to a series of connected pipes ( e . g . , PhD student , junior researcher , group leader ) that ‘leaks’ greater numbers of women than men at particular junctures [2] . A third possibility is that if women progress to research leadership roles more slowly than men—for example , due to facing extra challenges inside and outside the workplace [3 , 13 , 14]—the average woman would have a higher ratio of first to last author publications over her career than the average man [15] . A fourth possibility is that discussions over authorship are influenced by gender , such that women are less likely to be offered , or to request , the last author position [16] . Finally , students and other junior researchers might be less likely to select women supervisors , e . g . , because equivalent achievements by women are judged less favourably [17 , 18] or because senior women are less often publicly celebrated as leaders in their fields [19] . Indeed , the National Academies of Science , Engineering , and Medicine ( US ) concluded that the deficit of women in STEMM is not because too few women enter the field or because women are less committed to their STEMM careers , but rather because ‘assumptions and stereotypes about gender operate in personal interactions , evaluative processes and departmental cultures that systematically impede women’s career advancement in academic medicine , science and engineering’ [1] . Worryingly , highly male-biased disciplines tended to show especially slow improvement in the gender ratio with time ( Figs 1 and 2 , S9 Fig ) . For example , the arXiv category Physics presently has around 13% women in the last author position , but this figure is only rising by c . 0 . 1% per year , such that the best-fitting nonlinear model ( see Eq 1 ) predicts that it will be 258 years ( 95% CI 194–383 ) before the gender ratio of senior physicists comes within 5% of parity ( S3 Data ) . Additionally , the difference in gender ratio between first and last authors was weakest in male-biased disciplines ( Figs 1 and 2 , S10 Fig ) ; one possible explanation is that the gender gap among the newest recruits to these fields is only marginally smaller than the gender gap at senior levels . These results suggest that mostly male fields might attract fewer women graduates , lose women researchers to other careers at a faster rate , and/or have stronger gender biases that affect the relative publication rates of men and women . Thus , novel interventions appear necessary if we are to make progress in strongly gender-biased disciplines . Across countries , the gender ratio of all PubMed-indexed authors varied by >30% ( Fig 3 , S11–S18 Figs ) . Among the major research-producing countries , the STEMM gender gap was especially pronounced in Japan , Germany , and Switzerland . The most gender-equitable countries spanned Europe , South America , and Africa . Using data on gender equality and development collected by the United Nations , we found that countries in which children of both sexes attend school longer have more women authors , while countries with higher per capita income have fewer women authors ( S19 Fig; S2 and S3 Tables ) . Life expectancy , adolescent birth rate , percentage of women in parliament , and the education and labour force gender gaps were not significantly correlated with the STEMM gender gap . Though correlational , these results imply that wealth does not necessarily diminish gender inequality in the STEMM workforce , though access to education might . Cultural and historical factors are challenging to meaningfully capture in this type of analysis , but we suspect that they play a major role . Variation in author gender ratio between journals could reveal underappreciated biases against women and allow us to ask whether women’s research is published in equally visible , prestigious fora . For example , so-called ‘high impact’ journals tend to be well cited , widely read , and prestigious , and the same is true of many journals that specialise in reviews ( e . g . , Nature Reviews or Cell Press Trends journals ) . Both journal types also commonly reject more manuscripts , or publish more papers by invitation , which might disadvantage women [14 , 17] . By contrast , some Open-Access ( OA ) journals accept a comparatively high proportion of papers and are considered by some researchers to be less prestigious ( debated in [20] ) . We therefore tested whether impact factor and journal type correlated with author gender ratio . Journal impact factor ( standardised by discipline ) negatively correlated with the proportion of women authors . Review-focused journals also had fewer women authors than non-review-focused journals , and there were more women authors in OA than non-OA journals , particularly within review journals ( Fig 4; S5 Table ) . These results imply that women are disadvantaged and suggest remedial strategies . Prestigious journals tend to reject many submissions without peer review , and editors are usually aware of authors’ names ( and thus genders ) , even for journals that use double-blind peer review . Gender bias has been implicated in nonexperimental studies of peer review [21] and experimentally demonstrated in other academic contexts [14 , 17] , suggesting a need for double-blind editing and review . We hypothesized that women may be invited to submit academic papers less often than men [22] , given that this is the case for invited keynote lectures at some conferences [19] . A gender bias in the rate of invitations would directly contribute to the author gender gap and to differences in gender ratio between journals . By applying simulations to 3 , 067 invited papers detected in our dataset , we estimate that men are roughly 1 . 7–2 . 1 times more likely than women to be invited to submit papers ( see Methods; S4 Table ) . The gender ratio among authors of invited papers is even more male-biased than the gender ratio of last authors or single authors ( S4 Table ) . This suggests a need to scrutinise editorial practices [22] , appoint women editors [23] , and implement gender targets when using an invitation-based publishing model [19] . Another explanation for the elevated gender disparity we observed in higher-impact journals is that women submit a lower proportion of their manuscripts to prestigious journals . A recent study , upon finding that women’s papers passed peer review more frequently than men’s , hypothesised that women practice ‘better targeting of papers to a journal’ [24] . A more pessimistic hypothesis is that women are not encouraged to aim high by colleagues and mentors [1] or do not try because they believe themselves to have a lower chance of success [25] . These issues could be addressed via mentoring programs for staff and team leaders [1] and by taking steps [1 , 19] to promote women role models [26] . Because the variables examined here all potentially affect citation rate , our results partially explain previous reports [7–9] that women’s papers are cited less often than same-discipline papers written by men . The ‘citation gap’ would likely shrink if barriers to women publishing highly visible papers , e . g . , invited reviews in prestigious journals , were removed . In closing , our data reveal an authorship gender gap across STEMM , which is likely to remain for generations in certain fields . On a positive note , many disciplines are already close to having equal numbers of men and women authors , others are making steady progress towards parity , and no male-biased disciplines displayed a clear decline in the number of women authors . Women were especially common in the first authorship position across most fields , implying that , worldwide , increasing numbers of women are starting careers in STEMM . However , some STEMM disciplines ( e . g . Physics , Computer Science , Surgery ) clearly require additional interventions if parity is to be reached this century , and women were strongly underrepresented in authorship contexts ( namely last , single , and invited authors ) that are typically occupied by senior researchers , even in fields with gender parity in the overall pool of authors . We suggest that efforts to recruit and retain women in STEMM must be wide-ranging and could include dispelling the common but poorly evidenced belief that there are innate gender differences in STEMM aptitude [27 , 28]; reforming academic publishing and peer review [21]; ensuring women have equal access to informal professional networks [29]; affording greater recognition of the extra demands outside the workplace that traditionally fall on women when assessing candidates’ achievements [13]; guaranteeing women equal resources at work [9]; providing better access to parental leave [2 , 29] and additional provisions to help people return to work following a career break [30]; striving for a representative gender ratio of invited speakers at academic conferences [19]; and affirmative action during hiring . Our dataset has been publicly archived , and we hope it will prove useful to researchers , policy makers , journals , and scientific societies . Follow-up studies using the dataset could search for additional predictors of author gender ratio . For example , one could measure the impact of double-blind peer review or invitation-based submission models on the proportion of women authors . One could also seek to identify cultural and sociological factors common to countries with good gender balance in STEMM or measure the impact of policies implemented by countries , journals , or granting agencies aimed at promoting women in science .
The objective of our study was to assess the past , present , and future gender ratio of authors publishing in many different fields of STEMM . In particular , we sought to identify fields in which parity will not be reached for many years , assuming that present trends continue . The study began by downloading the author list for every single publication or pre-print indexed on the PubMed and arXiv databases . We then used the genderize . io database to assign genders to authors based on their given names and , where possible , determined the country in which each author was based from their academic affiliation . For the PubMed data , each journal was assigned to an academic discipline , using PubMed’s own classifications where possible . For each journal , discipline , and arXiv category , we fit a model to the data to estimate the present-day author gender ratio , its rate of change , and when ( or if ) gender parity will be reached . Additionally , we conducted several analyses searching for variables that explain variation in the author gender ratio . For example , we tested whether journal impact factor correlates with gender ratio , compared the author gender ratio of invited papers to that of papers submitted without an invitation , and searched for country-level correlates of the author gender ratio using metrics of equality and development collected by the United Nations . To facilitate easy access to our data , we have produced an interactive web application that allows one to view the past , present , and projected future author gender ratio for different combinations of journal , research discipline , country , and authorship position ( https://lukeholman . github . io/genderGap/ ) . The complete dataset is archived as a SQLite3 database at the Open Science Framework ( https://osf . io/bt9ya/ ) ) . Scripts used to collect and analyse the data , as well as a compact spreadsheet summarising our dataset , is archived at https://github . com/lukeholman/genderGapCode . The JavaScript and . json file underlying the web app is available at https://github . com/lukeholman/genderGap . We downloaded a local copy of PubMed’s MEDLINE database on 20 August 2016 using a shell script . For each unique article , we used R scripts to process the XML database entry and extract the first-listed given name of every author ( or the second given name , if the first was a single letter , as in B . Rosemary Grant ) . We also retrieved the paper’s title , the journal name , the addresses for all authors , the DOI , and the date on which the article was added to PubMed ( to the nearest day ) . From each author's affiliation , we attempted to determine the country in which they were based by pre-processing the addresses using libpostal ( github . com/openvenues/libpostal ) , followed up with custom R scripts that searched for the country name ( or inferred it from the state or city ) . Some major research-producing regions , e . g . , Taiwan and Hong Kong , were left separate from their countries . We filtered the PubMed data to exclude articles without authors and removed articles from journals that had fewer than 100 articles indexed on PubMed ( inspection suggested that these were usually not standard academic journals ) . On 17 October 2016 , we downloaded data for every preprint on arXiv using the R package aRxiv ( github . com/ropensci/aRxiv ) . For each article , we recorded ID , initial publication date , the authors’ given names , and the major and minor research specialties ( which are selected by the authors ) . Preprints lacking authors were discarded . Gender was assigned to given names using the genderize . io web server , which uses a large database of name–gender associations assembled by trawling social media websites ( >200 , 000 given names , with country-specific name–gender associations for approximately 80 countries ) . Whenever we knew an author’s country of affiliation , we specified the country when querying genderize . io to determine the gender of their name; this should reduce the misclassification rate ( e . g . , people named Kim are typically male in Denmark but female in the US ) . For each name , genderize . io returns a number ( the ‘gender score’ ) corresponding to the proportion of people with that name in the genderize . io dataset who are men or women; for example , 7% and 59% of people named ‘Chris’ and ‘Robin’ are women , respectively . Given the abundance of data , we elected to deal with this ambiguity by simply excluding names that were not associated with one gender with ≥95% frequency in the genderize . io database . This simple approach preserved 92 . 4% of the dataset ( 3 . 7 million authors were excluded , leaving 35 . 5 million ) . The gender score of the 3 . 7 million excluded names was not appreciably skewed towards zero or one ( gender score mean: 0 . 47 , median: 0 . 49 ) , indicating that this procedure did not bias our estimated gender ratios . To verify our computationally assigned gender dataset , we compared it to a similar , manually curated dataset from an in-prep paper by Michael McCarthy and colleagues ( University of Melbourne ) , in which author gender was ascertained via Google searches . Of the 372 authors that appeared in both datasets and were assigned a gender by us , there was only one error ( a man misclassified as a woman ) . Thus , we estimate our gender misclassification rate as 1/372 = 0 . 3% ( 95% CI 0%–1 . 7% ) . The manually collected data contained an additional 29 authors whose gender was left ‘unclassified’ in our dataset , either because their given names were associated with one gender <95% of the time ( 18 authors ) , were written as initials ( 6 authors ) , or were absent from genderize . io ( 5 authors ) . The manual dataset estimated the author gender ratio as 23 . 4% women ( n = 401 , 95% CI 19%–28% ) , while we estimated it as 24 . 4% ( n = 372 , CI 20%–29% ) ; thus , the accuracy and precision of our method was essentially the same as that afforded by manual assignment . See also S21 Fig for additional evidence that our methods produced accurate estimates of the gender ratio . Lastly , we note that within any given dataset , male-to-female and female-to-male misclassification errors will partially cancel out ( e . g . , 10 misclassified men and 6 misclassified women would have the same impact on the author counts as 4 misclassified men , making 16 errors behave like 4 ) , reducing the impact of errors . We next classified PubMed-indexed journals into STEMM research disciplines . Because subjective assignment of journals to disciplines could theoretically introduce bias , we used PubMed’s preexisting classifications wherever available . PubMed maintains a database of information on many ( but not all ) of the journals that they index ( ftp://ftp . nlm . nih . gov/online/journals/ ) , and in some cases , PubMed curators have manually assigned journals to a research discipline . Additionally , PubMed have assigned Medical Subject Heading ( MeSH ) terms to many journals; MeSH terms derive from text mining and describe a journal's content ( nlm . nih . gov/mesh ) . We assigned a single discipline to each journal using the manually assigned heading ( if present ) , a random PubMed category listed in the MeSH terms ( if the former was not present ) , and , as a last resort , keywords in the journal's title ( e . g . , journals with ‘Kidney’ in the title were added to the PubMed category ‘Nephrology’ ) . Assignment was done blind to the gender data; we focused on categorising journals that contained a lot of data and stopped once further categorisation became laborious , leaving 579 journals ( median: 293 papers each ) as ‘Unassigned’ . Although using the PubMed categorisations has the advantage of being objective ( preventing us from inadvertently biasing the results ) , a limitation of this approach is that the various subfields of medicine are precisely categorised , but other disciplines are not . For example , PubMed indexes many ecology journals but classifies them under broader headings , such as Biology . Readers interested in a discipline that is missing from our set can use our web app to search for data from relevant journals . We also used data from the Directory of Open Access Journals ( www . doaj . org ) to classify journals as OA or non-OA . Only journals in which all content is freely accessible were classified as OA; journals that allow OA publishing for an additional fee were classified as non-OA . We also obtained the 3-year impact factor of each journal tracked by Clarivate Analytics ( formerly the Institute for Scientific information [ISI] ) . Although we recognise that a journal’s impact factor is a weak predictor of the quality or citation count of any given paper in that journal , impact factor is ( for better or worse ) a widely used metric of journal prestige and the productivity of individual researchers , and journals with the highest impact factor in their discipline tend to be more well-established and widely read [31] . Thus , it is worthwhile to determine whether the author gender ratio differs between high and low impact factor journals , since this could affect the relative visibility of women authors or their perceived research productivity . Journals with a title containing the words “Review ( s ) ” or “Trends” were classified as review-focused journals , while the remainder were classified as non-review-focused . In initial trials , we also attempted to classify individual articles as OA or as reviews ( not just whole journals ) , but this proved difficult to do reliably . Our journal-level analyses are probably conservative , because OA or review-type articles that are published in predominantly non-OA and non-review-focused journals will be misclassified ( and vice versa ) , reducing the observed difference in author gender ratio . We obtained country-level metrics on gender inequality , wealth , and development from the United Nations Human Development Report 2015 ( http://hdr . undp . org/en/global-reports ) . We downloaded all the available UN variables , except the metrics of poverty ( which are unavailable for most major research-producing countries ) and the human development index and maternal mortality rate ( because these were >90% correlated with other variables in the set ) , resulting in seven predictors . We focused on metrics of gender equality and development due to our a priori expectation that these might be key predictors of the number of women in science . We analysed four types of authorship . We recorded the number of men and women listed as A ) first authors of multiauthor articles , B ) last authors of multiauthor articles , C ) authors of single-author articles , D ) authors covering all authorship positions of both single- and multiauthor articles ( termed ‘overall’ ) . To estimate the present-day gender ratio and its rate of change , we assumed that the proportion of women authors ( p ) scales with publication date via the logistic function: p=e0 . 5rt2e0 . 5rt+c ( 1 ) where t is the date ( to the nearest day , expressed as a decimal number of years before or after 1 January 2000 ) , r controls the steepness of the curve , and c varies its inflection point . This model assumes that the relationship between gender ratio and time is sigmoidal and progresses monotonically either towards gender parity or the complete disappearance of one gender . The model thus allows for nonlinear rates of change in the gender ratio ( which were common in our data ) and accommodates our strong prior expectation that gender-biased disciplines that are heading towards parity are likely to eventually plateau at parity ( as opposed to overshooting parity or reversing direction , as occurs if one uses alternative methods such as quadratic linear regression or generalized additive models ) . Some combinations of r and c also allow the relationship between t and p to appear almost flat over long timescales , allowing the model to accommodate an essentially unchanging gender ratio . We found the maximum likelihood pair of r and t values for a given set of data using R’s optim function and substituted them into the above formula to estimate the gender ratio on the present day ( i . e . , 20 August 2016 , the date when the data were retrieved from PubMed ) . We also estimated the present-day rate of change in the gender ratio ( by differentiating Eq 1 with respect to t ) and the projected number of years until gender parity . We estimated the 95% confidence intervals on these 3 variables using bootstrap resampling of the data ( n = 1 , 000 ) . To ensure accurate prediction , we only predicted the gender ratio , its rate of change , and time-to-parity on subsets of the data that contained at least 100 papers and which contained at least 50 authors of known gender per year for five or more years . Our PubMed gender data are almost entirely from 2002–present , because earlier PubMed entries generally list authors’ initials instead of their given names . arXiv was founded in 1991 , though 97% of entries were from 1998 or later . Programmatic gender assignment was mostly successful: 48 . 7 million authors were listed in the subset of PubMed articles that listed at least one author’s given name , and we were able to assign gender with ≥95% confidence to 35 . 5 million of these ( i . e . , 73% , or 76% if discounting authors who only gave their initials; S20 Fig ) . Because <3% of the authors in our PubMed dataset used initials rather than their full name , a gender difference in the tendency to publish under one’s initials would not appreciably skew our results . We detected 116 countries and territories in authors’ affiliations . Gender was assigned with ≥95% confidence to at least 70% of authors for 96/116 countries ( S20 Fig ) , though gender assignment was frequently unsuccessful for authors with affiliations in some East Asian countries , due to a high frequency of ( Romanized ) given names that are commonly used by both men and women . However , for most countries , failures to assign gender were infrequent enough that they could not add significant bias , even in the unlikely event that programmatic gender assignment was substantially more fallible for names of one particular gender . Gender was inferred with ≥95% confidence for 1 . 18 million authors ( 61 . 1% ) from 538 , 688 arXiv preprints . This lower success rate reflects the fact that arXiv does not mandate a consistent style for presenting authors’ names , and more authors only provided their initials—the success rate was 85 . 0% when excluding these authors . The following analysis was conducted on the subset of authors whose genders and countries were known , and the authors were from a country listed in the UN dataset ( n = 22 , 510 , 436 authors ) . We fit a linear mixed model with the percentage of women authors as the response variable , because initial tests showed that binomial generalized linear mixed models ( with author gender as the response variable ) did not converge ( using the lmer and glmer functions in the lme4 package for R , respectively ) . The model had the following formula: %womenauthors∼a1Position+a2Date+a3x1+a4x2+a5x3+a6x4+a7x5+a8x6+a9x7+ ( Date|Journal ) + ( Date|Discipline ) + ( Date|Country ) where x1–x7 are the seven UN predictor variables and terms in parentheses are random effects . ‘Date’ is publication date to the nearest year ( treated as a continuous variable ) , and ‘Position’ is authorship position ( first , last , middle , or single; first authors were treated as the reference level , and ‘middle’ authors were those whose names appeared between the first and last authors of multiauthor papers ) . To standardise the units of the model coefficients , publication date and all the UN predictor variables were transformed to have mean 0 and variance 1 . The notation ‘Date|Journal’ indicates that ‘Journal’ was treated as a random intercept and publication date as a random slope . Thus , the model estimates the fixed effects ( a1–a9 ) after controlling for differences in the gender ratio between journals , disciplines , and countries , where these differences potentially vary linearly with time . The marginal R2 of the model ( i . e . , the proportion of variance explained by the fixed factors ) was 0 . 08 , and the conditional R2 ( the proportion of variance explained by both fixed and random factors ) was 0 . 66 . We searched for papers that were published by invitation by examining the titles of all papers in the PubMed dataset . We defined a paper as ‘invited’ if its title began with the word ‘invited’ followed by another word , such as article , comment ( ary ) , discussion , editorial , essay , review , or paper; these second words were selected based on manual inspection of every title beginning with ‘invited’ . Note that this analysis presumably misses very many invited papers that do not have the word ‘invited’ in their titles . This limitation means that our analysis probably underestimates the true difference in gender ratio between invited and noninvited papers , but it seems unlikely that it could produce a spurious difference in gender ratio . We then wrote a simulation to test whether these invited papers had significantly fewer women authors than would be expected under the null hypothesis that invited papers have the same gender ratio as for all papers . In short , the simulation uses information from our dataset to ask , ‘Is the frequency of women authors on invited papers lower than expected , after controlling for variation in gender ratio due to journal , publication date , and ( optionally ) authorship position ? ’ Modelling these sources of variation is necessary because the frequency of invited papers differs across journals and across time , and our sample of invited papers contains an atypically high frequency of single-author papers , all of which could produce a spurious correlation between invitation and gender if not controlled for . To obtain the null expected author gender ratio for the 3 , 067 invited papers , we generated 10 , 000 random datasets , each of which had the same number of papers and same distribution of authors per paper as our sample of invited papers but which had randomly generated author genders . Each author’s gender was randomly generated by sampling from a binomial distribution with probability p ( Woman | Journal , Date , Position ) , i . e . , the probability of the focal author being a woman , given the journal in which that person was publishing , the date on which the paper appeared , and the authorship position of the focal person ( i . e . , first , last , middle , or single author—we assumed that the gender ratio for middle authors was the same as the overall gender ratio , unless otherwise stated ) . We obtained this probability by substituting the values of r and c estimated using the data ( i . e . , invited and noninvited papers together ) into Eq 1 . This gave us the expected distribution of gender ratios for the invited papers , under the null hypothesis that men and women are invited equally often , given the simulation’s assumptions . Because the results are sensitive to the simulation’s assumptions , we used four sets of assumptions of varying strictness when defining p ( Woman ) , which are outlined in S4 Table . The four assumptions all gave a qualitatively identical answer , namely that authors of invited papers are more likely to be men relative to our predictions under the null models . Differences in author gender ratio between disciplines , journals , and countries could be due to variation in the number of women researchers and/or to variation in the relative publication rates of men and women . It is also conceivable that we made errors when measuring the author gender ratio that escaped detection . To address these concerns , we tested whether the author gender ratio that we estimated from PubMed was correlated with estimates of the number of women working in various subdisciplines derived from surveys conducted by the US’s National Science Foundation ( NSF ) . Assuming that research disciplines in which many women work have more women authors and that our methods lack errors , we predict a strong positive correlation between the NSF data and our own . The NSF’s Survey of Graduate Students and Postdoctorates in Science and Engineering provides data on the gender ratio of PhD students and postdoctoral researchers working in a number of disciplines ( Tables 21 , 22 , and 38 at https://ncsesdata . nsf . gov/gradpostdoc/ ) . Unfortunately , we could not use NSF’s data on more senior researchers , because the NSF chose not to group senior researchers into the same fine-scale disciplines as the ones they use for early career researchers ( senior scientists were instead grouped into much broader categories , e . g . , ‘Life Sciences’ ) . Blind to the gender data , we manually matched each NSF research category with a similar discipline used by PubMed and tested for a correlation between the author data ( using US-based authors only ) and the NSF gender data for PhDs and postdocs using linear regression . We found that our gender ratio data were tightly correlated with the proportion of women PhD students and postdocs as measured by the NSF ( R2 = 0 . 48–0 . 64; S21 Fig ) ; in particular , the gender ratio among first authors was within 5% of the gender ratio of postdocs in most disciplines . Thus , it is indeed possible to use authorship data to reliably compare the gender ratios of different STEMM disciplines , at least for early career researchers , and we found no evidence that our methods introduced significant errors . However , this analysis revealed that our PubMed data sometimes underestimated the number of women early career researchers , particularly in female-biased disciplines or when using authorship positions other than the first ( S21 Fig ) . This implies that , among US academics , the average man has more publications than the average women ( especially last- and single-author publications ) . Note that the average male researcher has been publishing for longer than the average female researcher [2] , and S21 Fig does not account for gender differences in career length; thus , the gender difference in publication number is probably smaller than implied by the apparent shortfall in S21 Fig .
|
In most fields of science , medicine , and technology research , men comprise more than half of the workforce , particularly at senior levels . Most previous work has concluded that the gender gap is smaller today than it was in the past , giving the impression that there will soon be equal numbers of men and women researchers and that current initiatives to recruit and retain more women are working adequately . Here , we used computational methods to determine the numbers of men and women authors listed on >10 million academic papers published since 2002 , allowing us to precisely estimate the gender gap among researchers , as well as its rate of change , for most disciplines of science and medicine . We conclude that many research specialties ( e . g . , surgery , computer science , physics , and maths ) will not reach gender parity this century , given present-day rates of increase in the number of women authors . Additionally , the gender gap varies greatly across countries , with Japan , Germany , and Switzerland having strikingly few women authors . Women were less often commissioned to write ‘invited’ papers , consistent with gender bias by journal editors , and were less often found in authorship positions usually associated with seniority ( i . e . , the last-listed or sole author ) . Our results support a need for further reforms to close the gender gap .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Methods"
] |
[
"open",
"science",
"open",
"access",
"publishing",
"employment",
"publication",
"practices",
"engineering",
"and",
"technology",
"bibliometrics",
"sexual",
"and",
"gender",
"issues",
"sociology",
"social",
"sciences",
"careers",
"research",
"and",
"analysis",
"methods",
"meta-research",
"article",
"computer",
"and",
"information",
"sciences",
"peer",
"review",
"labor",
"economics",
"economics",
"research",
"assessment",
"science",
"policy",
"scientific",
"publishing"
] |
2018
|
The gender gap in science: How long until women are equally represented?
|
Leprosy is an endemic infectious disease caused by Mycobacterium leprae that predominantly attacks the skin and peripheral nerves , leading to progressive impairment of motor , sensory and autonomic function . Little is known about how this peripheral neuropathy affects corticospinal excitability of handgrip muscles . Our purpose was to explore the motor cortex organization after progressive peripheral nerve injury and upper-limb dysfunction induced by leprosy using noninvasive transcranial magnetic stimulation ( TMS ) . In a cross-sectional study design , we mapped bilaterally in the primary motor cortex ( M1 ) the representations of the hand flexor digitorum superficialis ( FDS ) , as well as of the intrinsic hand muscles abductor pollicis brevis ( APB ) , first dorsal interosseous ( FDI ) and abductor digiti minimi ( ADM ) . All participants underwent clinical assessment , handgrip dynamometry and motor and sensory nerve conduction exams 30 days before mapping . Wilcoxon signed rank and Mann-Whitney tests were performed with an alpha-value of p<0 . 05 . Dynamometry performance of the patients’ most affected hand ( MAH ) , was worse than that of the less affected hand ( LAH ) and of healthy controls participants ( p = 0 . 031 ) , confirming handgrip impairment . Motor threshold ( MT ) of the FDS muscle was higher in both hemispheres in patients as compared to controls , and lower in the hemisphere contralateral to the MAH when compared to that of the LAH . Moreover , motor evoked potential ( MEP ) amplitudes collected in the FDS of the MAH were higher in comparison to those of controls . Strikingly , MEPs in the intrinsic hand muscle FDI had lower amplitudes in the hemisphere contralateral to MAH as compared to those of the LAH and the control group . Taken together , these results are suggestive of a more robust representation of an extrinsic hand flexor and impaired intrinsic hand muscle function in the hemisphere contralateral to the MAH due to leprosy . Decreased sensory-motor function induced by leprosy affects handgrip muscle representation in M1 .
Leprosy , also known as Hansen's disease , is a chronic human granulomatous bacilliferous infection caused by the obligate intracellular organism Mycobacterium leprae [1] . Leprosy continues to be an important health problem worldwide , particularly in India , Brazil , Democratic Republic of Congo , Tanzania , Nepal , Mozambique , China and Nigeria [2 , 3] . The bacillus has a predisposition to infect cutaneous and peripheral nervous tissues , which allows infiltration into Schwann cells , resulting in nerve inflammation , most frequently in the eyes , hands and feet . This causes partial or total loss of sensory , motor and autonomic functions in the territory of the affected nerve resulting in skin anesthesia and dryness , as well as a decrease in proprioception and muscle paresis/hypotrophy [4] . Nerve damage may happen before , during or after treatment with the multidrug therapy ( MDT ) recommended by World Health Organization . In other words , even after bacteriological cure , leprosy can cause permanent physical disabilities and deformities . The installation of those deformities contributes to social exclusion , psychological disorders , and self-stigma , as has been recorded in studies about social participation and quality of life . Since 1985 , 14 million individuals have received MDT [5] . Despite these efforts , every year , many patients develop upper limb disabilities and are in need of rehabilitation services to control the chronic consequences of neural damage , such as claw hand , neuropathic pain and burns , requiring technical and scientific advances and a deeper understanding of the outcomes of either short or long-term rehabilitation . A considerable amount of literature has been published on leprosy . Several studies have focused on understanding the basic aspects of leprosy , such as genetics [6 , 7] , physiology of the bacilli [8 , 9] , biological markers [10] , kinesiology and biomechanical factors [11 , 12 , 13] , and public health concerns [14 , 15] . Until recently , however , there has been little discussion about the relationship between the peripheral upper-limb dysfunction caused by leprosy and changes in the motor cortex [16 , 17] . This phenomenon , known as plasticity , refers to the physiological and structural changes that occur in the central nervous system over time . The adult brain is capable of profound plasticity after peripheral lesions [18 , 19 , 20 , 21 , 22 , 23 , 24] . The potential for peripheral nerve injury to reorganize motor cortical representations was initially investigated in animal models [25] . Motor nerve injury is sufficient to produce changes in the primary motor cortex ( M1 ) of mammals , these changes appearing as early as a few hours , days and even months after peripheral nerve injury [23 , 26] . Motor cortex plasticity has also been demonstrated after peripheral lesions caused by amputation and/or phantom pain [27 , 28 , 29 , 30] , upper-limb muscle reconstruction and nerve transfer [21 , 31 , 32] , uni and bilateral heterotopic hand transplantation [33 , 34 , 35] , peripheral immobilization [36] and focal dystonia [19] . In most cases stated above , plasticity was accessed by means of transcranial magnetic stimulation ( TMS ) mapping . It remains unknown if and how the human motor cortex reorganizes after the typical gradual peripheral nerve damage provoked by leprosy . In the present study , we used single-pulse TMS mapping to evaluate motor organization in M1 after MDT in adult chronic leprosy patients with persistent hand disabilities . We hypothesized that patients with chronic damage in upper limb peripheral nerve caused by leprosy could present changes in corticospinal excitability as well as in hand grip muscle representation organization in M1 .
This study employed a cross-section design where a transcranial magnetic stimulation ( TMS ) mapping protocol was used to evaluate the cortical representation of selected hand muscles in M1 contralateral and ipsilateral to the most affected hand in chronic leprosy patients and healthy subjects in Hospital Federal Clementino Fraga Filho , RJ , Brazil , over the period of 2009 to 2013 . All participants provided written , informed consent , consistent with the Declaration of Helsinki . The Human Research Ethics Committee ( CEP-HUCFF/UFRJ from the Universidade Federal do Rio de Janeiro , UFRJ , under registry 143/09 ) approved this study . Six right-handed adult patients with chronic leprosy , grade 2 disability ( 4 males; 31 . 2 ± 4 . 7 years; mean age ± SD; Table 1 ) and 6 healthy controls , matched in gender and handedness ( 4 males; 27 . 2 ± 4 . 6 years ) , participated in this study . Handedness was determined using the revised Edinburgh Handedness Inventory [37] . Inclusion criteria were: both gender with long-term leprosy waiting for surgery to correct claw deformities in hands , with age between 18 to 45 years old . The exclusion criteria where: previous fracture in the upper extremity , use of central nervous system medications , demyelinating disorders , history of neurological deficits , stroke , diabetes , systemic disease or migraine headache , cardiac pacemaker placement , osteoarthritis , history of specific repetitive motor activity or any putative adverse reaction to TMS [38 , 39 , 40] . Two independent evaluators conducted a clinical assessment protocol for all participants , including anamnesis interview and active search for other information in the medical records . Dynamometry testing required all measurements to be taken with subjects sitting comfortably on a chair with their hand resting on an armrest and feet flat on the floor . The volunteers held a digital dynamometer ( EMG System do Brasil ) with the shoulder and wrist at a neutral position with the forearm supported and the elbow positioned at 90 degrees [41 , 42 , 43] . The importance of maintaining this position was explained to the subjects and repeated in both hands . In some cases , the evaluator assisted the patients to maintain the device in the proper position . The trials began with 60 s . of rest , after which the participants were asked to maintain maximal grip contraction for six seconds . A mean of three trials was collected with intervals of 60 s . between trials . The first and last second of each trial were discarded . To ensure maximum effort , a verbal cue was given to the subject while he performed the test . Leprosy neuropathy , despite being primarily demyelinating , frequently leads to axonal loss . Nerve conduction studies are considered the most objective method of assessing nerve function [44] to confirm the diagnosis/prognostic of neuropathy in leprosy patients . Neurophysiological examination of the nerves frequently shows that once axonal loss has been installed , nerve function is little affected by inflammatory , immune and/or bacterial events since chronic neuropathy has been established , inevitably leading to the well-known leprosy sequelae occurring at any time before and/or after leprosy diagnosis [45] . Patients in this study were examined in a nerve conduction study approximately 30 days before the TMS mapping . The ENMG protocol for the study of motor conduction of median and ulnar nerves was performed . The amplitude of the compound muscle action potentials ( CMAP ) as well as the latencies and motor conduction velocity ( MCV ) were measured . For the sensory conduction study in the same nerves , the compound sensory action potential amplitude ( CSAP ) , distal latency and velocity of sensory conduction velocity ( SCV ) were measured . The skin temperature was measured and maintained above 32 degrees Celsius . Prior to the TMS session , an image of the head of every subject was obtained using magnetic resonance imaging ( MRI ) using a neuronavigation system ( 3space Fastrack–Polhemus Isotrack II ) to ensure the accurate positioning of the TMS coil . At the beginning of the experiment , the 3D location of 200 points on the scalp was measured using an electromagnetic position sensor to co-register the MRI with the actual position of the subject’s head [46] . A plastic cap with grid marks spaced at 1 cm intervals was secured in position to serve as a reference for reproducible coil placement and external orientation . During the experiment , subjects remained awake , seated in a plastic comfortable chair with pillows placed under the forearms/hands , and his/her jewelry , glasses , watches and other potentially conducting or magnetic objects worn on the head or arms were removed to prevent interactions with the magnetic field , consistent with a TMS guide [39] . The skin surface over the forearm/hand muscles was washed , shaved and abraded with alcohol at 70% until an erythemic response appeared . To ensure consistent surface electromyography ( sEMG ) electrodes placement , the participant’s forearm/hand was measured with meter tape in each testing session . When possible , during all data collection , a manual muscle test was employed to isolate the target muscles , including the flexor digitorum superficialis ( FDS ) , abductor pollicis brevis ( APB ) , first dorsal interosseous ( FDI ) and abductor digiti minimi ( ADM ) , was conducted to determine the optimal placement of the electrodes [47 , 48] . Surface EMG recordings were obtained using surface 8 mm , Ag/AgCl electrodes ( Medtronic Adhesive Disposable Surface Electrodes ) placed on the skin over the muscle bellies , and the centers of the electrodes were placed approximately 1 . 5 cm apart . A ground electrode was placed ipsilaterally above the epicondylus , laterally at the elbow joint . Surface EMG signals were amplified and band pass filtered ( 1–5000 Hz , Biopac MP150 Systems Inc . ) . The signal was subsequently digitized at a sampling rate of 15 . 000 Hz ( A/D converter National Instruments—LABVIEW 7 . 0 ) and stored on a desktop computer for offline analysis . Custom-made MATLAB software ( 10—Mathworks , Inc . , Massachusetts , USA ) was used to measure the latency and peak-to-peak MEP amplitudes . The MEPs mean amplitudes recorded at each stimulated point was subsequently calculated and projected onto the brain to create a cortical muscle representation map . We visually inspected the EMG profiles to ensure that all muscles were electrically silent for each TMS pulse . When it was not the case , the trial was rejected and stimulation at that point was repeated . TMS mapping of four hand target muscles was performed using a MagVenture-MagPro R30 ( Tonica Elektronik A/S , Denmark ) connected to a figure-of-eight cooled coil ( wing diameter = 75 mm; peak magnetic field strength 2 . 2 T; peak electric field strength 660 V/m , biphasic pulse ) . The elicited electric field was directed in the lateral to medial direction with the coil held in a tangential position with the handle perpendicular to the midsagittal line . Before mapping , the FDS hot spot ( position on the scalp where FDS muscle responses could be reliably evoked with the lowest stimulator intensity and highest peak-to-peak amplitudes ) was located . Subsequently the resting motor threshold , which corresponds to the minimal intensity of stimulation at the hotspot eliciting MEPs larger than 50 μV in at least 50% of 10 trials , was determined according to the threshold-hunting paradigm [49] . Once these parameters were determined , the simultaneous mapping of four target muscles was performed , with the output intensity adjusted to 120% of the resting motor threshold . A total of 10 consecutive pulses were delivered at each stimulated site of the grid , beginning at the hotspot and moving in a spiral direction , with 3- to 5-second inter-pulse intervals . Mapping was complete when the locations adjacent to the active sites were identified as non-active/no MEP [21 , 38] . MEPs associated to muscle contractions were discarded . After data collection , the average latency and amplitude of the MEPs for each site was calculated . The mean amplitude obtained per coordinate was then normalized per participant by dividing the mean amplitudes by the maximum mean amplitudes collected in the hotspots . The center of gravity ( COG ) , map area ( number of active sites ) and muscle overlaps were measured from MEP normalized values . The COG was defined as the map location representing the amplitude-weighted center of the area of excitability [50] . The COG ( x , y ) coordinate was calculated as: COG=[∑aixi/∑ai , ∑aiyi/∑ai] Where , ai represents the mean amplitude , and xi and yi represents the stimulated coordinate position . The Euclidean equation was applied to determine the distance between COG locations in the same hemisphere , whereas the map area was defined as sum of active sites with >50 μV/MEP . The muscle overlap parameter was defined as the TMS stimulation points that generated simultaneous MEPs in all four-target muscles [51 , 50] . Descriptive and nonparametric analyses were performed using STATISTICA 7 . 0 ( StatSoft Inc . , Tulsa , USA ) and GraphPad Prism 6 ( GraphPad Software , Inc . , San Diego , USA ) . To study the handgrip ( dynamometry ) , TMS map area and COG we used a Wilcoxon matched-pairs signed rank test . The Mann-Whitney test was chosen to analyze the TMS motor threshold and MEP amplitudes . For all statistical tests , the alpha level was set to p<0 . 05 .
The ENMG exam was used to assess the bilateral ulnar/median nerve conduction in leprosy patients . Table 2 presents the results of the evaluation of the patients’ sensory and motor potentials . All tested patients presented a severe\complete impairment ( latency , amplitude and conduction velocity ) of both motor and sensitive fibers of the ulnar nerve ( innervating FDI and ADM muscles ) in one of the limbs , which was thus defined as “the most affected hand ( MAH ) ” . Furthermore , patients P1 , P2 , P4 and P6 showed a severe loss in the sensory component of the median nerve . Patient P1 also presented evidence of damage in the motor fibers of the median nerve . All the other patients presented normal values within the standardized ENMG thresholds for the median nerve . Indeed , the axonal injury caused by M . leprae courses with a distal to proximal pattern , with an initial ulnar nerve sensory impairment followed by injury in motor fibers as well , described in the literature as impairment of conduction of nerve impulse [52] and decreased amplitude of sensory-motor potentials [53] . If not diagnosed and treated in time the disease progresses towards affecting other nerves , such as the median . Firstly , we compared grip force in both upper-limbs of control subjects and found no significant difference ( Wilcoxon test , p = 0 . 156 ) . The average of the grip strength in both upper limbs of the control group ( standard value ) was then compared with that of patients . Lower grip strength was found for the MAH as compared to the control group ( p = 0 . 031 ) . Lower grip strength was also found in the MAH ( p = 0 . 031 ) as compared with the less affected hand ( LAH ) in the patient’s group . These results are shown in Fig 1 . The percentage of TMS machine output ( motor threshold ) , and the distance in time observed between the TMS stimulation artifact and the MEP onset ( latency ) were measured at the FDS muscle hotspot . Moreover , MEP average peak-to-peak ( amplitude ) was measured from the FDS , APB , FDI and ADM muscle hotspots . Table 3 shows individual values of these parameters obtained from patients and controls . We found a significant difference in motor threshold ( MT ) for the FDS muscle between the MAH and the control group ( p = 0 . 041 ) as well as between the LAH and the control group ( p = 0 . 019 , Fig 2A ) . When the MT of the hemisphere contralateral to the MAH was plotted as function of the ipsilateral hemisphere per subject , it was clear that the MT was lower in the contralateral than in the ipsilateral hemisphere for the patients , whereas this parameter was fairly balanced in the control group . Accordingly , such interhemispheric bias is absent in healthy volunteers [54] . Mann-Whitney test also showed significant differences for FDS amplitudes between the hemisphere contralateral to the MAH and the control group ( p = 0 . 026 ) , as well as for FDI amplitudes between the hemisphere contralateral to the MAH and those of the LAH ( p = 0 . 031 ) as well as between those of the hemisphere contralateral to the MAH and the control group ( p = 0 . 004 , Fig 3 ) . For each tested muscle , a map area was defined as the number of active sites with MEP average amplitudes equal or higher than 50 μV . In Patient ( P1 ) , diagnosed with serious right upper limb nerve damage , no MEP response could be elicited in the tested intrinsic muscles ( APB , FDI and ADM ) , even with 99% of the TMS machine output . Great variability in map area was observed for all other patients and control subjects . A Wilcoxon test was applied to compare the FDS , APB , FDI and ADM map areas between hemispheres in leprosy patients ( N = 4;P3 to P6; S1 Table ) . Despite the lack of significant differences , smaller motor representation areas were found for the FDI and ADM muscles , innervated by the ulnar nerve , in the hemisphere contralateral to the MAH , as illustrated in Fig 4 . In the same vein , no significant differences in overlap of motor representations were found for the four target muscles FDS , APB , FDI and ADM ( S2 Table ) . The position of the center of gravity ( COG ) of each muscle was compared within and between hemispheres . Differences in the absolute COGs between the hemispheres indicate whether these areas are equidistant from the midline , or whether any asymmetry exists . Moreover , the position of the COGs of each muscle in the same hemisphere might indicate either an overlap or a reorganization of the motor maps . The Wilcoxon test showed no significant differences between hemispheres neither in patients nor in paired controls . Likewise , no significant differences in distance between COGs were observed between the patients and control subjects for all pairs of muscles .
Measurement of handgrip strength has gained attention as a simple , non-invasive marker of muscle strength of upper extremities , suitable for clinical use . This assessment reflects the maximum strength derived from combined contraction of extrinsic and intrinsic hand muscles , leading to the flexion of hand joints [55] . Digital or analogical dynamometry has been recommended as an additional method to assess peripheral nerve function in leprosy , particularly in early ulnar impairment [41] . Patients with chronic ulnar/median nerve impairment were herein shown to exhibit a significant decrease in grip strength in the MAH when compared to the LAH as well as to control participants . Accordingly , Rajkumar , Premkumar and Richard [56] observed a high correlation between grip strength and daily life activity in 62 leprosy patients , where patients showing poor results in triple pinch strength could experience more difficulties in daily life activities . Likewise , in the majority of patients in the present study and as attested by the electroneuromyographic evaluation , the ulnar was the most affected nerve , causing weakness in hypothenar muscles and important dysfunctions in the first , fourth and fifth fingers . These effects were clearly most evident for the MAH . Target muscles ( FDS , APB , FDI and ADM ) were mapped employing the motor threshold of the FDS muscle at rest ( MT ) . This choice was based on the fact that the median nerve ( which supplies the FDS muscle ) exhibited control-like ENMG parameters in our cohort ( except for P1 , see Table 2 ) . The FDS muscle MT was globally higher in leprosy patients than in matched controls , with an average difference of 10 . 4% for the hemisphere contralateral to the MAH and 20 . 3% for that contralateral to the LAH . Altered MT could result from functional changes at any level of the motor output pathway , thus reflecting changes in the excitability level of neuronal elements within the corticospinal pathway , that is , cortical inhibitory or excitatory interneurons , corticospinal neurons as well as motor units [57 , 58 , 59] . Motor neuropathy derived from leprosy is known to impair hand functionality , resulting in grip strength decrease [56] and reduced nerve conduction [45] . These peripheral motor impairments are often accompanied by severe muscle atrophy [60 , 61 , 62] . Thus , one could suppose that the higher MT values found in leprosy patients as compared to the control group might reflect enhanced conduction resistance in the motor output pathway . Besides , changes in the sensory input of leprosy patients with severe sensory loss as abnormalities in the peripheral afferent inputs or in their central processing may interfere with motor output [63] . Indeed , several studies have shown that sensory deprivation resulting from dorsal root or dorsal column transections , skin anesthesia , peripheral neuropathy or inactivation of the somatosensory cortex in humans and non-human primates affect motor behavior [64 , 65 , 66 , 67 , 68 , 69 , 70 , 71 , 72] and result in changes in motor cortical representation . Therefore , the complete sensory loss experienced by the leprosy patients , as observed in the ENMG exam , may also be responsible for higher MT . Altogether , these factors may have led to higher MT herein found for the FDS muscle of leprosy patients as compared to control subjects . Patients had lower resting motor threshold ( MT ) for the FDS muscle in the hemisphere contralateral to MAH as compared to the hemisphere contralateral to the LAH . Accordingly , previous studies in humans suffering traumatic amputation of the upper limb showed that the MT of the amputated limb is lower ( average 10–17% ) in the hemisphere contralateral to the lost limb than that of the ipsilateral hemisphere [29 , 30 , 73 , 74 , 75 , 76] . The same holds for individuals who remained immobilized after a fracture of the upper limb [36] . MEP values collected in FDS hotspot in the hemisphere contralateral to the MAH were also higher than those of control subjects , consistent with the observations of Zanette et al , [36] . Higher MEP values and lower MT found in the FDS muscle in the MAH as compared to the LAH indicate that altered handgrip function induced by leprosy can result in a pronounced and long-term reorganization in M1 . These results are reinforced by the fact that the median nerve , which exhibited control-like ENMG motor parameters in our cohort , supplies the FDS muscle , from which MT values were collected . Such modifications should thus be rather due to cortical reorganization than to peripheral conduction change induced by leprosy . Smaller MEP amplitudes were found for the FDI muscle of the MAH as compared to the LAH and the control group . Although the map area of the tested intrinsic hand muscles was highly variable both within patients and in control subjects , smaller map areas were also found for the FDI and ADM muscles in the motor cortex contralateral to the MAH as compared to the LAH . These results are in line with those found herein for ENMG and grip strength , and furthers results obtained in hand allograft suggesting that the extent of intrinsic hand muscle representation in M1 associate with hand function [34] . During whole handgrip , the extrinsic muscles provide the major gripping force [77] . The FDS muscle , specifically , seems to be called upon in direct proportion to the required force . Besides , the major intrinsic muscles of whole handgrip are the interosseous , used as phalangeal rotators and metacarpophalangeal flexors [77] . If the function of the interosseous muscles , supplied mostly by the ulnar nerve , is affected by leprosy , then one could suppose that the FDS would take over the handgrip force . Thus , decrease in handgrip force might bear a correlate to the changes in corticospinal excitability shown for FDS ( an extrinsic and healthy forearm muscle , being possibly overused due to the chronic dysfunction ) and for FDI ( an important intrinsic hand muscle affect by ulnar infection ) . In conclusion , the decrease or loss of sensory afferent neurons and/or an impairment in the strength of peripheral muscles in the ulnar/median territory verified in leprosy patients radically alters the handgrip function leading to cortical motor reorganization in the corresponding affected hand muscles . Leprosy patients usually exhibit a mixed variety of peripheral nerve injuries with sensorimotor impairment , thereby increasing the cortical plasticity challenge and the variability of results ( for a discussion , see Reddy et al [20] ) . Individual factors might however contribute to the consequences of nerve damage . Future studies are needed to fully understand the plastic reorganization in leprosy as well verify cortical motor reorganization after repair procedures .
|
Leprosy is an endemic disease caused by Mycobacterium leprae that predominantly attacks both skin and peripheral nerves , resulting in persistent distal hand atrophy and the loss of sensory and autonomic functions . In this study , we employ a noninvasive tool named transcranial magnetic stimulation ( TMS ) to map the handgrip muscle representation in the primary motor cortex of patients affected by leprosy . The findings of this study support that the decrease or loss of sensory afferent neurons and/or impairment in the strength of peripheral muscles in the ulnar/median territory verified in leprosy patients alters the handgrip function leading to cortical motor reorganization in the corresponding affected hand muscles .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Primary Motor Cortex Representation of Handgrip Muscles in Patients with Leprosy
|
The South-to-North Water Diversion ( SNWD ) project is designed to channel fresh water from the Yangtze River north to more industrialized parts of China . An important question is whether future climate change and dispersal via the SNWD may synergistically favor a northward expansion of species involved in hosting and transmitting schistosomiasis in China , specifically the intermediate host , Oncomelania hupensis . In this study , climate spaces occupied by the four subspecies of O . hupensis ( O . h . hupensis , O . h . robertsoni , O . h . guangxiensis and O . h . tangi ) were estimated , and niche conservatism tested among each pair of subspecies . Fine-tuned Maxent ( fMaxent ) and ensemble models were used to anticipate potential distributions of O . hupensis under future climate change scenarios . We were largely unable to reject the null hypothesis that climatic niches are conserved among the four subspecies , so factors other than climate appear to account for the divergence of O . hupensis populations across mainland China . Both model approaches indicated increased suitability and range expansion in O . h . hupensis in the future; an eastward and northward shift in O . h . robertsioni and O . h . guangxiensis , respectively; and relative distributional stability in O . h . gangi . The southern parts of the Central Route of SNWD will coincide with suitable areas for O . h . hupensis in 2050–2060; its suitable areas will also expand northward along the southern parts of the Eastern Route by 2080–2090 . Our results call for rigorous monitoring and surveillance of schistosomiasis along the southern Central Route and Eastern Route of the SNWD in a future , warmer China .
Schistosomiasis is a neglected tropical disease that is known to have affected people in China for more than 2100 years , with presently ~800 , 000 infected and ~65 million people at risk of infection [1] . The challenge of combatting this disease lies in the wide distribution of its snail hosts and the broad range of domestic and wild mammals that act as reservoirs for human infections [2] . Chinese schistosomiasis is caused by the digenetic blood trematode Schistosoma japonicum , a parasitic flatworm that completes its life cycle through one intermediate ( i . e . the snail Oncomelania hupensis ) and diverse definitive ( i . e . mammals ) hosts . Over the past five decades , China has made remarkable progress in reducing S . japonicum infections in humans through a combination of chemotherapy and snail control , but schistosomiasis has re-emerged in recent years owing to changes in ecological and socio-economic factors , together with termination of the World Bank Loan Project on schistosomiasis control in 2001 [3] . Given that schistosomiasis is unlikely to be eliminated , considering whether and how future climates are likely to impact its transmission becomes increasingly important . Based on the environmental variables that associated with species’ occurrence records , ecological niche modeling ( ENM ) seeks to characterize environmental conditions suitable ( i . e . realized niche ) for a particular species and then identify where suitable environmental habitats are distributed in the space [4] , it is a powerful tool in studies of effects of global climate change on the geography of disease transmission [5] . Assumptions under which ENMs work best include equilibrium between species’ distributions and their ecological requirements , and conservatism of ecological niche [4] . Among them , niche conservatism providing support for using ENMs has been widely noticed , the degree to which plants and animals retain their ancestral ecological traits and environmental distributions ( 'niche conservatism' ) is hotly debated , in part because of its relevance to the fate of modern species facing climate change [6] . ENM tools , however , are also subject to issues including the need to balance goodness-of-fit against model complexity [7] , and the importance of considering uncertainty in model predictions [8] . These issues are particularly critical in studies involving transfer of models across space or time ( e . g . climate change effects ) . Recent efforts have developed methods to reduce model complexity and characterize uncertainty , and thereby improve model transferability in forecasting climate change effects [9–12] . These steps include species-specific tuning of settings ( rather than default setting ) to improve model performance [9 , 10] , evaluation using spatially independent training and testing data sets [12] , and integrating multiple predictions via ensemble approaches [11 , 12] . Oncomelania hupensis is the sole intermediate snail host of S . japonicum in China , which thus depends entirely on this snail species for transmission [13] . However , the taxonomy of O . hupensis in mainland China has been debated in view of marked morphological variation . Liu et al . recognized 5 subspecies [14] , whereas Davis et al . treated only 3 subspecies based on shell form , allozyme data , and biogeography [15] . However , Zhou et al . separated O . h . guangxiensis from O . h . hupensis based on molecular characters , and recognized 4 subspecies in mainland China [16] , which was later verified by Li et al . based on internal transcribed spacer ( ITS ) and 16S fragments [17 , 18] . Here , we consider the four subspecies [14 , 16 , 18 , 19]; at present , O . h . hupensis and O . h . robertsoni dominate transmission of S . japonicum , as control measures have reduced O . h . guangxiensis and O . h . tangi considerably [13] . These four subspecies differ in shell size and structure , breeding environment , growth rates , population genetics , and potential for infection by S . japonicum [17] . Previous attempts to predict spatial dimensions of transmission risk of schistosomiasis have characterized transmission environments of S . japonicum [20–22] or ecological requirements of O . hupensi [23 , 24]; these studies were generally conducted at local geographic scales and with limited temporal coverage . Several environmental correlates of S . japonicum transmission have been identified , including distance to snail habitat and wetlands , seasonal land surface temperature , and seasonal variation of vegetation indices [21 , 22] . Climate conditions explain much variation in transmission of schistosomiasis , especially at regional and continental scales [25 , 26] . Understanding ecological dimensions and potential distribution of O . hupensis is thus crucial [20] , and yet has not seen detailed analysis . The South-to-North Water Diversion ( SNWD ) project is a multi-decade mega-project in China . It is the biggest inter-basin transfer scheme in the world , aiming to channel 25 × 109 m3 fresh water annually from the Yangtze River in southern China to the more arid and industrialized north via two routes ( i . e . the Central Route and Eastern Route , Fig 1 ) . In the context of climate change , in which the geographic potential of O . hupensis may change , the relationship of such changes to planned SNWD corridors remains unknown . Surveillance sites were established during 2002–2010 across mainland China ( Fig 1 ) ; however , most sites were located along the Yangtze River at low elevations , focused on transmission by O . h . hupensis and O . h . robertsoni . The questions of whether future climate change and the SNWD project may synergistically favor expansion of some population of O . hupensis , and whether the existing surveillance sites are sufficient , necessitate the present study . In this study , we used a unique dataset of O . hupensis presences from more than 5 thousand villages to explore ecological dimensions and potential distributions of O . hupensis in mainland China . The aims of this study were to ( 1 ) compare climate spaces occupied by the four subspecies of O . hupensis , to ( 2 ) test whether climate niches were conserved during the four subspecies’ divergence ( i . e . climate niche conservatism evaluation ) , to ( 3 ) predict their potential distributions using state-of-the-art modelling techniques , to ( 4 ) investigate the potential impacts of future climate change and the SNWD project on O . hupensis . The overall purpose was to predict the S . japonicum transmission risk at present and under climate change in mainland China .
Occurrence data for subspecies of O . hupensis were assembled from Qian [27] . This national surveillance effort of schistosomiasis was carried out at the village level between the 1950s and 1980s across 12 Chinese provinces . In all , 5029 towns and villages reported presence of O . hupensis [27] . Rather than using centroids of infested counties , which reduces precision , we georeferenced individual villages using Google Maps . These points varied in terms of clumping , so we subsampled them to reduce sampling bias and spatial autocorrelation [28] , as follows . First , we arranged infested provinces according to sample density ( i . e . number of occurrence points divided by area of the province ) . The median served as the standard sampling effort , and all provinces presenting densities above that value were subsampled randomly to a lower density . In the end , we had 1996 occurrence points: 1402 O . h . hupensis , 470 O . h . robertsoni , 64 O . h . guangxiensis , and 60 O . h . tangi ( S1 File ) . Several approaches have been used to select environmental datasets for ecological niche modeling; the best environmental datasets would be ecological relevant to species in question [29] . At regional and continental scales , climatic factors have excellent predictive power in determining risk associated with disease transmission ( e . g . schistosomiasis [25 , 26] , West Nile virus [30] ) . Hence , we used subsets of the 19 bioclimatic variables developed by Hijmans et al . [31] , chosen as follow . First , variables that combined temperature and precipitation ( i . e . mean temperature of wettest quarter , mean temperature of driest quarter , precipitation of warmest quarter , precipitation of coldest quarter ) were excluded because they display artificial discontinuities between adjacent grid cells in some areas [32] . The importance of each of the remaining 15 variables was assessed by a jackknife analysis of variable importance in Maxent ( [33] , see below for Maxent detail ) , and unimportant variables were discarded . Highly correlated variables were then removed in SDMtoolbox , a python-based GIS toolkit for spatial analysis [34] . Eight variables ( S1 Table ) that showed ecological relevance ( regularized training gain >0 . 14 ) and low correlation with other variables ( Pearson correlation <0 . 9 ) were chosen in the end . All variables were analyzed at a spatial resolution of 2 . 5 minute . Climatic spaces occupied by the four subspecies were first compared along each environmental dimensions using violin plots , which combine the functions of boxplot and kernel density , providing a better indication of the shape of the data distribution . We used NicheA , a toolkit to create and visualized ecological niches in environmental spaces [35] , to visualize climate niches occupied by each subspecies in reduced multiple environmental spaces: we displayed the first three principal components derived from the 8 bioclimatic layers , and plotted minimum volume ellipsoids ( MVEs ) around occupied conditions . We quantified niche overlap between pairs of subspecies using Schoener’s D [36]; this metric ranges from 0 ( no overlap ) to 1 ( complete overlap ) , and was used to test niche identity and niche similarity between subspecies . Niche identity and similarity tests were performed to determine whether climate spaces occupied by the two subspecies were identical or exhibited significant difference , and whether these differences were caused by the environmental feature spaces [37] . Niche identity was tested by randomly allocating occurrence records within each pair 500 times , according to observed numbers of records , and comparing observed and simulated Schoener’s D estimates . In contrast , niche similarity was tested by shifting the centroid of the observed occurrence densities to a random location within the available environmental space 500 times , and comparing observed with the null distribution of simulated estimates of Schoener’s D [37]; climate variables measured at locations across the available backgrounds of subspecies were combined and projected onto the first two principal components using PCA_env package [37] . Smoothed densities of occurrences and available environments in each grid cell were calculated and compared among the four subspecies [37] . Background environments for climate niche comparisons and niche model calibration should include only areas that have been accessible to the populations under study [38] . We delimited this area by buffering a convex hull around known occurrences by 200 km ( Fig 1 ) in SDMtoolbox [34] . This approach reflects a compromise between including all environments that have been accessible to the species , and still covering a broad-enough extent to minimize extrapolation and detect climatic differences between presence and background records [39] . To forecast climate change effects , we used fMaxent ( fine-tuned Maxent , see below ) and ensemble approaches [11 , 12] to calibrate models under present conditions , which were then transferred onto climate conditions for 2050 and 2080 . Maxent is the most commonly used method in ENM , and it can fit arbitrarily complex models to explain relationships between environmental variables and occurrence data ( version 3 . 3 . 3k; [40] ) . However , because an excessively complex model will be extremely specific to input data and perform poorly when extrapolating , Warren and Seifert proposed using a sample-size-adjusted Akaike information criterion ( AICc ) as criteria with which to address overfitting; this approach does not control model fit directly , but rather uses AICc to choose appropriate settings [7] . We used the “ENMeval” package [9] to fine-tune Maxent models by seeking the minimum value of AICc among candidate models . ENMeval provides an automated way to execute Maxent models across a user-specified range of regularization multiplier ( RM ) values and features combinations ( FC ) . We set the RM range to 0 . 5–6 . 0 with increments of 0 . 5 , and used 6 FCs , to cover a broad range of model settings . The block method was used to partition occurrence data into four bins , 3 of which were used for training and the remaining one for testing ( bin combination , BC ) , which is desirable for studies involving model transferring [9] . In all , 2160 models ( 12 RMs × 6 FCs× 6 BCs × 5 occurrence groups ) were generated for the four subspecies and for O . hupensis as a whole . Ensemble models are used commonly in forecasting climate change effects , seeking to generate a consensus estimate that reduces individual model uncertainty by reflecting the central tendency of multiple models [11 , 12 , 41] . Here , outputs from six modelling algorithms , including generalized additive models ( GAM ) , generalized boosted models ( GBM ) , generalized linear models ( GLM ) , random forests ( RF ) , genetic algorithms ( GARP ) , and the fMaxent model described above were included in ensembles . Individual GAM , GBM , GLM , and RF models were developed using BIOMOD2 [42] , as implemented in R [43]; GARP models were developed in desktopGARP [44] . Details of implementation of each algorithm are provided in the supporting information ( S2 Table ) . Model ensembles typically use a weighted averaging approach , in which models are weighted according to their interpolative performance ( e . g . [30] ) . However , a recent assessment pointed to the challenge of balancing model interpolative accuracy against transferability [29 , 45] . Therefore , rather than using weighted averages , we used the PCA ( median ) method to identify the “central tendency” of individual model predictions [8 , 11] . The PCA measures , for each model , its ability to follow the general trend of predictions of the six models . This method calculates the median of the four individual models that had higher factor values among the six models [8 , 11 , 12] . The occurrence data used to fit the niche models were split randomly into two datasets , for calibration ( 70% of points ) and interpolation evaluation ( 30% of points ) . Performance of individual and consensus models was evaluated via a partial ROC ( receiver operating characteristic ) approach [46] . Comparing to traditional AUC ( area under the ROC curve ) , which was criticized because present data are more reliable than absence data in model evaluation [46] , the partial ROC approach takes the quality of occurrence points into account and weights more on omission error [46] ) . Here , AUC calculations were limited to ROC spaces over which models actually made predictions , and only omission errors <5% were considered ( i . e . E = 5%; [46] ) . Final model runs incorporating all point data were used for visualizations and risk assessments . A modified least training presence threshold based on E = 5% was applied to fMaxent model predictions for O . hupensis and O . h . hupensis to generate binary predictions . We did not generate threshold predictions in ensemble future projection because such predictions are not applicable and hard to interpret ( i . e . individual models for generating consensus models were different in present and future predictions ) . Future climate variables were downloaded from WorldClim [31] , the Consultative Group on International Agricultural Research ( CGIAR ) , and the research program on Climate Change , Agriculture and Food Security ( CCAFS ) . To reduce uncertainty regarding future climate conditions ( S1 Fig ) , rather than using the 13 original global climate models ( GCMs , S3 Table ) from the IPCC 5th Assessment , the PCA ( median ) protocol was also used to generate consensus “climate models” among the 13 GCMs for each climate dimensions for 2050–2060 and 2080–2090 ( S4 Table ) . The fMaxent and ensemble models based on present predictions were applied to these future conditions . Future climate models applied to the intermediate scenario of representative concentration pathways of 4 . 5 ( i . e . “RCP45”; [47] ) in which future anthropogenic greenhouse gas emissions were estimated to peak around 2040 . This scenario was chosen because it represents the middle range of available four scenarios , and as such is considered more realistic than models based on extremely high or extremely conservative scenarios [47] . Climatic similarities between present and future in 2050 and 2080 were assessed using mobility-oriented parity ( MOP ) metrics , a correction and simplification of multivariate environmental similarity surfaces [39] .
Different degrees of overlap were observed in the eight climate dimensions among the four subspecies ( S2 Fig ) . Oncomelania hupensis hupensis and O . h . robertsoni occupied similar temperature and precipitation dimensions in terms of annual mean temperature ( bio1 ) , mean diurnal temperature range ( bio2 ) , and annual precipitation ( bio12 ) , but not isothermality ( bio3 ) , temperature seasonality ( bio4 ) , mean temperature of warmest quarter ( bio10 ) , or precipitation of driest month ( bio14 ) ; O . h . guangxiensis and O . h . tangi occupied similar temperature and precipitation regimes in terms of annual mean temperature ( bio1 ) , temperature seasonality ( bio4 ) , mean temperature of warmest quarter ( bio10 ) , annual precipitation ( bio12 ) , and precipitation of driest month ( bio14 ) , but not mean diurnal temperature range ( bio2 ) or isothermality ( bio3 ) ( S2 Fig ) . The four subspecies showed diverse responses to precipitation seasonality ( bio15 ) . Minimum volume ellipsoids occupied by the subspecies overlapped broadly ( Fig 2 ) . The size of the MVEs corresponded roughly to the geographic range extent of each subspecies ( Figs 1 and 2 ) , with O . h . hupensis and O . h . robertsoni occupying larger volumes than O . h . tangi and O . h . guangxiensis . Niche overlaps between pairs of subspecies also corresponded roughly to their genetic distances estimated by 16S sequence ( Fig 2 and S3 Fig ) : i . e . the close relationship between O . h . hupensis and O . h . tangi coincided with the highest climatic niche overlap ( D = 0 . 215 ) among all pairs ( S3 Fig ) . Similar patterns were observed between O . h . tangi and O . h . guangxiensis ( D = 0 . 147 ) , but to a lesser extent ( Fig 2 and S3 Fig ) . The null hypothesis of niche identity was rejected in all pairwise comparisons ( S3 Fig ) . However , in analyses of niche similarity , the null hypothesis could not be rejected , except for O . h . robertsoni versus O . h . tangi ( S3 Fig ) . Results of niche identity and similarity thus suggest that , although the four subspecies occupy unique climate spaces , the nonequivalence of niche spaces derives from a background effect , and not from biological differences . Individual model performances varied across model algorithms and subspecies in interpolation validations ( Fig 3 ) . The machine learning methods ( i . e . fMaxent , GBM , RF ) generally showed better discriminant ability than regression models ( i . e . GAM , GLM ) ; GARP showed unstable performance ( Fig 3 ) . Similar to the machine learning models , consensus models showed good discriminant ability for the individual subspecies and for O . hupensis as a whole ( Fig 3 ) . Using all of the occurrence data , parameters of AICc-selected models ( i . e . fMaxent ) differed from default settings ( S4 Fig ) . Based on block partitions of occurrence data , mean AUCtest values of fMaxent models were 0 . 79 , 0 . 79 , 0 . 80 , 0 . 86 , and 0 . 94 for O . hupensis ( as a whole ) , O . h . guangxiensis , O . h . hupensis , O . h . robertsoni and O . h . tangi , respectively , with fMaxent models of O . h . hupensis ( AUCdiff = 0 . 09 ) and O . h . tangi ( AUCdiff = 0 . 02 ) showing less overfitting than the other three ( S4 Fig ) . In species-wide consensus models , the first principal component explained 46 . 7–72 . 2% of individual model variation ( Table 1 ) . Consensus models were discriminated by the first axis of the PCA , and each individual model was selected in consensus model processing ( Table 1 ) . Variation among individual model predictions spatially was observed in both present and future ( 2050 and 2080; S5 Fig ) . Some areas identified as suitable nonetheless corresponded to environments beyond the climate envelope of the calibration area at present , thus involving non-analog climate conditions ( S6 Fig ) . For example , fMaxent identified disjunct suitable areas around Beijing in northern China ( Fig 4 ) , but these areas involved model transfer into novel climate conditions ( S6 Fig ) , making their interpretation uncertain and unwise . Within the distribution of each subspecies ( Fig 1 ) , projection of present ENMs onto future climate datasets generally involved little extrapolation ( MOP metrics; S6 Fig ) . Transferring present-day models onto future climate scenarios , fMaxent models were more conservative than ensemble models ( Figs 4 and 5 ) : the western part of the predicted distribution based on consensus models was cleaner than predictions based on fMaxent , and the consensus method did not make the isolated predictions in the fMaxent model ( Fig 4 ) . Both fMaxent and consensus approaches identified a pattern of range expansion and suitability increase in O . h . hupensis ( Figs 4 and 5 ) . In O . h . robertsoni , both models identified an eastward shift , whereas in O . h . guangxiensis , a northward shift was indicated ( Figs 4 and 5 ) . In O . h . tangi , the two models showed contrasting predictions ( Figs 4 and 5 ) . Binary predictions were based on fMaxent outputs , as thresholding future predictions from ensembles is difficult . Overlapping the Central Route and Eastern Route of SNWD with the binary future predictions for O . hupensis and O . h . hupensis , the southern Central Route coincides with suitable areas for O . h . hupensis in 2050–2060 , and its suitable areas will expand northward along the southern Eastern Route by 2080–2090 ( Fig 6 ) . All of these areas are beyond the reach of present surveillance sites for schistosomiasis monitoring . Because a northward expansion of O . hupensis may occur considering future climate warming , these potential expansion areas need to be better covered by future surveillance efforts . Future surveillance efforts should also consider potential re-emergence of O . h . guangxiensis and O . h . tangi , as some areas of increasing suitability were noted for these two subspecies as well ( Fig 5 ) , although present intervention efforts have brought the snails to near extinction .
Limitations on materials and methodologies employed in this study need to be addressed here . While this paper focused on climate drivers , these factors occur in a complex milieu of other non-climatic drivers of snail distribution and parasite endemicity [21 , 22] , although the non-climatic factors usually functioned at a small scale . Although we adopted ensemble forecasting approach to minimize the uncertainty of individual model predictions , the uncertainty exists in consensus models [29] . Ecological niche conservatism is of increasing importance given the complex impacts of ongoing climate change on biodiversity [4 , 6] . Many studies have evaluated niche conservatism across diverse evolutionary time spans [4 , 6] . Future projections for species involved in disease transmission and likely to respond to climate change are usually fraught with uncertainties and complexities; however , these assessments are crucial in identifying appropriate mitigation strategies [26] . Here , we tested climatic niche conservatism among the four subspecies of O . hupensis across mainland China , and integrated state-of-the-art modelling techniques ( fMaxent and ensemble models ) to forecast climate change effects . Our results have important implications regarding genetic divergence of O . hupensis and likely climate change effects on schistosomiasis transmission in mainland China . The ecological niches of the four subspecies of O . hupensis were not identical , but we were unable to reject the null hypothesis that climatic niches are similar ( except O . h . robertsoni versus O . h . tangi ) . Although failure to reject the null hypothesis does not assure that the climatic niche has been conserved , no evidence indicates that they have not been conserved , and broad climate spaces overlapped among the four subspecies ( Fig 2 and S2 Fig ) . The relationship between niche overlap and phylogenetic relationships of the four subspecies further supports the idea that climate niches have been conserved ( Fig 2 and S3 Fig ) . The signal of climate niche conservatism suggests that factors other than climate likely account for the genetic divergence of O . hupensis populations . Li et al . suggested that genetic differentiation of O . hupensis in mainland China is ultimately structured by landscape ecology [18] , with populations falling into four different ecological settings ( Fig 1 ) : swamps and lakes in the Yangtze River Basin ( O . h . hupensis ) ; the mountainous region of Sichuan and Yunnan Provinces ( O . h . robertsoni ) ; the hilly , littoral part of Fujian province ( O . h . tangi ) ; and the karst landscape of Guangxi Autonomous Region ( O . h . guangxiensis ) . This landscape-level segmentation of the four subspecies is generally consistent with the foundational work of Liu et al . [14]: indeed , clear geographic barriers separate the four subspecies ( Fig 1; [14 , 16] ) . Climate niche divergence between O . h . robertsoni and O . h . tangi might relate to the long geographic distance separating them . Previous studies have found that long-term climate warming tends to favor geographic expansion of S . japonicum in mainland China , but most such risk assessments have relied solely on mechanistic approaches ( e . g . [23 , 25 , 48] ) . Although mechanistic models may be more desirable in that they estimate dimensions of the fundamental niche and in that they avoid problems with extrapolation [49] , correlative ENMs have practical advantages in terms simplicity and flexibility , particularly as regards parameterization [50] . Comparing with mechanistic models , which predict a broad northward and westward expansion of S . japonica [23 , 25 , 48] , correlative ENMs suggest a similar pattern , but with more detailed spatial predictions . Increased suitability and range expansion were observed consistently in O . h . hupensis , eastward and northward shifts in O . h . robertsoni and O . h . guangxiensis , and relatively stability status in O . h . gangi were observed in all our future model predictions ( Figs 4 and 5 ) . Most current surveillance sites are distributed along the Yangtze River , designated to monitor transmission by O . h . hupensis and O . h . robertsoni . However , in a climate change context , both of these subspecies are expected to expand or shift distributionally ( Fig 5 ) . Surveillance sites distribution will have to broaden in coverage to be able to detect these shifts . In addition , the potential of O . h . guangxiensis and O . h . tangi to re-remerge should also be considered , as sites presenting increased suitability were identified ( Fig 5 ) . The southern parts of the Central Route of South-to-North Water Diversion ( SNWD ) project will become suitable for O . h . hupensis in 2050–2060 , and suitable areas will expand northward along the southern parts of the Eastern Route of SNWD by 2080–2090: these areas are not covered by present surveillance efforts ( Fig 6 ) . Our results call for more rigorous monitoring and surveillance of schistosomiasis in the northern of potential expansion areas , although schistosomiasis currently has not been detected along either the southern Central Route or the Eastern Route; nonetheless , range expansion may open potential for emergence [48 , 51] .
|
The South-to-North Water Diversion ( SNWD ) project is designed to channel fresh water from the Yangtze River north to more industrialized parts of China . An important question is whether future climate change and dispersal via the SNWD may synergistically favor northward expansion of schistosomiasis in China . Our models indicated increased suitability and range expansion in Oncomelania h . hupensis in the future; an eastward and northward shift in O . h . robertsioni and O . h . guangxiensis , respectively; and relative stability in O . h . gangi . The southern Central Route of SNWD will coincide with suitable areas for O . h . hupensis in 2050–2060; its suitable areas will also expand northward along the southern Eastern Route in 2080–2090 . Our results call for rigorous monitoring and surveillance of schistosomiasis along the southern Central Route and Eastern Route of the SNWD in a future , warmer China .
|
[
"Abstract",
"Introduction",
"Methodology",
"Results",
"Discussion"
] |
[
"schistosoma",
"invertebrates",
"ecology",
"and",
"environmental",
"sciences",
"medicine",
"and",
"health",
"sciences",
"ecological",
"niches",
"helminths",
"china",
"atmospheric",
"science",
"geographical",
"locations",
"tropical",
"diseases",
"parasitic",
"diseases",
"animals",
"oncomelania",
"gastropods",
"neglected",
"tropical",
"diseases",
"snails",
"climate",
"change",
"schistosoma",
"japonicum",
"molluscs",
"people",
"and",
"places",
"helminth",
"infections",
"schistosomiasis",
"eukaryota",
"climatology",
"asia",
"ecology",
"earth",
"sciences",
"climate",
"modeling",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2017
|
Schistosoma japonicum transmission risk maps at present and under climate change in mainland China
|
After more than a decade of steadily declining notifications , the number of reported cholera cases has recently increased in Vietnam . We conducted a matched case-control study to investigate transmission of cholera during an outbreak in Ben Tre , southern Vietnam , and to explore the associated risk factors . Sixty of 71 diarrheal patients confirmed to be infected with cholera by culture and diagnosed between May 9 and August 3 , 2010 in Ben Tre were consecutively recruited as case-patients . Case-patients were matched 1:4 to controls by commune , sex , and 5-year age group . Risk factors for cholera were examined by multivariable conditional logistic regression . In addition , environmental samples from villages containing case-patients were taken to identify contamination of food and water sources . The regression indicated that drinking iced tea ( adjusted odds ratio ( aOR ) = 8 . 40 , 95% confidence interval ( CI ) : 1 . 84–39 . 25 ) , not always boiling drinking water ( aOR = 2 . 62 , 95% CI: 1 . 03–6 . 67 ) , having the main source of water for use being close to a toilet ( aOR = 4 . 36 , 95% CI: 1 . 37–13 . 88 ) , living with people who had acute diarrhea ( aOR = 13 . 72 , 95% CI: 2 . 77–67 . 97 ) , and little or no education ( aOR = 4 . 89 , 95% CI: 1 . 18–20 . 19 ) were significantly associated with increased risk of cholera . In contrast , drinking stored rainwater ( aOR = 0 . 17 , 95% CI: 0 . 04–0 . 63 ) , eating cooked seafood ( aOR = 0 . 27 , 95% CI: 0 . 10–0 . 73 ) , and eating steamed vegetables ( aOR = 0 . 22 , 95% CI: 0 . 07–0 . 70 ) were protective against cholera . Vibrio cholerae O1 Ogawa carrying ctxA was found in two of twenty-five river water samples and one of six wastewater samples . The magnitude of the cholera outbreak in Ben Tre was lower than in other similar settings . This investigation identified several risk factors and underscored the importance of continued responses targeting cholera prevention in southern Vietnam . The association between drinking iced tea and cholera and the spread of V . cholerae O1 , altered El Tor strains warrant further research . These findings might be affected by a number of limitations due to the inability to capture asymptomatic or mildly symptomatic infections , the possible underreporting of personal unhygienic behaviors , and the purposive selection of environmental samples .
Cholera is a highly contagious diarrheal disease , caused by infection of the Gram-negative bacterium Vibrio cholerae [1] . Areas with poverty , high population-density , poor sanitation , poor education levels , and lack of potable water are at risk for cholera outbreaks [2–4] . An estimated 1 . 3–4 . 0 million illnesses and 21 , 000–143 , 000 deaths are attributed directly to the disease , which is predominantly seen in Sub-Saharan Africa , South-East Asia , and the Americas ( i . e . , Haiti ) [5] . Consumption of contaminated water is thought to be the main mode of transmission [6–8] . Prior to implementation of control measures , Vietnam suffered a disproportionate burden of cholera . For example , between 1979 and 1996 , there were 56 , 050 reported cases of cholera and 1 , 272 deaths due to cholera in this country [9] . Oral cholera vaccination programs were introduced in highly endemic areas through a national expanded program of immunization in 1997 [10] . The program , in conjunction with improved personal hygiene and access to potable drinking water associated with both health promotion programs and economic growth , led to a substantial drop in the number of notified cholera cases [6 , 10] . However , these gains have not been sustained . Resource constraints [10] as well as the modest vaccine effectiveness , estimated to be 76% during the outbreak and 50% after 3–5 years [11 , 12] , have jeopardized the long-term impact of this program . There were a total of 3 , 646 cholera cases reported between 2006 and 2010 in Vietnam , which was nearly five times higher than the number of notification made during 2001–2005 ( 747 cases ) [13] . Of concern , from both a clinical and public health perspective , is the emergence and transmission of V . cholerae O1 , El Tor strains that produce the classical cholera toxin [14] and the circulation of strains genetically resistant to certain antibiotics [9 , 15] . Outbreaks in Hanoi and neighboring regions in northern Vietnam in 2007–2008 have been linked to contaminated food , especially dog meat , raw vegetables , and raw pig/duck blood [11 , 13] . However , the pathways for the transmission of cholera in other parts of the country have not yet been investigated . In this article , we report a cholera outbreak in the Mekong Delta province of Ben Tre in 2010 . We conducted a matched case-control study to identify risk factors for cholera infection , and examined environmental samples to identify contaminated food and water sources . These data are useful for informing targeted and effective responses to cholera outbreaks in Vietnam .
Ben Tre province is located in the Mekong Delta region of southern Vietnam . It has a population of 1 . 3 million people and a largely agricultural economy . This coastal province is about 1 . 25 meters above sea level and is almost surrounded by water ( Fig 1 ) . Due to low piped-water coverage , river water is generally used as one of the main sources of water among the population . The last mass oral cholera vaccine program was implemented in 2002 , and no cholera cases had been reported since 2005 [10 , 16] . However , in 2010 , an outbreak of cholera occurred in the province , starting in Mo Cay town on May 9 . A total of 71 cases were identified in the outbreak ( Fig 1 ) . We employed a purposive sampling technique to collect water samples from several types of water sources in Mo Cay and Giong Trom towns , where the first cholera cases were reported in this outbreak . They included 25 samples of river water obtained from the rivers closest to case-patients’ houses; six samples of indoor water , six wastewater samples , and two samples of drinking water in case-patients’ houses . In addition , 27 samples of fresh seafood were obtained from local markets where the case-patients lived . From June 4 through July 2 , 2010 , we conducted a matched case-control study with a target sample size of 60 case-patients and 240 controls . Because the outbreak occurred in a not well-defined population and required rapid investigation , the case-control study would be the most appropriate choice of study design to identify risk factors for cholera infection [17] . In this study , the matching by age groups , sex , and living areas was used [18] . The sample size of this study for 80% power at 5% level of significance was calculated on the basis of the estimated percentage of controls exposed to unsafe water ( 30% ) , the ratio of controls to cases ( 4 ) , and an odds ratio ( OR , 2 . 25 ) as per the method of Kelsey and colleagues [19] . The estimated magnitude of OR that we used to calculate the sample size came from the previous work of Hoge and his colleagues [8] . The smallest reported OR was selected to ensure that the study was large enough and had sufficient statistical power in identifying risk factors with an expected OR of 2 . 25 or higher . During this outbreak , all patients with acute watery diarrhea who sought diagnosis and treatment for their disease at local health facilities were immediately transferred to the nearest cholera treatment center . In all , there were four centers , which were established soon after the outbreak was confirmed and located in three distant local hospitals and one neighboring hospital ( Fig 1 ) . In these centers , case-patients were identified , quarantined , and treated , and they were considered eligible for inclusion if they had laboratory-confirmed V . cholerae identified through conditional culture of rectal swabs and resided permanently in Ben Tre province . For this study , case-patients identified since the beginning of the outbreak were consecutively recruited until the target sample size was reached . Case-patients who were under cholera treatment were prospectively recruited through the four cholera treatment centers , while case-patients who had already been discharged from these health facilities were contacted and recruited at their homes . For each case , four community-based controls [20] were selected and matched by commune , sex and 5-year age group to control for the potential confounding effects of these three factors . They were recruited at their homes on the same day as the interview of their matched case . Specifically , the control search began with random selection of four sub-communes in the area where the case resided by using Microsoft Excel . In each of these sub-communes , currently registered houses were numbered and one house was then chosen by hand drawing from the Vietnamese bingo game , a set containing 90 balls numbered 1–90 . Interviewers subsequently accessed the selected house , in which all household members were primarily screened by sex and age to identify a potential control . Controls were further screened by interviewers to ensure that they had not had acute diarrhea in the month prior to the interview . If no person in that house met the inclusion criteria , the interviewers went to the next house to the left until a matched control was found . No control persons that we approached refused to participate in the present study and be interviewed . Trained health-workers used a structured questionnaire to collect the study information during face-to-face interviews . This questionnaire was adapted from a cholera case investigation form previously developed and used in Ho Chi Minh City in 2008 . For the present study , it was extended and included a section for controls . Risk variables included were based on reports in the literature [8 , 21] and primary assessments conducted at the beginning of the outbreak . The questionnaire contained questions to elicit information on socio-demographic characteristics , exposure to people with diarrheal diseases , recent travel history , and detailed consumption of food and water . The logic and language of questions were tested locally in Ben Tre before being used to collect data . For children aged <6 years , interviews were conducted with their parents or guardians . Potential contaminated food sources were identified through dichotomous questions about the different types of food eaten , eating places , and cooking methods ( e . g . , well-cooked , stir fried , cooked rare , and raw ) within five days before the onset of diarrhea for case-patients or five days before being interviewed for control persons . Participants were also asked about their sources and characteristics of water used for drinking , cooking , and bathing , and their drinking habits ( e . g . , the use of boiled water and drinking water with ice ) via several dichotomous questions in the previous seven days . Clinical symptoms and health-seeking behaviors since the onset of diarrheal symptoms were also collected . Rectal swabs were collected from controls after completion of the interview . During the outbreak , rectal swabs of people with acute diarrhea were directly placed in Cary Blair Transport Medium and transported daily to the Pasteur Institute , Ho Chi Minh City ( PI-HCMC ) for identification of V . cholerae using a standard testing protocol [22 , 23] . All controls’ rectal swabs and environmental samples were stored at the Ben Tre Provincial Preventive Centre and transported fortnightly to the PI-HCMC for testing . Specifically , swab specimens were enriched in alkaline peptone water ( Oxoid ) and inoculated for 6–8 hours . We then cultured the bacterial solution obtained on thiosulfate citrate bile salts sucrose agar ( Merck ) and carried out biochemical and agglutination tests ( Denka Seiken ) . We further used polymerase chain reaction ( PCR ) to amplify the V . cholera isolations from the swab culture , environmental samples , to identify the genomic region encoding the O1/O139 genotypes , and to detect cholera toxins and other toxin-specific genes , specifically ctxA , ctxB , rtxC , and rstC , using previously published primers [24–28] according to the PCR method described by Nguyen et al . [29] . A 9 μl of amplicon was separated by electrophoresis on a 1 . 5% agarose gel and visualized with an ultraviolet light on the Gel Doc System ( Bio-Rad Laboratories ) . Test results were immediately reported to the attending physicians and the local preventive centers for personal treatment and prevention purposes , respectively . Statistical analysis was carried out using Stata version 14 ( StataCrop LP , College Station , TX ) . Information about the participants in the case-control study were initially explored using descriptive statistics ( including frequency and proportion for categorical variables and median and range for continuous variables ) , with comparisons between case-patients and controls made by McNemar’s chi-square tests . Children aged <6 years were classified as not having consumed a certain food or water if their parents or guardians reported that they did not consume it over the study period . To identify risk factors for cholera , conditional logistic regression with forward selection was used to estimate matched OR and 95% confidence intervals ( CI ) as per the method of Hosmer and his colleagues [30] . Initially , univariate analyses were employed to identify variables that were potential risk factors for cholera infection . Those variables with p<0 . 25 , along with those which have been acknowledged to have biologic plausibility for increasing the risk of cholera , were considered for inclusion in the multivariable model . The analysis began by fitting the base model , which included cholera variable , participants’ age , and a variable that yielded the lowest corresponding p-value in univariate analyses . The variable with the next lowest p-value was progressively added to the model , and its contribution to model fit was tested by the log likelihood ratio test . If the test yielded a p-value less than 0 . 05 , it was kept in the model . This process continued until the added variable made no significant improvement in model fit . As the level of education may behaviorally influence personal eating and drinking habits , we investigated potential effect modification by creating interaction terms between it and the other variables retained in the final model . In a sensitivity analysis , to accommodate missing data , we applied multiple imputation by chained equations , with an assumption of missingness at random . Ten imputed datasets were used with 1 , 000 iterations . We compared the risk factors found between the original and imputed data sets .
From May 9 ( epidemiologic week 19 ) through August 3 ( epidemiologic week 31 ) , 2010 , 71 cases of cholera were identified in all towns of Ben Tre province . The area with the highest number of cases was Mo Cay ( 46 cases ) , followed by Giong Trom ( 12 cases ) and Ba Tri ( 7 cases ) ( Fig 1 ) . The number of cases continued to increase , peaked during May 30 to June 5 ( epidemiologic week 22 ) , and then gradually decreased over time ( Fig 2 ) . No case-patient died during the outbreak . After confirmation of the outbreak on May 12 , 2010 , Ben Tre rapidly rolled-out its control activities in the entire province . It included measures to isolate and treat case-patients at four cholera treatment centers , clean and disinfect their houses with chloramine B , and provide chloramine B to enhance the use of chloramine-treated water among people living in the areas where the case-patients emerged . Case-patient’s close contacts were traced , given a single dose of prophylactic ciprofloxacin , and monitored their health . Education campaigns were additionally made to encourage people in the entire province to practice safe water , proper sanitation , and food safety . As reports of primary risk assessments suggested a potential link between drinking iced water and the illness , on June 11 , 2010 , the government of Ben Tre prohibited the manufacture , transportation , sale , or supply of commercial ice in the two most affected towns of Mo Cay and Giong Trom . In total , only V . cholerae O1 Ogawa , El Tor biotype was isolated from 71cholera cases . The ctxA , ctxB , rtxC , and rstC genes were exhibited in 97% , 100% , 100% , and 97% of all isolates , suggesting a presence of altered El Tor strains that combine characteristics of classical and El Tor strains . Due to resource constraints , only 193 rectal swabs taken from 240 controls were tested for cholera . V . cholerae was not detected in any controls’ rectal swabs . Two out of twenty five river water samples and one out of six wastewater samples were culture positive for V . cholerae . PCR analysis revealed the presence of V . cholerae O1 Ogawa carrying ctxA in all of these three positive-culture samples . V . cholerae was absent in two samples of drinking water , six samples of indoor water , and twenty-seven samples of fresh seafood . After excluding seven case-patients lost to follow-up and four case-patients whose disease occurred after the target sample size had been reached , a total of 60 case-patients occurring from May 9 through June 26 were consecutively included in a matched case-control study and matched to 240 controls . Forty case-patients were prospectively recruited through the cholera treatment centers and twenty case-patients were recruited at their homes . The case-patients were interviewed from June 4 through July 2 , 2010 . Nine case-patients and 32 control persons had missing data on at least one variable . Except for age , no significant discrepancies were recorded between the missing and non-missing subjects on variables ( Table A in S1 File ) . Demographic features and risk factors are listed in Table 1 . Two thirds of subjects were females . Participants’ ages ranged from two months to 83 years for the case-patients and from one month to 87 years for the controls . For those aged six years or older , 73% of the case-patients and 57% of the controls reported having little education ( primary education or illiterate ) . The median number of case-patient’s stools within a day prior to hospital admission was eight ( range: 2–20 ) , 55% had watery stools , 40% had abdominal pain , and 53% had vomiting . The mean time of admission to local hospitals was 0 . 7 days after the onset of the disease , and the mean time lag between case onset and enrolment of controls was 12 days among the case-patients . Fishpond or river toilets were frequent in both the case-patient’s and control’s houses ( 73% and 73% , respectively ) . A flush toilet in house was reported in 22% of the case-patients and 25% of the controls ( p = 0 . 404 ) . Thirty percent of the case-patients reported that their main sources of water for use were close to a toilet , whereas only 14% of the controls had a similar situation ( p <0 . 001 ) . A self-report of changes in the color , odor , appearance and taste of water used was seen in 15% of the case-patients and 8% of the controls ( p = 0 . 029 ) . Among the case-patients surveyed , 22% reported drinking iced tea within one week before the disease onset , compared with 3% of the controls in the week before being interviewed ( p <0 . 001 ) . A higher percentage of the case-patients than the controls consumed water with ice ( 81% vs . 68% , p = 0 . 001 ) , unboiled water ( 65% vs . 43% , p <0 . 001 ) , and sedimented river water ( 12% vs . 5% , p = 0 . 014 ) . Up to 52% , 33% , and 42% of the case-patients used sedimented river water as the source of water for bathing , brushing teeth/gargling , and cooking , respectively . These levels were nearly twice as high as those for the controls ( 30% , 16% , and 21% , respectively; all p-values <0 . 001 ) . The case-patients were less likely to use stored rainwater for both drinking ( 70% vs . 89% , p <0 . 001 ) and cooking ( 23% vs . 41% , p <0 . 001 ) , but they were more likely to drink bottled water ( 20% vs . 13% , p = 0 . 027 ) than the controls . The use of indoor tap water for both drinking and cooking was rarely seen in this sample . The case-patients were less likely to eat cooked seafood ( 62% vs . 81% , p<0 . 001 ) , steamed vegetables ( 12% vs . 31% , p <0 . 001 ) and raw vegetables ( 23% vs . 32% , p = 0 . 027 ) in comparison with the controls . While only 2% of the controls were living with people who had acute diarrhea , 15% of the case-patients did so ( p <0 . 001 ) . The case-patients had a greater percentage of travel out of town in the week prior to the disease onset than the controls ( 37% vs . 20% , p <0 . 001 ) . Risk factors for cholera infection identified through the multivariable conditional logistic regression are summarized in Table 2 . Individuals who reported drinking iced tea ( adjusted OR ( aOR ) = 8 . 40 , 95% CI: 1 . 84–39 . 25 ) , not always boiling drinking water ( aOR = 2 . 62 , 95% CI: 1 . 03–6 . 67 ) , living with people who had acute diarrhea ( aOR = 13 . 72 , 95% CI: 2 . 77–67 . 97 ) , having the main household sources of water for use close to a toilet ( aOR = 4 . 36 , 95% CI: 1 . 37–13 . 88 ) , and having little or no education ( aOR = 4 . 89 , 95% CI: 1 . 18–20 . 19 ) were significantly associated with an elevated risk of cholera . The risk of cholera was lower among persons who drank stored rainwater ( aOR = 0 . 17 , 95% CI: 0 . 04–0 . 63 ) , ate cooked seafood ( aOR = 0 . 27 , 95% CI: 0 . 10–0 . 73 ) , and ate steamed vegetables ( aOR = 0 . 22 , 95% CI: 0 . 07–0 . 70 ) . No significant interactions were recorded throughout the analysis . Only the level of education become insignificant in another regression analysis based on a multiple imputation approach ( Table B in S1 File ) .
In this investigation of the 2010 Ben Tre cholera outbreak , most risk factors identified have been previously described and associated with exposure to unsafe water ( such as , not always drinking boiled water , main source of water close to toilet ) or with proximity with another possible case who had acute diarrhea . Our results revealed a significantly increased risk of cholera among individuals who reported drinking iced tea in the week before the onset of their illness . We also observed substantial cholera risk reductions in some subpopulations , specifically those who drank stored rainwater and ate cooked seafood or steamed vegetables . Our study suggested that the spread of cholera in Ben Tre was quite considerable , but not nearly as high as had been seen in Hanoi [11] and other Southeast Asian settings [31 , 32] . Early outbreak detection , community-engaged health promotion , large-scale distribution of chloramine B to treat surface water , its low population density ( an estimated 533 people/km2 ) , and a prohibition of ice block production business during the outbreak could have contributed to the relatively smaller size of the outbreak in Ben Tre province . In agreement with previous studies [14 , 26 , 32 , 33] , we found evidence that V . cholerae O1 , altered El Tor was prevalent among the case-patients . There were no reported deaths attributable to any strains of cholera among the case-patients in this outbreak . This finding could help to address the previously reported clinical concern that the spread of these strains with increased severity would lead to more cholera deaths in some areas of the world [34] . We know of no previous studies that have reported increased risk of infection specifically associated with iced tea consumption . Locally , iced tea is made by adding cooled boiled tea to a glass or bottle of ice . Ice is typically bought from street vendors rather than prepared at home in rural or semi-rural areas of Vietnam because less than one-eighth ( 12% ) of rural households own a refrigerator [35] . A previous hospital-based case-control study conducted in Thailand demonstrated that consumption of ice was significantly associated with increased risk of cholera in a bivariate analysis but was not significant in the multivariable model [8] . In our outbreak investigation , we were not able to obtain commercial and household ice for testing for the presence of V . cholerae in ice . We were therefore unable to ascertain the underlying microbiological mechanism leading to the association between drinking iced tea and cholera . A previous study , conducted in Jakarta , Indonesia , showed that a large percentage of ice ( 35% ) and beverages ( 24% ) were contaminated with V . cholerae [36] . In Vietnam where untreated wells and surface water ( i . e . river water ) have been commonly used for making commercial ice , it is possible for V . cholerae to be introduced into commercial ice , and such contamination could have triggered this resurgence of cholera in Ben Tre province . The observation that the decline in the weekly reported number of cholera was seen after the local government prohibited on the manufacture , transportation , sale , or supply of commercial ice in the outbreak area supports this hypothesis . In future outbreaks , further epidemiological and microbiological investigation of the association between commercial ice and cholera is warranted . The microbiological quality of the water used at ice manufacturing plants should be tested regularly . Our other risk factors for cholera included living with people who had acute diarrheal illness and not always boiling drinking water . These are well-described predictors of cholera in endemic countries [8 , 37] . The association between having the main source of water for use being close to the toilets indicates that pollution of drinking water sources by contaminated feces remains a problem in Vietnam and could lead to further spread of disease in future cholera outbreaks [38] . As expected , we found that drinking stored rainwater , eating cooked seafood , and eating steamed vegetables were protective against cholera , and these results are consistent with findings from several previous studies [21 , 39] . Local public health authorities should rapidly roll out a response that involves appropriate case management and greater promotion of proper sanitation , safe water , and food safety in the event of a cholera outbreak [40] . Our observation that individuals who reported little or no education were at a higher risk of cholera is consistent with the results of a previous case-control study in Harare City , Zimbabwe in which having attained less than secondary education was found to be a risk factor for cholera [41] . Poor education is part of the cycle of poverty , which includes crowded living conditions , malnourishment , poor household sanitation and personal hygiene practices [42] , which are well-known to be important risk factors for cholera [43] . This study has several limitations . First , the number of cholera cases reported herein might not have reflected the actual burden of cholera in Ben Tre province . The vast majority of people infected with V . cholerae are asymptomatic or mildly symptomatic , and thus might not seek healthcare services [44] . Second , detection of V . cholerae in some water samples obtained only from Mo Cay and Giong Trom should be interpreted with caution , because the degree to which these samples are representative of environmental sources in other towns throughout the province where case-patients occurred is unknown . Third , selection of controls from the commune where the case-patients resided can limit exploration of some important geographical and cultural risk factors , such as exposure to river water , due to similar exposures between the controls and the case-patients . Fourth , reduced recall of information may have occurred among case-patients recruited and interviewed late . To reduce the need to recall distant exposures among both the case-patients and controls , study questions were specific to food/water and location . Fifth , about 20% of controls’ rectal swabs were not tested for cholera , potentially misclassifying case-patients as controls . However , such misclassification was unlikely as 80% of controls’ rectal swabs tested were negative for V . cholerae . Sixth , as with other behavioral studies that depend entirely on face-to-face interviews , this study may under-report some important risks , such as food and personal hygiene practices . To reduce this bias , we selected well-trained , experienced , and unprejudiced interviewers after the interview training . Lastly , our study may not have enough power to detect rare risk factors for cholera . Despite these limitations , this present study has important implications for Vietnam’s cholera responses . This emergence of cholera due to V . cholerae O1 , altered El Tor emphasizes the need for increased efforts to prevent the spread of cholera in southern Vietnam . Along with traditional approaches that focus on enhancement of safe water , sanitation , and food safety , combined with periodic provision of oral cholera vaccines , a water quality monitoring system at ice-making plants should be established . It is vital to ensure the quality of the water supply , reduce the introduction of V . cholerae into ice , and subsequently lower the risk of cholera in Vietnam , a tropical setting where consumption of iced drinks is common .
|
Cholera is a highly contagious , acute diarrheal illness , which poses a profound health threat in many parts of the less developed world . The majority of cases are reported from Sub-Saharan Africa , South-East Asia , and the Americas ( i . e . , Haiti ) where infections are primarily transmitted through ingestion of contaminated water . Today in the era of widely available rehydration therapies and antibiotics , deaths due to cholera are quite rare . Despite this , early detection of contaminated water sources is crucial for directing early interventions for curbing community-wide transmission . The authors found evidence linking an outbreak of cholera in southern Vietnam to consumption of unsafe water , especially drinking iced tea . This finding suggests the need for a water-monitoring system at ice-making plants . Further research is needed to confirm the biological link between iced tea consumption and cholera infection . Larger studies should also be conducted to understand the clinical consequences of infection with the new cholera agent ( V . cholerae O1 , altered El Tor strains ) .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"ecology",
"and",
"environmental",
"sciences",
"surface",
"water",
"water",
"resources",
"pathology",
"and",
"laboratory",
"medicine",
"rivers",
"pathogens",
"vibrio",
"tropical",
"diseases",
"microbiology",
"geographical",
"locations",
"diet",
"diarrhea",
"bacterial",
"diseases",
"vibrio",
"cholerae",
"nutrition",
"signs",
"and",
"symptoms",
"aquatic",
"environments",
"gastroenterology",
"and",
"hepatology",
"bodies",
"of",
"water",
"neglected",
"tropical",
"diseases",
"bacteria",
"bacterial",
"pathogens",
"hydrology",
"infectious",
"diseases",
"cholera",
"natural",
"resources",
"medical",
"microbiology",
"marine",
"and",
"aquatic",
"sciences",
"microbial",
"pathogens",
"vietnam",
"people",
"and",
"places",
"tea",
"freshwater",
"environments",
"diagnostic",
"medicine",
"asia",
"earth",
"sciences",
"beverages",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2017
|
Cholera returns to southern Vietnam in an outbreak associated with consuming unsafe water through iced tea: A matched case-control study
|
Envenomation of humans by snakes is a complex and continuously evolving medical emergency , and treatment is made that much more difficult by the diverse biochemical composition of many venoms . Venomous snakes and their venoms also provide models for the study of molecular evolutionary processes leading to adaptation and genotype-phenotype relationships . To compare venom complexity and protein sequences , venom gland transcriptomes are assembled , which usually requires the sacrifice of snakes for tissue . However , toxin transcripts are also present in venoms , offering the possibility of obtaining cDNA sequences directly from venom . This study provides evidence that unknown full-length venom protein transcripts can be obtained from the venoms of multiple species from all major venomous snake families . These unknown venom protein cDNAs are obtained by the use of primers designed from conserved signal peptide sequences within each venom protein superfamily . This technique was used to assemble a partial venom gland transcriptome for the Middle American Rattlesnake ( Crotalus simus tzabcan ) by amplifying sequences for phospholipases A2 , serine proteases , C-lectins , and metalloproteinases from within venom . Phospholipase A2 sequences were also recovered from the venoms of several rattlesnakes and an elapid snake ( Pseudechis porphyriacus ) , and three-finger toxin sequences were recovered from multiple rear-fanged snake species , demonstrating that the three major clades of advanced snakes ( Elapidae , Viperidae , Colubridae ) have stable mRNA present in their venoms . These cDNA sequences from venom were then used to explore potential activities derived from protein sequence similarities and evolutionary histories within these large multigene superfamilies . Venom-derived sequences can also be used to aid in characterizing venoms that lack proteomic profiles and identify sequence characteristics indicating specific envenomation profiles . This approach , requiring only venom , provides access to cDNA sequences in the absence of living specimens , even from commercial venom sources , to evaluate important regional differences in venom composition and to study snake venom protein evolution .
The evolution of venoms among the advanced colubroid snakes has had tremendous adaptive significance and has allowed this clade to diversify rapidly and occupy a diversity of niches globally [1] . Snake venoms are complex glandular secretions that may contain 2–100+ protein/peptide components with a myriad of biological activities , ranging from potent neurotoxins to rapid-acting myotoxins to hydrolytic enzymes [2] . These toxins are synthesized and stored in a cephalic venom gland which allows immediate deployment as a chemical weapon , also necessitating intricate storage and protective mechanisms [3] . Venoms likely allowed a transition from mechanical capture and processing of prey to one dependent on chemical means [4] , and during the approximately 100 million year history of snakes [5] , a diversity of biochemical compositional “strategies” have evolved [6 , 7] . Resulting venom phenotypes can therefore be significantly different , even among closely related species [2] , and these different phenotypes are often correlated with dietary variables or foraging strategies [8–10] . Determining detailed venom composition among differing lineages of snakes can provide important connections linking phenotypic variation to specific selective pressures , and linking venom composition to snakebite envenomation effects . The application of transcriptomic methods has provided insight into venom protein post-transcriptional regulation , as well as documenting isoform diversity and molecular evolutionary trends within large multigene venom protein superfamilies [11–15] . Venom gland transcriptomics has evolved from the generation of ESTs ( expressed sequence tags ) [16–22] to more comprehensive next generation sequencing ( 454 pyrosequencing or Illumina ) of total venom gland cDNA ( complementary DNA ) [23–28] . However , these methods both currently rely on venom gland tissue to obtain venom protein cDNAs , requiring access to venomous snake tissues and animal euthanasia . The ability to acquire venom protein cDNA sequences from venom has been documented [29–32] , but this source has not been fully exploited because mRNA yields have been highly variable and very low , and cDNA amplification has previously focused only on known venom protein transcripts and only partial transcripts were amplified . Extracellular messenger RNA has been demonstrated to be unusually stable , for at least several years within lyophilized venom [30] . This alternative source to obtain venom protein cDNAs is a less destructive method because sacrifice of animals is avoided . It also increases the availability of venom protein cDNA sequences to researchers that have limited access to venom gland tissues , as in the case of rare or difficult to acquire snake species , or due to limitations on animal euthanasia protocols . Further , this approach provides the opportunity to generate both venom transcriptomic and proteomic profiles , essentially a genotype-phenotype map , using only the same venom sample from one individual . The standardized venomics approach of characterizing venoms by separating venom components by HPLC ( high performance liquid chromatography ) , followed by trypsin digestion and tandem mass spectrometry , primarily relies on protein identification from databases such as MASCOT [33] . Mass spectral matching to determine peptide sequences is an efficient and less expensive alternative to N-terminal sequencing ( Edman degradation ) . An advantage to this venomic approach is that it is at least semi-quantitative , allowing inter- and intrapopulational variation in amounts of specific proteins to be estimated . However , many venom proteins , in particular those from rear-fanged venomous snakes , are not present within current databases or are poorly represented , making it difficult to use this methodology to characterize these venoms [9 , 33 , 34] . The incorporation of transcriptomics into venom proteomics has resulted in venom protein-locus resolution and has been labeled next generation snake venomics [18 , 23 , 28 , 34] . Species-specific venom gland transcriptomes aid in the identification and characterization of venom profiles by providing custom databases for tandem mass spectrometry ( MS/MS ) spectra matching , allowing for the identification of additional peptide sequences , specific isoforms , and novel venom proteins [16 , 24 , 28 , 34 , 35] . This study provides support for an approach to obtain species-specific venom protein transcript sequences , including those from rear-fanged venomous snakes , using relatively little starting material ( 2 mg of lyophilized venom or 100 μl of fresh venom ) that does not require venom gland tissue . cDNA derived from this method has great potential to fill gaps within databases and to aid in the characterization of understudied snake venoms . Acquiring full-length venom protein sequences can also help to identify protein characteristics indicative of serious envenomation profiles , because venom toxins commonly utilize conserved structural folds but produce diverse pharmacological activities . For example , many venoms contain enzymatic phospholipase A2s ( PLA2s ) of low toxicity [36 , 37]; some contain PLA2s with potent myotoxic activity , and a limited number contain neurotoxic PLA2s ( crotoxin-like complexes or other asparagine-6 containing PLA2 sequences ) . Current methods for examining venom protein pharmacological activity can be labor intensive , requiring multiple protein purification steps and functional assays . Phospholipase A2 functionality based on clustering with other PLA2s sharing similar sequence has been demonstrated to be a potential in silico alternative for predicting specific PLA2 activity [37] . Knowledge of sequences and sequence similarities to others with noted activity and that are currently promising drug leads can help guide new drug exploration or even help to identify unique characteristics of prey-specific toxins that allow them to specifically target selected prey receptors , as in the case of rear-fanged snake venom three-finger toxins ( 3FTxs ) . Rear-fanged snake venoms have not been as well-studied as front-fanged snake venoms , largely due to the difficulties extracting venoms from these snakes and the fact that the large majority of envenomations from these snakes are not life threatening [38–41] . However , rear-fanged snake venoms potentially contain proteins that could serve as novel pharmaceutical drug leads or in other applications , such as proteolytic enzymes for protein fragmentation for mass spectrometry [42 , 43] . Rear-fanged snake venoms have also demonstrated unique evolutionary trajectories , including the presence of the only prey-specific toxins yet identified within snake venoms [8 , 44 , 45] . The aim of this study was to obtain cDNA sequences of abundant venom proteins to predict protein activities or envenomation symptomology . From the rattlesnake PLA2 transcripts identified , sequences characteristic of known viper neurotoxic PLA2s were observed , demonstrating the utility of using these methods to obtain predictive venom activities and envenomation profiles . In addition , this technique could be used to screen for novel sequences that could be of potential biomedical development based on predicted activities , and to explore the evolutionary trajectories in complex multigene superfamilies .
Venom was collected from front-fanged vipers Crotalus scutulatus scutulatus ( Mohave Rattlesnake; SE Arizona ) , Crotalus cerastes ( Sidewinder; S Arizona ) , Crotalus oreganus cerberus ( Arizona Black Rattlesnake; E Arizona ) , Crotalus oreganus concolor ( Midget Faded Rattlesnake; S . Wyoming ) , and Sistrurus miliarius barbouri ( Florida Pigmy Rattlesnake; central Florida ) by placing an RNase Away ( Thermo Fisher Scientific Inc . , U . S . A . ) -treated 100 μl capillary tube over each fang and gently massaging the gland; 100 μl of venom was then immediately added to 1 mL of TRIzol ( Life Technologies , CA , U . S . A . ) . Venom from Crotalus simus tzabcan ( Middle American Rattlesnake; Yucatán Peninsula , México ) was extracted into a sterile beaker and 10 μl , 25 μl , 50 μl and 100 μl of venom were each immediately added to 1 mL of TRIzol; the remaining venom was then centrifuged ( 9500 x g for 5 minutes ) , lyophilized and stored at -20°C until used . Venom from Crotalus molossus nigrescens ( Mexican Black-tailed Rattlesnake; Morelia , México ) was collected in the field , desiccated , and stored at -20°C until used . Lyophilized venom from Crotalus pricei ( Twin-spotted Rattlesnake; SE Arizona ) was collected from a captive snake and stored frozen ( -20°C ) with desiccant for approximately 20 years; lyophilized Crotalus basiliscus ( Mexican West Coast Rattlesnake; W . México ) venom was purchased from the Miami Serpentarium ( Lot#CB15SZ ) and lyophilized venom from an elapid snake , Pseudechis porphyriacus ( Red-bellied Black Snake; E Australia ) , was a gift from Venom Supplies Pty Ltd ( Tanunda , South Australia ) . Venoms from the rear-fanged snakes Boiga irregularis ( Brown Treesnake , Guam ) , Boiga dendrophila ( Mangrove Snake; Indonesia ) , Boiga nigriceps ( Black-headed Catsnake; Indonesia ) , Trimorphodon biscutatus lambda ( Sonoran Lyre Snake; Portal , AZ ) , and Alsophis portoricensis ( Puerto Rican Racer; Guana Island , British Virgin Islands ) were extracted using the method of Hill and Mackessy ( 1997 ) with subcutaneous injections of ketamine-HCl ( 20–30 mg/kg ) followed by pilocarpine-HCl ( 6 mg/kg ) . Venom was collected by placing RNase Away-treated 100 μl capillary tubes over each enlarged rear maxillary tooth [46] , and venom ( 100 μl ) was then added to 1 mL of TRIzol . For the rear-fanged snakes Boiga cynodon ( Dog-toothed Catsnake; Indonesia ) , Oxybelis fulgidus ( Green Vine Snake; Central America ) , and Ahaetulla prasina ( Asian Vine Snake; Indonesia ) , venom was collected using the same protocol without RNase Away treated capillary tubes , centrifuged ( 9500 x g for 5 minutes ) , lyophilized , and stored at -20°C until used . The 3’ RACE System kit , PCR SuperMix High Fidelity polymerase , custom oligonucleotides , DNase I , and Escherichia coli DH5 α competent cells were purchased from Life Technologies , CA , U . S . A . The plasmid Quick Clean 5M Miniprep Kit was from GenScript , Inc ( Piscataway Township , NJ , U . S . A ) , and the pGEM-T Easy Vector System and Wizard SV gel and PCR clean-up system from Promega , Inc . ( Madison , WI , U . S . A . ) All other reagents were purchased from Sigma ( St . Louis , MO , U . S . A ) . Many of the snake venoms were collected manually from venomous snakes maintained in the University of Northern Colorado Animal Resource Facility in accordance with protocols #9204 and 1302D-SM-S-16 ( evaluated and approved by the UNC IACUC ) and collecting permits from state and federal agencies ( Arizona Game and Fish Department #MCKSY000221 and #SP727017; Colorado Parks and Wildlife #15HP974; U . S . Fish and Wildlife Service #MA022452-0 ) . Snakes are maintained for venoms in accordance with guidelines published by the American Society of Ichthyologists and Herpetologists [47] . RNA was purified from 2 mg of lyophilized venom or 100 μl of freshly collected venom that had been added to 1 mL of TRIzol following the recommended TRIzol RNA protocol: after incubation for 5 minutes , 200 μL of chloroform was added to each tube , tubes were centrifuged at 12 , 000 x g for 15 minutes , aqueous upper phases were transferred to new RNase-free tubes , and 500 μL 100% isopropanol added to each aqueous phase to precipitate RNA . Tubes were incubated at room temperature for 10 minutes and centrifuged at 12 , 000 x g for 10 minutes . Supernatant was removed and the resulting RNA pellet ( not visible ) washed with 1 mL 75% ethanol . Another centrifuge step at 7 , 500 x g for 5 minutes was performed and supernatant poured off . A -20°C overnight incubation in 300 μL 100% ethanol with 40 μL 3 M sodium acetate was then performed to increased RNA yields , and the following day tubes were centrifuged at 10 , 000 x g for 15 minutes , supernatant removed , and total RNA resuspended in 16 μL nuclease-free H2O . To evaluate the effect of different amounts of lyophilized and fresh venom on cDNA yields , 5 mg , 10 mg , and 20 mg of lyophilized venom or 10 , 25 , 50 or 100 μL of crude fresh venom from C . s . tzabcan was added to TRIzol reagent and processed as above . For rear-fanged snake venoms , extraction methods resulted in retention of significant amounts of contaminating DNA; therefore , an Amplification Grade DNase I digestion after RNA isolation was performed at room temperature for 15 minutes , followed by the addition of 1 μl 25 mM EDTA ( pH 8 . 0 ) and a 15 minute 65°C incubation . cDNA synthesis from total RNA was accomplished using the 3’ RACE System following the manufacturer’s protocols . The oligo ( dT ) adaptor primer provided with the kit initiated reverse transcriptase cDNA synthesis and effectively selected for polyadenylated mRNAs . Sense primer sequences were designed from conserved signal peptide regions for each venom protein superfamily ( Table 1 ) . To identify conserved signal peptide sequences , multiple sequence alignments within MEGA v6 . 06 [48] using MUSCLE [49] were performed for each venom protein superfamily , with representative sequences obtained from the NCBI ( National Center for Biotechnology Information ) nucleotide database . Each sense primer was used in a reaction with the 3’RACE system AUAP antisense primer 5’-GGCCACGCGTCGACTAGTAC-3’ . For Mojave toxin , published sense and antisense primers were used for both acidic and basic subunits [50] . Twenty-three μL of PCR SuperMix High Fidelity polymerase was used with 1–2 μL of cDNA template and 0 . 5 μL of each primer ( sense and antisense ) . PCR was performed with seven touchdown cycles of 94°C for 25 seconds , 52°C for 30 seconds , and 68°C for two minutes . Thirty additional cycles followed with 94°C for 25 seconds , 48°C for 30 seconds , and 68°C for two minutes with a final 68°C extension for five minutes . The amplified products were observed on a 1% agarose gel , and bands of the estimated transcript sizes ( based on previous published transcripts ) were excised and then purified using the Wizard SV gel and PCR clean-up system . Amplified and purified cDNA was ligated into the pGEM-T Easy Vector System and transformed into Escherichia coli DH5 α competent cells following the manufacture’s recommended protocol . Transformed E . coli were grown on nutrient rich agar plates overnight at 37°C with ampicillin , IPTG and β-galactosidase for white/blue colony selection . Recombinant plasmids were selected from agar plates , and E . coli colonies picked for viper PLA2 sequences were as follows: three colonies were picked for Crotalus cerastes , three for S . m . barbouri , four for C . m . nigrescens , six for C . o . cerberus , six for C . basiliscus , eight for C . pricei , twelve for C . o . concolor , and fifteen were picked for C . s . tzabcan . In addition to the PLA2 sequences , eight colonies for serine proteases , ten colonies for C-type lectins , and eighteen colonies for metalloproteinases were chosen from C . s . tzabcan venom to obtain transcripts for all major venom protein families . For rear-fanged colubrid snake 3FTxs , the following number of E . coli colonies were selected: three were picked for T . b . lambda , three for A . prasina , six for O . fulgidus , four for B . nigriceps , ten for B . cynodon , twenty for B . dendrophila , and nineteen were picked for B . irregularis . A sampling of 3FTxs from Boiga sp . was chosen because of the three currently identified prey-specific 3FTxs , two have been found in Boiga species [8 , 44] . Three-finger toxin sequences from rear-fanged snakes , especially those from Boiga sp . and O . fulgidus , can provide insight into the evolution of 3FTx prey-specific binding affinities . Three colonies were also picked for metalloproteinase transcripts in A . portoricensis venom . For the elapid snake P . porphyriacus , five colonies from amplified PLA2s were picked . The numbers of colonies picked varied depending on the number of expected isoforms within each snake venom protein family and also because some primers were still being evaluated for specificity ( only a few colonies were selected in these cases ) . Each E . coli colony was placed into 2 mL LB broth with 1 μL/mL ampicillin , and shaken overnight at 37°C . Plasmid copies for each E . coli colony were than purified using the Quick Clean 5M Miniprep Kit and were sequenced at the DNASU facility ( Arizona State University , AZ , U . S . A ) using Big Dye V3 . 1 chemistry with samples processed on an Applied Biosystems 3730XL Sequence Analysis Instrument . Sequences were viewed with 4Peaks software ( http://nucleobytes . com/index . php/4peaks ) and base pairs with acceptable quality scores ( Phred score >20 ) were retained for analysis . Redundant sequences were removed . Sequences were identified with BLASTx ( Basic Local Alignment Search Tool ) on the NCBI server , limiting the search to “Serpentes ( taxid: 8570 ) ” proteins . Protein identities were considered significant if they fell below an e-value threshold of e-4 and shared sequence similarity to other known snake venom proteins . Sequences were translated to their corresponding amino acid sequence and trimmed in MEGA v6 . 06 [48] , then aligned with MUSCLE [49] and manually checked . Sequence alignment figures were generated using BoxShade 3 . 3 . 1 ( http://mobyle . pasteur . fr/cgi-bin/portal . py ? #forms::boxshade ) . All full-length CDS sequences , including all rattlesnake PLA2s , C . s . tzabcan serine proteases , C . s . tzabcan C-type lectins , and rear-fanged snake three-finger toxin sequences , were submitted to GenBank ( accessions KU666900-KU666937 ) . Phylogenetic analysis was completed with MrBayes v3 . 2 . 4 [51] using models selected by PartitionFinder v1 . 1 [52] . PartitionFinder v . 1 . 1 models selected were favored using Akaike Information Criterion . These datasets were then run in duplicate using MrBayes v3 . 2 . 4 with the default of three heated and one cold chain for 1x107 generations , sampling every 1 , 000 generations , and with the first 10% discarded as burn-in . Tracer v1 . 6 ( http://tree . bio . ed . ac . uk/software/tracer/ ) was used to check for run convergences . Consensus tree figures were prepared with FigTree v1 . 4 . 0 ( http://tree . bio . ed . ac . uk/software/figtree/ ) .
Venom RNA concentrations were determined using both a Nanodrop 2000 ( Thermo Fisher Scientific , NY , U . S . A ) and a Qubit 2 . 0 Fluorometer ( Life Technologies , CA , U . S . A ) with a high sensitivity RNA assay kit . Various RNA isolation kits and reagents were used to determine which method produced the greatest RNA yields and amplification success ( Table 2 ) . The TRIzol RNA isolation protocol described in the methods was found to produce the most consistent results when isolating venom RNA and amplifying transcripts; however , the Direct-zol RNA kit ( Zymo Research , CA , U . S . A ) and mirVana miRNA isolation kit ( Life Technologies , CA , U . S . A ) were also successful . Dynabeads from Life Technologies , CA , U . S . A were tried using a previously published technique for isolating extracellular mRNA within venom [29 , 30] , as well as a FastTrack MAG mRNA isolation kit ( Life Technologies , CA , U . S . A ) and a RNeasy mini kit ( QIAGEN , CA , U . S . A ) , but cDNA amplification did not produce visible PCR products ( Table 2 ) . The TRIzol reagent Nanodrop readings for RNA yields of different rattlesnake venoms varied from 69 ng/μl ( 1 . 1 μg of total RNA isolated from 2 mg of lyophilized rattlesnake venom ) to 683 . 3 ng/μl ( 10 . 9 μg of total RNA isolated from 100 μl of fresh rattlesnake venom ) ( Table 2 ) . Rear-fanged snake venoms consistently showed slightly higher yields ( 10 . 3 μg– 13 . 6 μg of total RNA ) as determined by Nanodrop; however , this appeared to be mostly due to the 260 nm readings of contaminating DNA from saliva during venom collection , and all RNA isolated from rear-fanged snakes required a DNase I digestion before PCR to prevent nonspecific amplification . When fresh venom was used , 100 μl of C . s . tzabcan venom yielded a total RNA amount ( Nanodrop ) of 6 . 1 μg , 50 μl yielded 8 μg , and 25 μl yielded 10 . 4 μg . However , when the same volume amounts were used for cDNA synthesis and amplification , successful amplification of PLA2 transcripts tended to decrease with decreasing venom input ( Fig 1A ) . When different lyophilized venom amounts were used , 20 mg yielded 6 . 1 μg , 10 mg yielded 5 . 2 μg , 5 mg yielded 6 . 5 μg , 2 mg yielded 5 . 3 μg , and 1 mg yielded 8 . 4 μg of total RNA ( Nanodrop ) . As seen with the total RNA amounts from fresh venom , the Nanodrop amount of total RNA from lyophilized venom did not demonstrate a clear relationship to PLA2 transcript amplification success , and 2 mg produced the highest concentration of PLA2 amplicons ( Fig 1B ) . All Nanodrop readings showed low 260/280 and 260/230 ratios , indicating low purity . The typical 260/280 ratio observed was 1 . 5 and 1 . 6 for 260/230 . Qubit results revealed values < 20 ng/μl , below instrument detection , for all measured samples ( Table 2 ) . It was possible to isolate RNA from both fresh venom and from lyophilized venom that had been stored at -20°C ( including after 20 years of storage ) , as well as from venom desiccated in the field and venom purchased from a commercial venom supply source . Both front-fanged and rear-fanged venomous snakes were found to have extracellular RNA within their venoms; this is the first report of mRNA in the venom of rear-fanged venomous snakes , and RNA was isolated from both freshly collected and lyophilized rear-fanged snake venoms . As proof of concept , cDNA amplicons from C . s . scutulatus venom , obtained using published primer sequences [49] , were used to confirm presence of the two Mojave toxin subunits ( acidic and basic chains; Fig 1C ) . These basic and acidic subunits were both sequenced and found to be 100% identical to the published sequences for C . s . scutulatus ( PA2A_CROSS and PA2Ba_CROSS ) [53 , 54] , demonstrating that cDNAs of mRNA within venom can be used to detect the presence of specific expressed venom protein transcripts . In this case , the presence and abundance of crotoxin/Mojave toxin-like acidic and basic subunits is strongly indicative of neurotoxic envenomation symptoms characteristic of human envenomations by these rattlesnakes . 3’RACE with sense primers designed from conserved sequences of the signal peptide or the 5’UTR ( untranslated region ) of transcripts ( Table 1 , S1 Fig ) were used to amplify cDNAs for a diversity of PLA2s , metalloproteinases , serine proteases , C-type lectins , and 3FTxs from viperid , elapid and rear-fanged snake venoms . This is the first time that cDNA derived from venom transcripts has been used to obtain unknown sequences for such a diversity of venom protein families . For Middle American Rattlesnake venom ( C . s . tzabcan ) , full-length cDNA sequences were successfully amplified for the major venom proteins present within this rattlesnake’s venom [55] in spite of limited colony sampling ( Table 3 ) . A partial venom gland transcriptome was assembled , focusing on venom protein transcripts that significantly contribute to envenomation symptomology , including metalloproteinases , serine proteases , and C-type lectins . There are likely many more unique C-type lectins , serine proteases , and metalloproteinase transcripts within C . s . tzabcan venom , but the intent here was to demonstrate the presence of diverse , intact venom protein transcripts ( Fig 2 ) . Greater diversity of C-type lectins has been identified within other rattlesnake venom gland transcriptome assemblies using next generation sequencing [23 , 25 , 27] , but the vast majority of these C-type lectins transcripts were found to be present in very low abundance . The diversity of PLA2 isoforms appeared to vary for each rattlesnake species ( Table 4 ) . For C . pricei , only one unique PLA2 sequence was discovered in eight selected clones , while C . m . nigrescens had three unique sequences found in the selection of only four clones . There was no clear positive trend between the number of colonies sequenced and the number of unique sequences ( Pearson’s correlation test; df = 6 , r = 0 . 6684 , p = 0 . 0699 ) ; instead , isoform diversity appeared to be dependent on the species ( and likely on the abundance of the most prominent isoform ) . The number of clones sampled in this study was relatively low , and an increase in the number of clones sequenced should increase the chance of observing less abundant isoforms and determining the total number of isoforms present in each venom [19 , 56] . Of the fifteen PLA2 clones selected from C . s . tzabcan , two unique sequences were similar to sequences from crotoxin or Mojave toxin-like acidic A chain ( Fig 3 ) . One sequence ( C_s_tzabcan1 ) was the most abundant , with six identical clones , and it was 99% identical in amino acid sequence to crotoxin acidic A chain ( PAIA_CRODU ) from the South American Rattlesnake ( Crotalus durissus terrificus ) , while the second sequence ( C_s_tzabcan4 ) had only one clone and was 88% identical to the Mojave toxin acidic ( A ) chain ( PA2A_CROSS ) from the Mojave Rattlesnake ( C . s . scutulatus ) . This less abundant acidic chain sequence revealed that isoform variation within the acidic ( A ) chains of the toxin also exists for C . s . tzabcan . A crotoxin-like basic ( B ) chain was also sequenced from C . s . tzabcan venom , with three clones that were 100% identical in amino acid sequence to crotoxin subunit CBc from C . d . terrificus ( PA2BC_CRODU ) , providing molecular evidence that C . s . tzabcan from this study has an abundance of available PLA2 transcripts to form a neurotoxic complex similar to that of C . d . terrificus ( Fig 3 ) . The PLA2 cDNAs obtained from C . basiliscus also had multiple clones containing crotoxin-like basic B chain sequences ( C_basiliscus3 ) , which were 100% identical in amino acid sequence to crotoxin basic subunit CBc of C . d . terrificus ( PA2BC_CRODU ) ( Fig 3B ) . However , crotoxin-like acidic subunit sequences were not discovered , either due to a lack of sufficient sampling or absence from this venom sample . Crotoxin-like protein complexes have been previously observed in C . basiliscus venom [57] . Mojave toxin-like PLA2 sequences were obtained from C . o . concolor venom ( Fig 3 ) , which has been previously recognized as containing a neurotoxic PLA2 complex ( concolor toxin ) similar to Mojave toxin , identified from Ouchterlony immunodiffusion , immunoelectrophoresis , ELISA , and Western blot analyses [58 , 59]; however , the full sequence has not been published . Concolor toxin acidic ( A ) chain was found to share 100% sequence identity with Mojave toxin acidic ( A ) subunit ( PA2A_CROSS ) from C . s . scutulatus; however , the concolor basic ( B ) subunit was found to be more similar to crotoxin subunit CBc ( PA2BC_CRODU ) from C . d . terrificus , sharing 99% amino acid sequence identify with this crotoxin basic ( B ) chain . Interestingly , C . o . cerberus and C . m . nigrescens venoms also had PLA2 sequences that contained the asparagine-6 ( N6 ) substitutions associated with myotoxic/neurotoxic activity , also a feature of the basic ( B ) subunits of crotoxin and Mojave toxin ( Fig 3 ) . All other PLA2 sequences were similar in sequence to acidic PLA2s that show edema-inducing activity and myotoxicity , corresponding to the envenomation symptomology seen in bites from these species . The S . m . barbouri PLA2 sequence reported here was found to be 100% identical to the amino acid sequence of a previously reported PLA2 from S . m . barbouri venom ( ABY77929 . 1 [60] ) . There were only two unique PLA2 transcripts identified in five colonies selected from an elapid ( P . porphyriacus ) venom ( Table 3 ) . However , of the identified PLA2 sequences , 4/5 clones were 100% identical in mature protein sequence ( P . porphyriacus1 ) to a previously identified PLA-1 precursor ( AAZ22667 . 1 ) from P . porphyriacus venom gland [61] , and the other unique isoform ( P . porphyriacus2 ) was 99% identical to PLA2 pseudexin B chain ( PA2BB_PSEPO ) , also from P . porphyriacus [62] . Again , sequences determined from venom-derived mRNAs are identical to previously reported venom protein sequences ( Fig 4 ) , validating this method . Full-length venom protein transcripts were also identified from rear-fanged snake venoms . Thirty full-length 3FTx sequences were obtained using a degenerate sense primer designed from multiple sequence alignments with published non-conventional 3FTx sequences ( Fig 1D , S1B Fig ) . Three-finger toxin transcripts were found in the venoms of T . b . lambda , A . prasina , O . fulgidus , B . nigriceps , B . cynodon , B . dendrophila , and B . irregularis ( Fig 5 ) . Although none of the three 3FTx sequences from T . b . lambda venom were 100% identical to previously published 3FTx sequences from this species [63] , they did cluster with the previous T . b . lambda sequences within a well-supported clade also containing other 3FTx sequences from New World rear-fanged venomous snakes ( Fig 6 ) . A greater diversity of 3FTxs within T . b . lambda venom is possible considering that only three clones were picked for this study and all were unique 3FTx sequences; differences observed could also be due to locality-specific transcript variation . Two unique 3FTx sequences were found in A . prasina venom and only one unique 3FTx sequence in O . fulgidus . The sequence from O . fulgidus venom was not identical to the previously characterized fulgimotoxin , which was based on N-terminal Edman degradation sequencing [45]; however , it did show 95% amino acid sequence identity . Only two unique 3FTx sequences were found in the venom of B . cynodon ( from ten selected colonies ) and one in B . nigriceps venom ( from four selected colonies ) . Many of the clones sequenced from rear-fanged snake venoms were of poor quality and were culled , so more sequences are likely present . Fifteen unique 3FTx sequences were revealed in B . dendrophila venom , but the majority were missing complete signal peptide sequences and therefore were omitted from further analysis because it is unknown if these transcripts produce proteins that are secreted in the venom gland and are active components of B . dendrophila venom . Six unique 3FTx sequences were found in B . irregularis venom , but none were 100% identical to either irditoxin subunits [8] , although one sequence was 96% identical to irditoxin subunit B . One of the 3FTx B . cynodon clones was 97% identical ( amino acid sequence ) to irditoxin subunit B and another sequence had 83% amino acid sequence identity with irditoxin subunit A; these toxins also clustered together with irditoxin in the 3FTx phylogenetic tree ( Fig 6 ) . These results suggest that B . cynodon venom likely contains a prey-specific heterodimeric 3FTx complex similar to irditoxin . One unique metalloproteinase sequence was amplified , cloned and sequenced from the venom of the Puerto Rican Racer ( Alsophis portoricensis ) . Although this sequence was not similar to alsophinase , a previously characterized metalloproteinase in A . portoricensis venom [43] , it was similar in sequence to other rear-fanged and elapid P-III metalloproteinase cDNA sequences ( Fig 7 ) . The complete metalloproteinase sequence was amplified , based on the observed amplified product size , but longer venom protein transcripts ( >2 , 000bp ) required multiple sequencing reactions that were not performed for this analysis , and therefore only the partial sequence is shown in the alignment ( Fig 7 ) . While optimizing primers , other sequences were incidentally amplified from both rattlesnake and rear-fanged snake venom , including complete 60S ribosomal sequences . These sequences , from C . cerastes and A . portoricensis venom , were 99% identical to the predicted Burmese Python ( Python bivittatus ) 60S ribosomal protein L7a ( XP_007420634 . 1 ) and L15 isoform X1 ( XP_007421748 . 1 ) , respectively . There were also 40S ribosomal protein sequences amplified from C . s . scutulatus which showed 100% sequence identity with the 40S ribosomal protein S9-like isoform X1 from P . bivittatus ( XP_007439934 . 1 ) . Cathelicidin-OH antimicrobial peptides ( XP_007442672 . 1 ) were identified from C . o . cerberus and B . irregularis venom . These sequences were observed in both rattlesnake and rear-fanged snake venoms , demonstrating that other complete transcripts , in addition to venom protein transcripts , exist within venoms ( S2 Fig ) . Sequences that were similar to crotoxin/Mojave toxin acidic ( A ) subunits in C . o . concolor and C . simus tzabcan venoms formed one well-supported clade ( 1 . 0 posterior probability ) , and sequences that were similar to crotoxin/Mojave toxin basic ( B ) subunits discovered in C . o . concolor , C . simus tzabcan , and C . basiliscus clustered with other known neurotoxic N6 PLA2 homologs ( 1 . 0 posterior probability; Fig 8 ) . Other PLA2s from C . o . concolor , C . simus tzabcan , and C . basiliscus venoms , and also from C . pricei , C . cerastes , C . m . nigrescens , C . o . cerberus , and S . m . barbouri venoms , clustered within an acidic hemolytic PLA2 clade shared with other rattlesnakes ( 0 . 96 posterior probability; Fig 8 ) . It has been experimentally determined that even neurotoxic PLA2s can also exhibit anticoagulant activity , and this appears to be a common characteristic of many venom PLA2 enzymes .
Concentrations of total extracellular RNA within snake venom were observed to be moderate but variable using Nanodrop . Using a Qubit instrument , RNA concentrations were all below the instrument detection limit ( < 20 ng/μl ) ( Table 2 ) . Qubit readings provide better accuracy because the fluorescent dye is highly selective for RNA and not DNA , which has also been detected in venom [64] . In addition , common contaminants do not affect Qubit readings . The more accurate Qubit readings revealed a much lower concentration of extracellular RNA within venom , lower than Nanodrop concentrations reported in this and in a previous study [30] . However , in spite of low RNA concentrations , venom protein cDNAs could still be amplified successfully from both front and rear-fanged venomous snakes . By using extracellular messenger RNA from venom to obtain full-length venom protein transcripts , this method can be used without the need to sacrifice living animals to obtain venom gland tissue . It was also possible to successfully amplify full-length venom protein transcripts from venom that was fresh , lyophilized or stored at -20°C for 20+ years , desiccated over silica gel in the field , or obtained from a commercial venom supplier . This represents a significant advance over previous attempts to amplify venom-derived mRNAs , which typically produced only partial sequence transcripts [29 , 30] . Venom protein genes experience an accelerated rate of nucleotide substitution [65 , 66] , making it difficult to design sense and antisense pairs of primers to amplify unknown venom proteins sequences , which is why complete venom gland transcriptomes from gland tissue are usually assembled . However , venom proteins demonstrate high conservation of nucleotide signal peptide sequence and/or 5’UTRs ( Fig 2 ) . By designing degenerate sense primers from these conserved nucleotide sequences and performing 3’RACE , the successful amplification of a diversity of transcript sequences for the major venom protein families ( metalloproteinases , serine proteases , C-type lectins , phospholipase A2s , and three-finger toxins ) responsible for clinically significant snakebite were obtained from venom . This approach also allowed determination of unique , currently unknown full-length toxin sequences for many front-fanged and rear-fanged species in all of the major clades of venomous snakes ( Viperidae , Elapidae , Colubridae ) . This use of degenerate primers to amplify unknown full-length venom protein sequences within a superfamily from snake venom can be employed to screen sequences within each species for toxins of interest , to examine novel mutations within a venom protein superfamily , or to provide an inexpensive method to obtain complete amino acid sequence for a protein under investigation . Venom gland transcriptomes generated from next-generation sequencing ( Roche 454 or Illumina ) provide more comprehensive transcriptome profiles and identify the complete repertoire of transcripts within each venom protein superfamily [23 , 25 , 27] . An abundance of unique 3FTx transcripts identified in rear-fanged snake venom gland transcriptomes generated by next-generation sequencing has been reported , with over fifty 3FTxs transcripts in the case of Boiga irregularis [28] . The number of unique venom protein PLA2 sequences discovered in viper venom gland transcriptomes completed with next-generation sequencing ranges from 4–9 [23 , 25 , 27]; therefore , the number of unique sequences obtained in this study was by no means a comprehensive evaluation of all transcripts within these protein superfamilies . However , using established procedures that are readily accessible to many researchers , such as 3’RACE and the selection/sequencing of E . coli clones , it was possible to identify the major transcripts present for each venom protein superfamily explored in this study . The approach used here allows for researchers interested in a single venom protein superfamily to obtain selectively all highly abundant transcripts for that protein superfamily . This approach is cost-effective and does not require the computing resources/bioinformatics needed for next generation sequencing transcriptome assemblies . Because venom protein cDNA sequences are obtained from venoms , this method also allows for the assembly of a genotype-phenotype map , using only venom as source material . Phospholipase A2 enzymes and 3FTxs were chosen as the main focus of this study because they constitute very large venom protein superfamilies that exhibit a diversity of activities , including neurotoxic , myotoxic , cardiotoxic , anticoagulant and hemolytic activities [2 , 67 , 68] . These venom proteins are ideal for structure/function studies as well as protein engineering studies , because a variety of activities and functional sites are possible using the same conserved protein structural scaffold . Also , 3FTxs and PLA2s are venom proteins that are observed in abundance in snake venoms , and both are toxins that contribute significantly to serious snake envenomation symptomology . Presence of crotoxin or Mojave toxin PLA2 heterodimeric complexes result in phenotypically neurotoxic venoms , and the absence or presence of these complexes result in distinctive venom types that have been labeled type I and type II . Type I venoms have higher metalloproteinase activity and lower toxicity , and type II venoms have low metalloproteinase activity and high toxicity/neurotoxicity [69] . There can be variation in the occurrence of crotoxin/Mojave toxin complexes within a species , as is seen among different populations of C . horridus , C . scutulatus and C . simus throughout their range [55 , 70 , 71] . This study shows that it is possible to detect the acidic and basic subunit transcripts of these neurotoxic PLA2 complexes within venom . In the case of C . s . tzabcan utilized in this study , several isoforms of acidic and basic crotoxin-like subunits were observed . The neurotoxicity of C . s . tzabcan venom varies with snake locality [58]; because the specific locality of the C . s . tzabcan used in this study was unknown , sequencing venom protein transcripts present within venom was a successful approach to evaluating venom phenotype . This technique can also be used to analyze the amino acid sequences of toxins in unexplored venoms , and this study is the first to report the complete sequence for both subunits of concolor toxin from venom of C . o . concolor . Although it has been known for some time that this neurotoxic complex is present in C . o . concolor venom [58 , 59] , the presence of acidic and basic crotoxin/Mojave toxin homologs confirmed that a PLA2-based neurotoxin was present in this type II venom . The viper PLA2 Bayesian sequence similarity tree revealed some distinctive clusters that corresponded with experimentally characterized PLA2 protein activities . For example , analysis of PLA2 sequences from C . o . cerberus , a subspecies with type I venom , demonstrated that its PLA2 clustered within the acidic hemolytic PLA2 clade , as is typical of many low toxicity rattlesnake PLA2s . Mojave and crotoxin-like PLA2 clusters for acidic and basic subunits were also separate from the clades that contained Old World viper neurotoxic PLA2 complexes ( basic and acidic subunits of a heterodimer PLA2 from Vipera nikolskii and vaspin subunits from Vipera aspis aspis ) , suggesting the possibility of a separate evolutionary origin for Old World and New World neurotoxic heterodimeric PLA2 complexes . Venom 3FTxs and PLA2s can have multiple different , active sites , and individual toxins are rarely tested for all possible activities or substrates , so it is difficult to predict protein activities or to determine if misclassifications are occurring with predictive methods based solely on sequences similarities [37] . Nevertheless , sequence similarity clustering did successfully identify crotoxin and Mojave toxin homologs , PLA2s that are associated with serious neurotoxic envenomation symptomology , in known and previously uninvestigated venoms . Two 3FTx transcripts were discovered in Boiga cynodon venom that were very similar in sequence to the two subunits of the heterodimeric , prey-specific iriditoxin from B . irregularis venom , indicating the presence of another lizard and bird-specific neurotoxin within the venom of a closely related species . Full-length venom protein transcripts obtained from venom can therefore be used to screen for particular toxins or venom phenotypes . As more full-length transcripts become available , high throughput methods such as next-generation proteomic ( and transcriptomic ) characterization of venoms that lack profiles , including most rear-fanged snake species and many understudied front-fanged snake species , will be greatly facilitated . The methods described here provided full-length venom protein transcripts from venoms representing the three major families of venomous snakes , making it is possible to determine snake venom genotype-phenotype relationships without the need to sacrifice living snakes . By requiring only venom to obtain venom protein cDNAs , the approaches detailed here will provide access to cDNA-based protein sequences in the absence of living specimens , from commercial and other venom sources , and will facilitate study of snake venom protein composition and evolution , and in turn , provide greater predictability of the development of regionally-specific reactions following snakebite envenomation .
|
This work demonstrates that full-length venom protein messenger RNAs are present in secreted venoms and can be used to acquire full-length protein sequences of toxins from both front-fanged ( Elapidae , Viperidae ) and rear-fanged ( Colubridae ) snake venoms , eliminating the need to use venom glands . Full-length transcripts were obtained from venom samples that were fresh , newly lyophilized , old , field desiccated or commercially prepared , representing a significant advance over previous attempts which produced only partial sequence transcripts . Transcripts for all major venom protein families ( metalloproteinases , serine proteases , C-type lectins , phospholipases A2 and three-finger toxins ) responsible for clinically significant snakebite symptoms were obtained from venoms . These sequences aid in the identification and characterization of venom proteome profiles , allowing for the identification of peptide sequences , specific isoforms , and novel venom proteins . The application of this technique will help to provide venom protein sequences for many snake species , including understudied rear-fanged snakes . Venom protein transcripts offer important insights into potential snakebite envenomation profiles and the molecular evolution of venom protein multigene families . By requiring only venom to obtain venom protein cDNAs , the approach detailed here will provide access to cDNA-based protein sequences from commercial and other venom sources , facilitating study of snake venom protein composition and evolution .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"sequencing",
"techniques",
"medicine",
"and",
"health",
"sciences",
"toxins",
"pathology",
"and",
"laboratory",
"medicine",
"split-decomposition",
"method",
"vertebrates",
"animals",
"toxicology",
"toxic",
"agents",
"multiple",
"alignment",
"calculation",
"reptiles",
"forms",
"of",
"dna",
"rna",
"isolation",
"molecular",
"biology",
"techniques",
"dna",
"research",
"and",
"analysis",
"methods",
"venoms",
"sequence",
"analysis",
"sequence",
"alignment",
"molecular",
"biology",
"complementary",
"dna",
"snakes",
"biochemistry",
"biomolecular",
"isolation",
"nucleic",
"acids",
"computational",
"techniques",
"squamates",
"genetics",
"biology",
"and",
"life",
"sciences",
"rattlesnakes",
"amniotes",
"organisms"
] |
2016
|
Full-Length Venom Protein cDNA Sequences from Venom-Derived mRNA: Exploring Compositional Variation and Adaptive Multigene Evolution
|
RuvAB and RuvABC complexes catalyze branch migration and resolution of Holliday junctions ( HJs ) respectively . In addition to their action in the last steps of homologous recombination , they process HJs made by replication fork reversal , a reaction which occurs at inactivated replication forks by the annealing of blocked leading and lagging strand ends . RuvAB was recently proposed to bind replication forks and directly catalyze their conversion into HJs . We report here the isolation and characterization of two separation-of-function ruvA mutants that resolve HJs , based on their capacity to promote conjugational recombination and recombinational repair of UV and mitomycin C lesions , but have lost the capacity to reverse forks . In vivo and in vitro evidence indicate that the ruvA mutations affect DNA binding and the stimulation of RuvB helicase activity . This work shows that RuvA's actions at forks and at HJs can be genetically separated , and that RuvA mutants compromised for fork reversal remain fully capable of homologous recombination .
DNA replication and recombination are two processes that are now recognized as more closely connected than originally suspected . It is well documented that replication defects induce the formation of recombination substrates , such as double-stranded DNA ends or single-stranded DNA regions ( ssDNA ) . Depending on the nature of the replication defect , such recombinogenic DNA structures form at blocked replication forks and/or behind forks , on the newly replicated daughter chromatids ( [1]; reviewed in [2]–[4] ) ( Figure 1A ) . In addition , replication and recombination can be directly coupled by enzymes that recognize two different targets , one specifically produced during replication and the other during recombination . The best-documented example is the bacterial PriA protein , which promotes replication restart ( i ) independently of recombination by its virtue of recognizing replication forks and ( ii ) during double-stranded DNA end recombinational repair by its virtue of recognizing D-loop structures ( reviewed in [5] ) . Another example is the RuvAB complex , originally identified for its activity on Holliday junctions ( HJs ) , four-DNA arm recombination intermediates ( reviewed in [6] , [7] ) , and recently proposed to also act on inactivated replication forks [8] ( Figure 1 ) . RuvA and RuvB are nearly ubiquitous bacterial proteins , with a well-conserved structure and function in distantly related species [9]–[11] . During homologous recombination , a RuvA tetramer binds a HJ formed by RecA-catalyzed strand exchange and two RuvB hexamers assemble on two opposite arms of the HJ to form the tripartite RuvAB-HJ complex . RuvB belongs to the AAA+ ( ATPase Associated with various cellular Activities ) family of enzymes and acts as a molecular motor for branch migration . Binding of the dimeric endonuclease RuvC leads to the formation of a RuvABC complex that resolves HJs to produce recombinant molecules . Band shift experiments and structural studies of RuvA complexes with synthetic HJs indicate that two tetramers can eventually assemble to sandwich a HJ and form an octameric RuvA complex [12]–[15] . RuvABC are essential for recombinational repair of DNA lesions in bacteria , and the ruvA ruvB operon is induced by DNA damage , via the SOS response [16] . In addition to its crucial role in processing HJs during homologous recombination , RuvAB binds fork structures in vitro [17]–[19] , and was recently proposed to act at certain inactivated replication forks in vivo [8] . Indeed , inactivated replication forks that occur in several replication mutants are converted into HJs by the annealing of newly synthesized leading and lagging strand ends , a reaction called replication fork reversal ( RFR ) [20]; reviewed in [3] , [4] ( Figure 1A ) . HJs formed by RFR , as those formed by homologous recombination , are resolved by RuvABC . Notably , RuvAB was shown to be essential for the formation of HJs at blocked forks in some replication mutants , including the dnaEts mutant affected for the catalytic subunit of the main E . coli DNA polymerase Pol III . We proposed that RuvAB binds to certain inactivated replication forks and catalyzes their conversion into HJs [8] ( Figure 1B ) . As the two functions of RuvAB in E . coli , resolution of HJs and RFR , involve interactions with two different target molecules , we searched for mutants that have lost only one of these functions . We describe here the isolation and characterization of two ruvA mutants that still promote homologous recombination while they have lost the capacity to reverse forks .
The ruvA100::CmR mutant is sensitive to UV irradiation , and UV resistance is restored in the presence of pGB-RuvA+ while the pGB2 vector has no effect ( Table 1 ) . The UV sensitivity of the ruvA , ruvB and ruvC mutants results from the lack of resolution of recombination intermediates , therefore reflects the recombination defect of these mutants [2] , [21] . The products of a ruvA mutagenic PCR were cloned in pGB2 , used to transform the ruvA100 mutant , and the UV sensitivity of cells carrying recombinant plasmids was monitored . Seventeen clones were tested , nine remained UV sensitive , therefore contained a plasmid unable to complement the UV repair defect of the ruvA mutant . The remaining eight recombinant plasmids that carry a ruvA allele functional for UV repair were isolated and tested for their capacity to promote RFR in a dnaEts mutant ( dnaE486ts , Table S1 ) . The dnaE486ts mutant is completely defective at 42°C , and partially affected at 37°C for the main E . coli DNA polymerase , Pol III . Because its slight growth defect at 37°C is suppressed by preventing SOS induction [22] , recF derivatives of dnaEts were used for the screening of RFR deficient ruvA mutants . RFR takes place at dnaEts-blocked forks and renders RecB essential for viability . Consequently , growth of a dnaEts recF mutant at 37°C is prevented by recB inactivation [23] . However , because RFR requires RuvAB , inactivation of ruvA or ruvB restores the growth of dnaEts recB recF mutants and introduction of a functional ruvA gene in a dnaEts recF recB ruvA100 mutant is lethal [8] ( Table 1 ) . Three of the plasmids conferring UVR to a ruvA100 mutant were lethal in dnaEts recF recB ruvA100 cells at 37°C , therefore presumably carried a wild-type ruvA gene . The other five plasmids allowed variable levels of viability and therefore expressed candidate RFR-defective RuvAs . To ascertain whether the ruvA alleles in these plasmids were deficient for RFR , fork breakage was measured directly . In dnaEts recB mutants , resolution of HJs formed by RFR leads to an increase in the level of linear DNA in vivo , which can be quantified by pulse field gel electrophoresis ( PFGE ) , as only linear DNA can enter PFG [20] . Because RuvAB promotes fork reversal and RuvABC resolves the resulting HJ , the level of linear DNA resulting from fork breakage is high in dnaEts recF recB cells ( ∼60% ) and low in the dnaEts recF recB ruvA100 mutant ( ∼10% ) [8] ( Table 1 ) . As expected , fork breakage was increased in dnaEts recF recB ruvA100 cells by the presence of pGB-RuvA+ ( Table 1 ) . Fork breakage remained low in the presence of 4 candidate plasmids: 12–16% for pGB-ruvAz60 and pGB-ruvAz80 , 32–38% for pGB-ruvAz26 and pGB-ruvAz87 ( Table 1 ) . These ruvA alleles can promote HJ resolution , since they fully complement the UV sensitivity of a ruvA null mutant . Therefore , the defect in fork breakage suggests that they are affected for fork reversal . Sequencing of ruvA on pGB-ruvAz26 and pGB-ruvAz87 showed that these were double mutants ( E68G H136R and N79D N100D , respectively ) . pGB-ruvAz60 and pGB-ruvAz80 carried 7 mutations each , 6 of which were identical ( Table 1 ) . This result was not surprising since all plasmids derived from the same PCR cloning experiment . In order to identify the mutations in ruvAz60 which are necessary and sufficient to abolish RFR , the 7 mutations were introduced individually or in combination on a pGB-ruvA plasmid ( see Supplementary Material ) . The capacity of the ruvA mutant alleles to promote homologous recombination was monitored by measuring the UV resistance that they confer to ruvA100 cells . Their inability to catalyze RFR was deduced from the viability and the low level of fork breakage that they confer to dnaEts recF recB ruvA100 mutants ( Table 2 ) . Sub-cloning of different ruvA gene regions and site-directed mutagenesis showed that three mutations were necessary and sufficient for the RFR defect ( H29R K129E F140S , pGB-ruvAz3 , Table 2; Figure 2 ) . pGB-ruvAz3 and pGB-ruvAz87 ( N79D N100D ) were used for further studies . A recF mutant background was used for the original screening experiment because inactivation of recF improves the viability of dnaEts recB ruvA cells and we have shown that the recF mutation has no effect on RFR in dnaEts cells [8] . As expected , the ruvA alleles identified as deficient for RFR in a recF null background were also unable to promote fork breakage in the presence of RecF ( dnaEts recB ruvA100 , pGB-ruvAz3 and pGB-ruvAz87 , Table 3 ) . A RecF+ context was therefore used for the subsequent experiments . Although ruvA and ruvB genes form an operon , ruvB is expressed in the ruvA100::CmR mutant , as only wild-type RuvA protein is required for suppression of the recombination defects ( Table 1 ) . In dnaEts recB ruvA100 [pGB-ruvA] mutants , ruvB is expressed from the chromosomal locus downstream of the ruvA100::CmR insertion , whereas ruvA is expressed from its own promoter on the plasmid which has about 10 copies per cell . The imbalance between ruvA and ruvB expression could play a role in the RFR defect conferred by ruvAz3 and ruvAz87 mutations . To test this possibility , we cloned ruvB downstream of these ruvA alleles on the pGB-ruvAz3 and pGB-ruvAz87 plasmids . Co-expression of ruvB restored a high level of breakage in dnaEts recB ruvA100 cells expressing ruvAz3 or ruvAz87 ( Table 3 ) . In addition , a high level of breakage was observed when the ruvAz60 allele was inserted into the chromosome ( Table 3 ) . These observations indicate that the mutant RuvAz-RuvB complexes are defective for RFR only if the ruvB gene is expressed from a single chromosomal copy downstream of the ruvA100::CmR mutation . The insertion of the CmR gene in ruvA most likely reduces the amount of RuvB protein synthesized . The capacity of ruvAz alleles to catalyze homologous recombination was analyzed by different assays . Mitomycin C is a DNA damaging agent that causes various DNA lesions and mitomycin C treatment prevents growth of a ruvA null mutant , defective for recombinational repair [24] ( Figure 3A ) . Introduction of pGB-ruvAz3 or -ruvAz87 plasmids in the ruvA100 mutant restored the same level of resistance to mitomycin C treatment as pGB-RuvA+ ( Figure 3A ) , indicating that these mutant ruvA alleles promote recombinational repair of mitomycin C lesions . Inactivation of ruvA decreased conjugational recombination about 3–5 fold [25] ( Figure 3B ) ; this defect was also suppressed by the ruvAz3 or ruvAz87 alleles ( Figure 3B , RecG+ ) . Full suppression of the UV sensitivity of single ruvA100 mutants by pGB-ruvAz3 and -ruvAz87 plasmids was observed at a wide range of UV doses ( Figure 3C ) , and at 42°C , the temperature used for fork-breakage measurements ( data not shown ) . The UV resistance conferred by these alleles was also tested in different mutant backgrounds . The ruvA100 mutation decreases the survival of UV-irradiated recR mutants deficient for the recombinational repair of gaps ( Figure 3D ) . Introducing pGB-ruvAz3 or -ruvAz87 plasmids fully suppressed the UV-repair defect caused by the ruvA100 mutation in a ruvA100 recR double mutant ( Figure 3D ) . recG inactivation affects the viability of ruvA , ruvB or ruvC mutants and renders them extremely deficient for homologous recombination [25] ( Figure 3F ) . It was proposed that RecG provides an alternative way of resolving HJs in vivo [26] . Expression of ruvAz3 or ruvAz87 in a ruvA100 recG double mutant suppressed the viability defect ( not shown ) and the sensitivity to UV irradiation ( Figure 3E ) . Accordingly , conjugational recombination was not significantly different in ruvA100 recG mutants carrying pGB-RuvA+ , -ruvAz3 or -ruvAz87 ( Figure 3B , recG− ) . These findings indicate that the mutant Ruv proteins promote HJ resolution in a recG context . RusA is a HJ resolvase that is only expressed in E . coli when the rusA ORF is activated by the insertion of an upstream IS element ( rus-1 mutant , [27] ) . RusA resolves HJs in vitro but is devoid of detectable branch migration activity [28] . By allowing RusA resolvase synthesis , the rus-1 mutation suppresses the recombination defects of ruvA mutants in vivo . However , suppression is partial in ruvC mutants in which RuvA protects HJs from RusA action [27] . As shown in Figure 3F , ΔruvABC rus-1 cells were resistant to UV irradiation as expected ( HJs are resolved by RusA ) , and expression of wild-type RuvA from pGB2-RuvA+ made them UV sensitive ( RusA-catalyzed resolution is prevented by RuvA binding to HJs ) . In contrast with pGB-RuvA+ , plasmids carrying the ruvAz3 or ruvAz87 allele did not compromise the survival of ruvA100 rus-1 UV-irradiated cells , suggesting that these mutant RuvA proteins are not capable of protecting recombination intermediates from resolution by RusA ( Figure 3F ) . Co-expression of ruvB and ruvAz87 only slightly prevented RusA action , suggesting that the HJ-binding defect of the RuvAz87 protein is mainly independent of the amount of RuvB . In contrast , co-expression of ruvB and ruvAz3 fully prevented RusA action ( Figure 3F ) . This observation indicates that the HJ-binding defect of RuvAz3 can be suppressed by increasing the cellular level of RuvB , suggesting a defect in RuvAz3-RuvB protein interactions . Expression of RusA allows HJ resolution in dnaEts recB ruvA100 rus-1 , but , because RuvA is required for RFR , the level of linear DNA remains significantly lower in the absence of RuvA than in its presence ( compare JJC4196 containing pGB2 and pGB-RuvA+ , Table 3; Baharoglu et al , 2006 ) . RuvAz3 and RuvAz87 did not restore a high level of linear DNA in the presence of RusA , unless these ruvAz alleles were made capable of RFR by the presence of ruvB on the plasmid ( compare JJC4196 containing pGB2-ruvAz and pGB-RuvA+ , Table 3 ) . This observation confirms that the fork breakage defect in the presence of RuvAz3 or RuvAz87 does not result from a defect in HJ resolution , but rather from a defect in HJ formation . A slight increase in the percentage of DNA entering PFG is observed upon RusA expression ( compare JJC3723 and JJC4196 containing pGB2 , Table 3 ) . This may result , at least in part , from RusA-resolution of HJs made behind replication forks by recombination at gaps , which would preventing linear DNA migration if left unresolved [8] , [29] . Wild-type RuvA and mutant RuvAz3 and RuvAz87 proteins were over-expressed in E . coli and purified . Both mutants behaved as wild type at all chromatographic steps during purification . The oligomeric states of the mutant RuvA proteins were analyzed by SDS-PAGE without boiling the protein samples [30] . Both RuvAz3 and RuvAz87 showed a tetramer band under denaturing gel conditions ( data not shown ) , which indicates that the structural organization of the RuvA mutants was not affected by the mutations . Binding to a substrate that mimics a Holliday junction ( X12 ) was measured by electrophoretic mobility shift assays ( EMSA ) . In the presence of EDTA , wild-type RuvA formed two complexes: complex I , which contains one bound RuvA tetramer per HJ and complex II , in which the junction is sandwiched between two tetramers [31] , [32] ( Figure 4A ) . The proportion of DNA bound by RuvAz3 and RuvAz87 proteins was only slightly lower than that with the control wild-type RuvA protein . However , both mutant RuvA proteins could only form complex I ( Figure 4A ) . In the presence of Mg2+ , where only octameric complexes can be observed [31] , RuvAz3 was partially and RuvAz87 totally unable to promote band-shifts ( Figure 4B ) . Therefore , both proteins are slightly affected for binding to HJ DNA , and strongly affected for the formation and/or the stability of octamers on the junction . The mutant RuvA proteins were tested for branch migration of ×12 in the presence of wild-type RuvB protein ( Figure 4C ) . Both mutants were able to support branch migration of ×12 but at significantly higher concentrations than wild type RuvA . RuvAz87 exhibited a relatively high branch migration activity with RuvB , while RuvAz3 was more defective in branch migration . Therefore RuvB compensates for the binding defect of RuvAz87 but only partially for that of RuvAz3 . These results indicate that whereas RuvAz87 is more affected than RuvAz3 for HJ binding , RuvAz3 is more affected for RuvB binding and/or activation . To confirm this idea , RuvB helicase activity was compared in the presence of wild-type and mutant RuvA proteins using the property of RuvA to stimulate RuvB in a classical helicase assay , in which RuvB displaces an oligonucleotide annealed to a ssDNA circular molecule ( Figure 5 ) . Both RuvA mutant proteins were capable of RuvB stimulation , but whereas 100 nM RuvA were required for RuvB to unwind 50% of the annealed oligonucleotide , 150 nM of RuvAz87 and 250 nM of RuvAz3 were required . Therefore , RuvAz3 was more deficient than RuvAz87 for the stimulation of the RuvB helicase activity . To test the binding of the mutants to DNA substrates that mimic replication forks , band shift experiments were performed with an entirely double-stranded fork ( F2 ) or a partially single-stranded fork ( F1 ) . Both RuvA mutant proteins were completely defective for fork binding , both in EDTA and in Mg2+ buffer ( Figure 6A and 6B , data not shown ) . The single-strand binding protein SSB , which covers ssDNA regions at replication forks in vivo , did not stimulate binding of either wild-type or mutant RuvA proteins to forked DNA with an appropriate length of ssDNA regions ( not shown ) . Although it did not act on F1 , RuvB unwound F2 in the presence of RuvA ( Figure 6C ) . This reaction was very inefficient in the presence of RuvAz3 and RuvAz87 compared to the wild-type RuvA control ( Figure 6C ) . In conclusion , the two mutated proteins are deficient for fork binding and allow only a weak RuvB action on fork structures . Altogether , these in vitro experiments indicate that both mutated proteins are particularly deficient for binding to forked DNA substrates . In addition , RuvAz87 is weakly , and RuvAz3 more strongly affected for RuvB activation .
In this work , we isolated and characterized two ruvA mutants that are fully capable of resolving HJs made by homologous recombination after UV irradiation , mitomycin C treatment and Hfr conjugation , while they do not reverse forks at dnaEts-blocked forks . In agreement with the mutant strains Rec+ phenotype , purified mutant proteins bind HJs nearly as efficiently as wild-type protein . The in vivo RFR defect mainly correlates with in vitro defects in RuvA octamer formation on HJs and binding to forks . In addition , RuvAz3 is affected for RuvB helicase activation in vitro , which could be a major cause of the ruvAz3 mutant defects in vivo , as these defects can be suppressed by over-producing RuvB . The RuvA polypeptide consists of three distinct domains , I , II and III . The major core domains , I ( residues 1–64 ) and II ( residues 65–140 ) , form the central part of the RuvA tetramer and provide a platform for DNA binding . Domain III ( residues 156–203 ) , which is linked to domain II by a flexible linker , is involved in RuvB contact and branch migration [10] , [15] , [33] , [34] . Both mutations in ruvAz87 are in domain II and are very likely to affect primarily DNA binding: N100D lies between two helix-hairpin-helix structures that contact DNA and N79D is within the first of these structures [10] , [15] ( Figure 2 ) . Mutations in ruvAz3 affect the three domains ( Figure 2 ) . H29R lies within domain I , in a region thought to be involved in RuvA interactions within the tetramer and in DNA binding [33] . K129E is at the end of the second helix-hairpin-helix in domain II , in a region involved in the association of two tetramers to form RuvA octamers [14] , [18] . Finally , F140S is the last residue before the flexible linker and may affect the positioning of domain III within the RuvAB complex [35] . Previous studies of RuvA mutants indicated that a combination of three mutations at residues 122 , 127 and 130 , which disrupt RuvA tetramer-tetramer interface , inactivated recombinational repair in vivo [18] . Although the purified proteins do not form octamers on HJs in vitro , both ruvAz3 and ruvAz87 mutants remain capable of recombinational repair , which suggests that either these mutant proteins form octamers in vivo , or octamer formation is not a pre-requisite for homologous recombination . It should be noted that the defect in RuvAz3 and RuvAz87 octamer formation in vivo may be responsible for the lack of protection of recombination intermediates from RusA , and may play a role in the RFR defect . In the complex with a HJ , a RuvA tetramer contacts four double-stranded DNA arms and two RuvB hexamers [34] . In the model that we propose , when RuvAB binds to a replication fork to initiate RFR only three RuvA polypeptides in the tetramer are engaged in DNA contacts , including one with single-stranded DNA , and only one RuvB hexamer is present [8] ( Figure 1B ) . Such a complex might be intrinsically unstable , so that mutant proteins with a decreased DNA affinity would , as the RuvAz mutants described here , retain the ability to bind HJs but lose fork binding . A second RuvA tetramer sandwiching the junction could strengthen interactions with both the fork and the RuvB hexamer . In this case , the defects in octamer formation , fork binding and RuvB activation observed in vitro would all contribute to the RFR defect in vivo . One of the forces driving the evolution of RuvA , among others such as recombination between diverged sequences or recombination with small DNA fragments , might be to promote RFR . Consequently , RuvA could have acquired a capacity to bind DNA and interact with RuvB exceeding the needs of conjugational recombination and lesion recombinational repair . Our observation that mutations that affect RuvA activity do not necessarily inactivate homologous recombination may explain why , when a collection of 40 ruvA mutants was made by alanine replacement of conserved residues , 34 mutants conferred a normal level of UV resistance and 6 only showed a recombinational repair defect [33] . Although RuvAB are among the best-conserved recombination proteins in prokaryotes [9] , they do not have close homologues in eukaryotes . The nature of the enzymes that catalyze HJ branch migration and/or resolution in eukaryotes is a subject of debate , possibly because different activities can be involved , depending on the organism and on whether meiotic or mitotic recombination is considered . Rad51 is the functional and structural homologue of RecA in eukaryotes and two mammalian Rad51 orthologues , named Rad51C and XRCC3 , were identified as components required for coupled HJ branch migration and resolution in cell extracts [36] , [37] . In addition , several purified proteins could catalyze HJ branch migration in vitro: the RecQ helicase family members BLM and WRN [38]–[40] and Rad54 [41] . In yeast , the Mus81-Eme1 complex from Schizosccharomyces pombe catalyzes the resolution of synthetic HJs in vitro and is thought to resolve meiotic recombination intermediates ( [42] , and references therein ) . Interestingly , proteins that act on Holliday junctions are most often able to target alternative structures , at least in vitro . Mus81 , which cleaves nicked HJs , also cleaves fork and D-loop structures ( reviewed in [43] ) . BLM and WRN helicases , which displace HJs , also unwind fork and D-loop structures ( [44]; reviewed in [45] ) . Similarly , the E . coli RecG protein promotes HJ branch migration and unwinds D-loops , R-loops and forks [46]–[48] . Recently , the yeast Rad5 protein was shown to promote fork reversal in vitro and this reaction may account for the physiological role of Rad5 during post-replicative repair of UV lesions [49] . Fork reversal by Rad5 did not require RPA and did not involve a single-stranded DNA intermediate , which may well be the case with RuvAB . The mammalian BLM and WRN proteins and the bacterial RecG protein were also shown to be able to convert fork structures into HJs in vitro [50]–[53] . However , when and where exactly these reactions take place in vivo remains to be determined . It is tempting to speculate that RFR in E . coli replication mutants involves interactions of RuvA or RuvB with replication fork-associated proteins . Indeed , several proteins that act at replication forks were shown to interact with fork-associated proteins such as SSB [54] or the polymerase clamp ( reviewed in [55] ) . In addition , RuvB is closely related to clamp-loader subunits , i . e . the DNA Pol III δ′ subunit and the replication factor C in eukaryotes [11] , [56] , and also homologous to RarA ( Mgs1 in yeast ) , a universally conserved protein associated with replication forks [29] , [57] , [58] . In conclusion , this work shows that it is possible to genetically separate the two functions of the RuvAB complex , RFR and branch migration/resolution of homologous recombination intermediates . Mutations that weaken the function of RuvA inactivate only RFR , possibly because this reaction is more demanding than HJ branch migration .
Strains were constructed by classical P1 transduction [59] and are described in Table S1 . Details of constructions are described in supporting information material . Sequencing of ruvA genes in plasmids and chromosome was performed using “Genetic Analyzer” 3100 ( Applied Biosystem ) automatic sequencer . Oligonucleotides used for sequencing are shown in the supplementary material . Plasmid constructions are described in supplementary material . For the construction of the mutagenic pGB-ruvAm pool , a protocol derived from Fromant et al , was used [60] . ruvA was amplified using a mutagenic PCR reaction containing 10 mM dGTP , 10 mM dCTP , 10 mM dTTP , 2 mM dATP , 5 mM MnCl2 and ExTaq ( Takara ) polymerase . A first denaturation step at 94°C for 10 min was followed by 25 cycles of denaturation at 94°C for 30 s , annealing at 55°C for 30 s , elongation 72°C for 3 min 30 s . The PCR product was purified using Qiagen PCR purification kit and cloned in pGB2 as described . Separations of mutations in pGB-ruvAz60 were done as described in supplementary material . Wild-type RuvA and RuvB proteins were purified as described in previous studies [18] . Some steps were modified or added for the purification of the mutant RuvA proteins , as described in supplementary material . The concentration of RuvB was determined by absorbance at 280-nm wavelength using an extinction coefficient of OD280 , native = 16 , 900 M−1 cm−1 [61] . All of the other protein concentrations were determined by the Bradford method using the protein assay reagent from Bio-Rad with BSA as a standard . 5′ IRD700-labelled and unlabelled oligonucleotides were purchased from MWG . Oligonucleotide sequences and oligonucleotide assembly are shown in Table S2 . Annealing reactions and substrate purification were performed as described [62] , using 2 µg of unlabelled and 1 µg of labeled oligonucleotides in each reaction . The helicase substrate was obtained by annealing 10 pmol of IRD-700-labelled 52-mer IT . 300 to 10 pmol φx174 virion ssDNA ( NEB ) in 10 mM Tris-HCl pH 7 . 5 , 10 mM MgCl2 , 50 mM NaCl [63] . The mixture was denatured at 100°C for 3 min , incubated at 68°C for 30 min and slowly cooled down at room temperature . Purification was performed on a 5–20% sucrose gradient and fractions collected after centrifugation at 4°C , 45000 rpm for 3 h . Substrates were visualized and quantified using the Li-cor Biosciences ODYSSEY infrared imaging system . Binding reactions in conditions without magnesium were performed as described [18] and analysed by PAGE in 0 . 5× Tris-borate EDTA buffer . For binding assays in the presence of magnesium , 3 mM MgCl2 was added to the reaction buffer , which did not contain EDTA . Electrophoresis was performed in 0 . 5× Tris-borate buffer supplemented with 200 µM MgCl2 and using buffer recirculation . Reaction products were analyzed using the Li-cor Biosciences ODYSSEY infrared imaging system . Branch migration reactions ( 20 µl final volume ) were performed in 20 mM TrisHCl pH 7 . 5 , 2 mM ATP , 2 mM DTT , 100 µg/ml BSA , 1 . 5 mM MgCl2 . The reactions contained ∼5 ng labeled synthetic junction with 250 nM RuvB and various concentrations of RuvA protein . Proteins were diluted in 20 mM Tris HCl pH 7 . 5 , 1 mM EDTA , 0 . 5 mM DTT , 100 µg/ml BSA , 150 mM NaCl , 10% glycerol . Branch migration reactions were performed as described [18] and visualized as described above . DNA helicase reactions were performed under the same conditions as branch migration assays described above . Reaction products were analyzed by electrophoresis in 1 . 2% agarose gel in 1× TAE buffer and visualized as described above . UV irradiation was performed as described [8] . For mitomycin C treatment , cells were grown at 37° in LB to an OD600 = 0 . 5 C , mitomycin C was added to the culture at a final concentration of 2 µg/ml and incubation continued at 37°C for 90 minutes . An untreated culture was used as control . Appropriate dilutions were plated on LB plates and incubated over-night at 37°C . Ratios of cfu of mitomycin C treated over cfu of untreated cells were calculated . Conjugations were performed as described using JJC145 as Hfr donor [29] , donor and recipient cells were mixed for 25 min . Selective medium was M9 minimal medium supplemented with leucine , proline , threonine and arginine ( 2% final concentration each ) and 10 µg/ml Cm . Quantification of pulsed field gels was performed using in vivo 3H-thymidine labeled chromosomes as previously described [20] .
|
DNA replication is the process by which DNA strands are copied to ensure the transmission of the genetic material to daughter cells . Chromosome replication is not a continuous process but is subjected to accidental arrests , owing to the encounter of obstacles or to the dysfunctioning of a replication protein . In bacteria , inactivated replication forks restart but they are most often remodeled before restarting . Interestingly , enzymes involved in homologous recombination , the process that rearranges chromosomes , are also involved in fork-remodeling reactions . The subject of the present study is RuvAB , a highly conserved bacterial complex used as the model enzyme for resolution of recombination intermediates , which we found to also act at blocked forks . We describe here the isolation and characterization of ruvA mutants that have specifically lost the capability to act at inactivated replication forks , although they remain fully capable of homologous recombination . The existence of such ruvA mutants , their properties and those of the purified RuvA mutant proteins , indicate that the action of RuvAB at replication forks is more demanding that its action at recombination intermediates , but have nevertheless been preserved during evolution .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
] |
[
"molecular",
"biology/dna",
"replication",
"molecular",
"biology/recombination",
"biochemistry/protein",
"chemistry",
"genetics",
"and",
"genomics/chromosome",
"biology",
"biochemistry/replication",
"and",
"repair",
"genetics",
"and",
"genomics/gene",
"function",
"microbiology/microbial",
"growth",
"and",
"development",
"molecular",
"biology/dna",
"repair"
] |
2008
|
ruvA Mutants That Resolve Holliday Junctions but Do Not Reverse Replication Forks
|
Upon cell adhesion , talin physically couples the cytoskeleton via integrins to the extracellular matrix , and subsequent vinculin recruitment is enhanced by locally applied tensile force . Since the vinculin binding ( VB ) sites are buried in the talin rod under equilibrium conditions , the structural mechanism of how vinculin binding to talin is force-activated remains unknown . Taken together with experimental data , a biphasic vinculin binding model , as derived from steered molecular dynamics , provides high resolution structural insights how tensile mechanical force applied to the talin rod fragment ( residues 486–889 constituting helices H1–H12 ) might activate the VB sites . Fragmentation of the rod into three helix subbundles is prerequisite to the sequential exposure of VB helices to water . Finally , unfolding of a VB helix into a completely stretched polypeptide might inhibit further binding of vinculin . The first events in fracturing the H1–H12 rods of talin1 and talin2 in subbundles are similar . The proposed force-activated α-helix swapping mechanism by which vinculin binding sites in talin rods are exposed works distinctly different from that of other force-activated bonds , including catch bonds .
Talin physically links integrins to the contractile cytoskeleton [1 , 2] . The talin head ( TH ) , residues 1–432 , has binding sites for integrin β-tails [3] , PIP kinase γ [4] , focal adhesion kinase ( FAK ) [5] , layilin [5] and actin [6] ( Figure 1B ) . Binding of the talin head to the cytoplasmic tail of β-integrins can cause integrin activation [7] . The 60 nm long talin rod ( TR ) , residues 433-2541 , is composed of bundles of amphiphatic α-helices [8 , 9] ( Figure 1 ) . The talin rod contains up to eleven vinculin binding sites [10] ( Figures 1B and S1 ) , including five located within fragment H1–H12 , residues 486–889 , which is studied here ( Figure 1 ) . All of these five binding sites are buried inside helix bundles ( native talin shows considerably lower affinity to vinculin compared to peptide fragments isolated from talin ) . In addition to the VBSs , the talin rod has binding sites for actin [11] and for integrins [12] . Upon cell adhesion , talin rapidly accumulates in focal contacts prior to vinculin recruitment [13] . Only talin , but not vinculin or paxillin , is recruited to the clustered integrins , if β3 integrins are activated not by ligand binding but by manganese ( Mn2+ ) , i . e . , in cases where integrin activation occurs without the application of force and integrin is thus not part of a force-bearing protein network [14] . Indeed , the recruitment of vinculin to cell adhesion sites has been shown to be force-dependent [15–19] . Vinculin recruitment strengthens cell adhesions [20] , and reduces focal adhesion turnover [21] . In focal adhesions , vinculin is in an activated state while it is autoinhibited in the surrounding cytoplasm [22] . Important to the model proposed here is that vinculin activation can also be induced in solution by isolated talin peptides containing VBSs [23] . As talin , vinculin is an alpha-helical protein sharing considerable structural similarity . The x-ray structure of full-length vinculin [PDB: 1TR2] reveals three head domains , VH , VH2 , and VH3 , that are connected to the tail domains , VT2 and VT , via a flexible , proline-rich linker . Vinculin's talin binding site is located in the VH head domain , which is composed of two four-helix bundles sharing one helix [24] ( Figure 1A ) . Since VH can bind to VT , thereby confining the molecule to an autoinhibitory closed conformation , vinculin activation is needed to increase its affinity to talin and actin [25 , 26] . While it was suggested that vinculin recruitment to focal adhesions is force-regulated [15–19] , no high resolution structural mechanism has been derived experimentally by how force can activate talin's buried vinculin binding sites . What is known , however , is that fragmentation of the talin rod into smaller isolated helix bundles can activate talin's VBS [27] , and that point mutations that destabilize helix-helix interactions can have a similar activating effect [27] . Finally , vinculin binding to talin's VBS can cause significant loss of secondary structure of the adjacent non-VB-helices [28 , 29] . Alternatively , it has been hypothesized that binding of PIP2 lipids to the talin rod may also lead to the activation of VB sites as the presence of PIP2 slightly increased the binding of vinculin to talin [30] . No mechanism has been suggested though how activation by PIP2 could directly explain force-upregulated vinculin recruitment to newly formed cell adhesion sites . To establish a potential mechanism of force-regulated activation of talin's VBSs , we used steered molecular dynamics ( SMD ) to study the force-induced conformational changes in the N-terminal talin rod fragment H1–H12 ( a model created by merging two experimentally determined overlapping structures covering residues 486–889 [28] ) . This computational method is valuable since no experimental techniques are available to obtain high resolution information how the structure of proteins is changed when they are mechanically stretched . Access to such information is essential to learn how force might activate proteins by switching their structure-function relation ( as reviewed in [31 , 32] ) . Therefore , known equilibrium structures of the talin rod were solvated here computationally in a box filled with explicit water molecules . Conducting the simulations in the presence of explicit water rather than in an unstructured dielectric medium is important , since the access of free water molecules to often buried force-bearing hydrogen bonds or salt bridges can regulate mechanical stability [31–34] . After equilibration for 1 ns , constant force was applied to either the termini of the rod fragment , or alternatively to putative force-bearing interfaces which might stabilize this rod fragment in the intact talin molecule . We then characterized the structural changes that lead to the sequential exposure of the vinculin binding sites . A mechanism of how stretching of the talin rod might activate vinculin binding is proposed .
When applying various constant forces locally to the terminal atoms of the H1–H12 bundle ( 100 pN , 200 pN , 300 pN , and 400 pN ) , the intial end-to-end distance of the talin rod H1–H12 of 3 . 2 nm increases rapidly with time due to progressive unfolding ( loss of secondary structure ) of the terminal helices ( Figure 2B ) . The ease by which they unravel even at low forces indicates that relatively little energy is required to sequentially break the backbone hydrogen bonds , turn-by-turn , that stabilize their helical secondary structure [39 , 40] . As shown in Figure 2A , a first short-lived plateau can be seen that has a lifetime of less than 1 ns if we pull at 100 pN ( Intermediate State I1 ) . Just little activation is then needed to begin the unfolding of both of the terminal helices , H1 and H12 , which leads to an extension with respect to the resting length of altogether 15 nm , until a second plateau is reached ( Intermediate State I2 ) . Notice though that both helices do not initially unfold completely . Only the N-terminal half of H1 ( residues 499–503 ) is unraveled in state I2 , due to a bent in H1 at Ser502 . The bending site seems to be defined by the side chain of residue Ser502 , which competes for the backbone hydrogen bond between the Ile499 and Met503 ( see Figure 3B for presentation of such an event ) . Similarly for the H12 , which is a vinculin binding helix , its C-terminal residues 874–879 are unraveled while its N-terminal part of the helix remains buried in the helix bundle . Further extension results from the sequential turn-by-turn unraveling of the N-terminal end of H1 ( D2; in all of 8 local force simulations; Figure 2 ) , and the H12 by ∼3 additional turns ( Intermediate State I3 ) which leads to a total extension of 28 nm while pausing in the third plateau in Figure 2A . This intermediate I3 , which is seen to last for ∼15 ns at 200 pN , is characterized by a completely unraveled H1 , whereas residues 868–879 of H12 are unraveled while residues 849–867 of H12 are still in a helical conformation and in tight contact with the rest of the H9–H12 bundle . At higher forces , we see a rapid transitioning into I4 ( multiple unraveled helices ) where both H2 ( ∼8 turns unraveled ) and H12 ( ∼6 turns unraveled ) are mostly unraveled . I4 is characterized by an extension of 35–36 nm . Once the major energy barrier to rupture I4 is passed ( as described below ) , the remaining bundle H2–H11 rapidly breaks down into the bundles H2–H5 , H6–H8 and H9–H12 ( T4 in Figure 2 ) . Each of these three helix bundles unfolds independently from each other at later times as discussed below . Note that the two vinculin binding helices , H11 and H12 , are already completely unraveled when T4 is reached , i . e . , they convert early on into an extended polypeptide chain held under tension . To assess whether the just described intermediates are characteristic only for an isolated talin fragment where force is applied to the terminal atoms , or also for the case where force is distributed over the cross-section of the talin rod , we ran constant force simulations in which we distributed the force ( 200 pN , 250 pN , 300 pN , 400 pN ) along the Cα-atoms of the terminal helices H1 and H12 as shown in Figure 3 . Additional constant velocity simulations were carried out under otherwise similar conditions , but by applying force distributed along the C-terminal H12 ( see Methods ) and harmonically constraining H1 ( Figure S2 , only stimulation with harmonical constrained helix ) . Distributing the force over a putative interface had a major stabilizing impact on the resulting unfolding trajectories and the forces needed to activate the VBSs ( Figures 2 and 3 ) : The most notable difference is that the talin rod H1–H12 could withstand considerable higher forces before breaking apart ( Figure 3C ) , stably exceeding our simulation window of 20 ns in a prestretched intermediate state I1 ( Figure 3 ) if pulled with a constant force ( 200 pN ) . Earlier , this force had resulted in the unraveling of the terminal helices within the first 5 ns ( Figure 2A ) if force was only applied to the terminal atoms . Distributing the force over the terminal helices thus stabilized the entire H1–H12 bundle against force-induced breakage and terminal helix unraveling: unfolding of H1 was not detected as a first major unfolding event even at higher forces ( 7 simulations , 250–400 pN ) . Plotting the sequence by which hydrogen bonds between the helix bundles broke ( Figure 4 ) reveals that straining the talin rod leads to the appearance of some new hydrogen bonds ( for example bond 8 between residues Asn559 and Glu738 ) while others break ( for example bond 5 between residues Ser658 and Pro725 ) . When H1–H12 is stretched with a constant force of 250–400 pN distributed over the terminal helices , interhelical bundle contacts start to loosen up after a few ns leading to a slight opening of the interface between the bundles H9–H12 and H1–H8 ( I2 ) . This passage leads to the breakdown of some hydrogen bonds in the interface between H1–H8 and H9–H12 , namely between residues Gln887-Ser752 , Gln886-Gln755 , His784-Glu733 , His788-Gln733 ( referred to as bonds 19 , 20 , 21 , 25 in Figure 4 ) . Also the bonding between bundles H1–H5 and H6–H8 is weakened significantly by breakdown of hydrogen bonds between residues Asn559-Gln735 , Ser729-Asp548 , Gln715-Asn559 ( referred to as bonds 1 , 2 , 3 in Figure 4 ) . Further pulling results in a gradual opening of the interface between the bundles and water starts to penetrate between the helices ( Figure 3E ) . The stress applied leads finally to a bending of H9 around residue 770 due to breakage of the backbone hydrogen bond formed between residues Gly768 and Thr772 ( Figure 3B ) . After this , H9 stays attached to H1–H8 on the N-terminal end during the short-lived intermediate state I3 seen in many but not in all the simulations ( 5 of 7 simulations 250–400 pN ) . After passing the I3 intermediate , a rapid disintegration of the talin rod is seen ( T3− ) which then allows for the sequential unfolding of the now separated bundles . Common to all simulations is the break-up of the talin rod , H1–H12 , into smaller helix bundles that have far smaller mechanical stabilities . We observed that the first split occurred between H1–H8 and H9–H12 ( 6 times in 7 simulations where constant force is applied over terminal helices ) , after which H1–H8 splits into H1–H5 and H6–H8 ( Figures 2 , 3 , and 5 ) . In the other case observed only once in seven simulations , H1–H5 separated first followed by the H9–H12 separation . The break-up into these well-defined bundles as a major event is most clearly seen if the force is distributed over the terminal helices ( Figure 3 ) . Once the force-bearing interfaces are broken apart , the bundles have no stabilizing effects upon each other any longer . Consequently , the force is thus transmitted at these later times through the N- and C-terminal atoms of the resulting helix bundles . We thus simulated separately the unfolding of the helix bundles H1–H9 ( crystallographically determined structure of talin rod fragment , PDB 1SJ8 ) , H2–H8 ( intermediate found in terminal atom pulls ) , and H9–H12 ( intermediate seen only in force-bearing interface simulations ) . The starting end-to-end distances of the terminal atoms of these helix bundles prior to stretching were 3 . 6 nm ( H1–H9 ) , 7 . 9 nm ( H2–H8 ) and 1 . 5 nm ( H9–H12 ) . Most noticeable when looking at all their unfolding pathways ( Figure 5 ) is the lack of well pronounced plateaus and thus clearly distinguishable intermediate states , especially in the case of H9–H12 . If pulled apart at 300 pN , the H1–H12 bundle breaks into bundles H1–H8 and H9–H12 after 7 ns ( Figure 3 ) . This indicates that different structural changes are happening in parallel at different positions . Furthermore , when simulating the mechanical stabilities of alternate talin rod fragments that were not truncated along a ‘natural' bundle-bundle interface , for example simulating H1–H9 instead of H2–H8 , it is interesting to note that H9 is the least stable helix in those cases ( Figure 5 ) since it belongs structurally to the H9–H12 bundle . Unraveling of H9 is the first major unfolding event of H1–H9 even if the force is distributed over terminal helices , since H9 is easily detached from the H1–H8 bundle . Activation of the VBS requires that the VB helices H4 , H6 , H9 , H11 , and H12 , are at least partially exposed to water . As seen in Figure 3D , these five VB helices are buried in the H1–H12 bundle under equilibrium , and become sequentially exposed only after the fragmentation of the talin rod into smaller helix bundles has occurred . The asterix defines the point where the strain-exposed solvent-buried surface area of a VB helix is equal to the solvent-buried area of the helix when complexed to the vinculin head . The asterix thus marks the unique points in the unfolding trajectory in which each of the VB helices gets activated ( Figure 3D ) . Breaking the talin rod apart thus defines the highest energy barrier that has to be overcome to initiate the exposure and activation of the VBSs , and the implications thereof will be discussed below . To better understand the molecular mechanism behind the interactions that regulate helix bundle separation , we analyzed the hydrogen bonding pattern between the defined helix bundles as shown in Figure 4 . The blue fluctuatuions represent bonds formed between H1–H5 and H6–H8 , while the green represents bonds formed between H6–H8 and H9–H12 . While most hydrogen bonds fluctuate between formation and breakage even during the equilibration , only a few of the side-chain hydrogen bonds , like the bond 2 , are longer-lived . This , together with the fact that there appears to be considerable statistical variability in bond breaking events between different simulations suggests that at least most of these side-chain hydrogen bonds are not force-bearing . While it cannot be excluded that one or the other of these bonds slightly contributes to the mechanical stabilization , hydrophobic contacts between the helix bundles seem to play the dominant role in upregulating the mechanical stability of the N-terminal part of the talin rod . A separate SMD analysis of the talin rod fragments H1–H5 was done recently by applying force to the polar side chains T498 , S501 and S502 close to the N-terminus of H1 and Q635 , Q646 , E650 , and Q653 close to the C-terminus of H5 , assuming that the force-bearing interactions were mediated by side chain hydrogen bond formation across the interfaces of adjacent α-helix bundles [41] . When using an implicit water model in which the protein structure was solvated in a dielectric medium , they observed a rotation of VBS1 ( H4 ) under applied force and suggested this to be a potential activating mechanism . They also observed that the H4 rotation was strongly reduced when repeating the simulation in the presence of explicit water molecules ( these computationally more elaborate conditions were used in our simulations as well ) . We thus analyzed for how long the polar side chains of H5 are hydrogen-bonded across the interface formed between H1–H5 and H6–H8 . Our simulations of H1–H12 reveal that water penetrates into the interface H1–H5 and H6–H8 , thus breaking these side-chain hydrogen bonds , even before we can see a major force-induced structural change within the H1–H5 bundle ( Figure S3 ) . Among the residues that were previously used to model the contact interface between bundles H1–H5 and H6–H8 [41] , only residues Gln635 ( bonded to Ser714 ) and Gln646 ( bonded to Lys721 ) are hydrogen bonded across the interface , and those side-chain bonds show quite low stability during our simulations ( Figures 4 , bonds 6 and 14 , and S3 ) , suggesting that those polar residues are also not force-bearing during the activation process of the H4 helix . To compare the mechanical properties of the talin1 and talin2 rods , H1–H12 , we generated a homology model of talin2 based on the talin1 structure ( Figure S4 ) . Homology modeling is a reasonable approach since their sequences are highly similar ( 74% identical ) in this region . Furthermore , a sequence analysis of the hydrogen bonding partners when comparing talin1 and talin2 from human , mouse , and chicken ( Figure 4 ) revealed that 29 of the 43 residues participating in hydrogen bonds between helix bundles are fully conserved , eleven of the residues ( Ser658 , Ser688 , Thr693 , Thr720 , Ser754 , Gln755 , Gln762 , Arg765 , Glu780 , His788 , Lys869 ) are similar and only three residues are non-conserved ( Arg692 , Ser752 , Gly766 ) ( Figure S1 ) . The RMSD for backbone atoms after energy minimization and thermalization was 1 . 3 Å . Also for talin 2 , the fragmentation of the rod H1–H12 constituted the major energy barrier as seen in constant force and constant velocity SMD simulations , in which the force was distributed over force-bearing interfaces . The split of the rod occurred in identical positions as described above for talin1 , namely between the bundles H1–H5 , H6–H8 and H9–H12 . Also the sequence of early events was similar: First , the H9–H12 bundles separated , followed by the rupture of the H1–H8 fragment . Our simulations of early events further indicate that the major energy barrier of fragmenting the rod are not greatly different between the talin1 and talin2 rods , H1–H12 , at least within the stochastic variability that is intrinsic to single molecule studies ( Figure S5 ) . Finally , it is important to note that the SMD simulations are carried out on time scales that differ significantly from those at which biological molecules are stressed . Since the force needed to unfold a protein is logarithmically dependent on the pulling velocity , significantly smaller forces may be able to cause the here described structural rearrangements at physiological timescales . Unfolding forces measured by SMD in nanosecond timescale are thus significantly higher compared to those measured using AFM on millisecond timescales , yet , SMD has correctly predicted in the past the relative mechanical stabilities of some protein domains and the position of key energy barriers [32–34] .
Five VB helices have been identified in H1–H12 ( Figures 1 and S1; Table 1 ) [10] , namely H4 , H6 , H9 , H11 , and H12 . Breaking the H1–H12 talin rod into smaller bundles is the prerequisite for the strain-dependent decrease in the buried surface area of these amphiphatic VB helices ( Figure 3 ) . Starting with the first split where H1–H8 separates from the H9–H12 bundle ( Figure 3 ) , we observe a significant decrease in the buried surface areas of the three VB helices ( H6 , H9 , and H12 ) , while the buried surface area of the VB helix H4 is decreasing only at later time points when helix bundle H1–H5 is fragmented into smaller pieces . Note that H6 , H9 and H12 are located at the interfaces of the H1–H5 , H6–H8 and H9–H12 bundles . Since H11 is buried in the interior of the subbundle , there is only a slight drop in the buried area observed during the simulated time frame . Is the increased solvent exposure sufficient to activate the VBSs ? Considerable experimental evidence has demonstrated that bundle fragmentation by clipping off helices can indeed expose the VB sites that are otherwise buried in the full length talin rod under equilibrium conditions . The isolated H1–H12 bundle shows a higher affinity to vinculin than does the full-length talin or talin rod domain [30] , and mutations stabilizing the H9–H12 bundle result in decreased vinculin binding [30] . NMR studies have revealed that the helix H10 of the isolated H9–H12 bundle unfolds into a flexible random coil when the fragment complexes with VH [28] . Taken together , these results support the model where the H9–H12 bundle is stabilized by the neighboring helices which bury the VB helices , H9 , H11 and H12 . Similar trends are seen for the VB helix H4 . Experiments show that the H1–H5 bundle does not bind to vinculin but removal of H5 from the bundle activates the otherwise buried VB helix H4 [30] . Taking together all of these experimental and computational findings , we propose the following model of how tensile force acting on talin can activate its binding to vinculin: What drives the force-induced association of talin helices with vinculin once they have been activated by water exposure ? Key to the VB helices are hydrophobic residues that are located on one face of these amphiphatic α-helices which share the consensus ( LxxAAxxVAxxVxxLIxxA ) , where x is a variable amino acid residue [10] . As shown for talin's isolated VB helices H4 , H11 and H12 ( the only ones for which sufficient data are available ) , association with the vinculin head is energetically favored since the vinculin head is thermodynamically stabilized by complexation with VB helices [23 , 26] , and has considerable structural homology to the respective equilibrium talin bundle structures [10 , 27 , 28] . At equilibrium , the buried surface areas of talin's VB helices complexed with talin versus the vinculin head , respectively , are 1537 Å2 vs . 1310 Å2 for H4 ( residues 607–631 ) , 1190 Å2 versus 1144 Å2 for H11 ( residues 821–842 ) , and 1706 Å2 versus 1291 Å2 ( residues 853–876 ) ( Table 1 ) . Accordingly , the affinities of isolated VB helices with vinculin ( Kd = 3–33 nM ) [47] differ considerably from that of full length talin ( Kd = 8 . 9 μM ) [30] . It has thus been proposed earlier that once released from the hydrophibic core of the talin rod , the H4 ( VBS1 ) and potentially the other VB helices are available to induce ‘bundle conversion' thereby displacing the intramolecular interaction of the vinculin head to its tail [24 , 27 , 28 , 48] . This suggestion was further supported by findings that vinculin can be activated by helical peptides that contain the VBSs , subsequently leading to an increased affinity of the vinculin tail to actin [22 , 23] . Following these earlier suggestions derived for unstrained systems , we now propose that the complementary design between the talin and vinculin structures facilitates force-induced α-helix swapping: the VB helices find a thermodynamically more stable environment in talin under equilibrium conditions ( Table 1 ) , the biologically inactive state , but as soon as talin is sufficiently strained , the association with the unstrained vinculin head is energetically preferred once a collision has occurred and occurs upon a collision with the vinculin head . Vice versa , the vinculin head is thermodynamically less stable in the absence of a swapped talin helix [30] and thus forms the auto-inhibitory complex with its tail [23–26] . We further suggest a biphasic vinculin binding behavior: the maximum probability of the α-helix swapping is reached once the VB helix has broken away from the other talin helices and exposes the hydrophobic residues of the otherwise structurally intact VB helices to water ( Figure 6 ) . Once a VB helix is sufficiently stretched and starts to lose its secondary structure , helix swapping might again be inhibited . Of major physiological importance is furthermore the insight that the talin rod is engineered such that the activation of vinculin binding is force and time dependent , thus acting as a hierarchical force-time integrator . Not all VB helices are activated at the same time since the VB helices are located in sub-bundles that break up sequentially due to their differential mechanical stabilities ( Figures 3 and 5 ) . A hierarchy thus exists in which the VBSs are force-activated ( as indicated by asterixes in Figure 3D ) : the earliest to be exposed are the VB helices H6 and H12 , which are located in the weakest helix bundle interface . The next to be exposed are the VB helices H9 and H11 located in the C-terminal H9–H12 bundle , and the last one is VB helix H4 , which is located in the sub-bundle H1–H5 . If the talin rod is kept under constant force , the number of VB helices that have been activated will initially increase with time and the amount of force applied . Once a vinculin complex has been formed during the lifetime of the activated helix , the complexed helix might be protected against complete unraveling due to the additional interactions formed between a VBH and the unstrained vinculin head . In contrast , when applying force to the terminal ends of H1–H12 , the VB helices H11 and H12 lose already their secondary structure before the bundle is completely fractured , potentially leading to their early deactivation . Vinculin recruitment to talin thus initially increases if talin is incorporated into a force-bearing network formed when a cell adheres to a surface or matrix fibrils [15–18 , 20 , 29 , 30] . However , since each VB helix has its unique biphasic time response , increased force thus accelerates the sequence of these events , but not necessarily the total number of activated VBSs at later time points , particularly at low vinculin concentrations . The comparative analysis carried out here for the rod fragment H1–H12 of talin1 and talin2 shows that they share a similar sequence of early unfolding events that lead to the fragmentation of the N-terminal part of the talin rod . Talin1 and talin2 are not only differentially expressed in various tissues but also localize in different parts of a cell [49] . While we do not see major changes between the mechanical stabilities of the H1–H12 fragments of two talins , i . e . , in their likelihood of being fragmented by tensile force , it is important to note that the talin rod has six additional VBSs in the structurally unresolved C-terminus and that we do not have any information so far regarding the differential mechanical stresses to activate them ( see Figures 1B and S1 ) . Unclear is also to what extend functional differences of talin1 and talin2 might result from mutations in their binding sites to other proteins . The mechanism described here might not be unique to the talin-vinculin bond but might be more widespread among proteins that are composed of α-helical bundles . First of all , once an amphiphatic helix is broken off from a helix bundle by the application of tensile force , it might be stabilized by insertion into either hydrophobic pockets of other proteins or even into the lipid bilayer [27] . Alternatively , other proteins that form helix bundles might also bind vinculin in a force-regulated manner . α-actinin , for example , has also a VB helix that can form a similar structural complex with vinculin [23 , 26] . Similarly to talin , the VBS in α-actinin is buried in the native structure [50] . Identifying the repertoire of mechanisms by which forces can upregulate adhesive interactions has led to the recent discovery of catch bonds where a receptor-ligand interaction is enhanced when tensile mechanical force is applied between a receptor and its ligand ( for review see [51 , 52] ) . In contrast , the force-activated helix swapping mechanism proposed here requires that the force is applied to just one of the binding partners , thereby activating bond formation with a free ligand . Also in contrast to catch bonds , the ligand does not necessarily have to be part of the force-bearing protein network at the time the swap is initiated . While force-induced helix swapping thus primarily upregulates the bond formation rate , the catch bond mechanism primarily extends the lifetime of an already existing complex .
The talin rod domain H1–H12 ( residues 486–889 ) of mouse talin was obtained from the Protein Data Bank ( http://www . pdb . org/ [PDB theoretical models section: 1XWX] ) [28] . This structure is an energy-minimized model derived from combining two talin fragments derived from NMR ( H9–H12 , residues 755–889 [PDB: 1U89] [28] ) and X-ray ( H1–H9 , residues 482–789 [PDB: 1SJ8] [27] ) . Another talin structure was also studied ( H1–H9 [PDB: 1SJ8] ) . Talin fragments H2–H8 ( residues 523–757 ) and H9–H12 ( residues 755–882 ) were obtained from the 1XWX talin rod structure by removing the excess atoms . Homology model for talin2 H1–H12 was generated based on talin1 [PDB: 1XWX] using SWISS-MODEL protein structure homology-modeling server [53] and subjected to WhatCheck analysis . The human talin2 sequence was used; human and mouse talin2 are virtually identical in the region of H1–H12 , and there are only 2 sequence differences: the residue corresponding to human talin2 Glu604 is Asp in mouse , and the residue corresponding human to talin2 Thr762 is serine in mouse talin2 . The sequence alignment was prepared using ClustalW ( http://www . ebi . ac . uk/clustalw/ ) . The H1–H12 sequences of talin1 and talin2 have 74% identity . All the simulations were carried out in explicit water . The periodical TIP3 water box was created using the program VMD plugin solvate [54] The system was neutralized by adding Na+- or Cl−-ions to the system using VMD autoionize-plugin . The systems used in the simulations are summarized in Table S1 . Since efficient arrangement of water is a pre-exquisite for validity of the results , we used Solvate 1 . 0 ( http://www . mpibpc . mpg . de/groups/grubmueller/start/software/solvate/docu . html ) to find possible hydration sites inside the protein , and energetic evaluation of those positions was performed using Dowser [55] . Positions energetically favoring hydration were analyzed in the H1–H12 which was subjected to a 1 ns equilibration after solvation with the VMD solvate plugin . A further analysis revealed that 1 ns equilibrations were sufficiently long to allow water to penetrate into all energetically favored locations . A 12 Å cutoff was used for the van der Waals interactions , switching function starting from 10 Å . The PME ( Particle Mesh Ewald ) method was used to calculate long-range electrostatics without a cutoff using grid spacing lower than 1/Å3 . Each system was minimized with a conjugate gradient method using NAMD [56] . Initially , only the solvent and ions were allowed to move for 4 , 000 steps . Next , the entire system was allowed to move for another 4 , 000 steps . After minimization , the system was heated from 0 K to 300 or 310 K in 30 ps under Berendsen pressure control ( 1 atm ) and subsequently equilibrated for 1 ns under constant pressure ( Berendsen pressure control at 1 atm ) and temperature ( tCouple method ) . Structures equilibrated for 1 ns were used in all SMD simulations . Constant force was applied either to both termini , or alternatively to Cα-atoms of the terminal helices ( residues 495–514 and 853–872 ) and the force vector was then calculated using residues in the middle of the helices ( residues 504 and 865 ) . The system coordinates were saved every picosecond , and the system energies were recorded every 0 . 1 ps . Constant velocity SMD simulations were performed by applying moving springs with a spring constant of 5 kcal/mol/Å2 [30] to the Cα-atoms of residues 853–872 in the C-terminal helix H12 . The springs were moved along the force vector calculated according to the vector connecting the Cα-atoms of residues 504 and 865 with a velocity of 1 , 10 , or 100 Å/ns , while the Cα-atoms of the residues 495–514 of the N-terminal helix H1 were constrained using a harmonic energy constraint function with a spring constant of 5 kcal/mol/Å2 . The force along the pulling vector was obtained from the absolute forces measured during the simulation by calculating the projection of the force vector to the pulling vector . Pressure was maintained at 1 atm by the Berendsen pressure control method implemented in NAMD with the following parameters: BerendsenPressureCompressibility 0 . 0000457 , BerendsenPressureRelaxationTime 1000 , BerendsenPressureFreq 4 . The temperature of the system was maintained at a defined temperature using the tCouple method implemented in NAMD with a tCouple coefficient of 1 . The resultant trajectories were analyzed using VMD version 1 . 8 . 5 [54] . The pictures were rendered using Tachyon rendering system implemented in VMD and further processed with programs GIMP and Adobe Photoshop CS2 . Surface-accessible surface area ( SASA ) was measured using VMD with a 1 . 4 Å scanning probe . The vinculin head complexed to talin VBSs was first hydrogenated using the program psfgen . The structures were then subjected to surface analysis without further processing . The analysis was done for both the whole protein and individual helices . Calculations over simulation trajectories were performed for 10 ps spacing between frames . The buried areas of VB helices were calculated according to Eq . 1 . To avoid the burying effect of the backbone of the neighboring talin polypeptide chain , one residue before and after each helix was excluded from the analysis . This makes direct comparisons to vinculin–VBS complexes more accurate , since VBS-helices are individual peptides in these complexes and therefore not connected to a protein scaffold as are helices in the talin rod . Sequence alignments for human , mouse and chicken talin1 and talin2 proteins were done using program ClustalW using default parameters . The RefSeq database ( http://www . ncbi . nlm . nih . gov/RefSeq/index . html ) accession numbers used are as follows: human talin1 ( NP_006280 ) ; human talin2 ( NP_055874 ) ; mouse talin1 ( NP_035732 ) ; mouse talin2 ( XP_486227 ) ; chicken talin1 ( NM_204523 ) . The sequence of chicken talin2 was combined from database sequences ( XP_413760 ) and ( XP_413761 ) . Computation was carried out on the Gonzales cluster at ETH Zürich ( 2 . 4 GHz AMD Opteron 250 processors ) and at CSCS Cray XT3 ( 2 . 6 GHz AMD Opteron processors ) . For 1 ns of simulation time ( ∼100 , 000 atoms ) , ∼1100 cpu hours were needed on both clusters . Overall , >200 ns of simulation data were analyzed in this study .
|
For cell survival , most eukaryotic cells need to be mechanically anchored to their environment . This is done by transmembrane proteins , including integrins , which externally bind to the extracellular matrix and on the cell interior to the contractile cytoskeleton via scaffolding proteins . One essential scaffolding protein is talin , which binds to integrins via its head and to the cytoskeletal filament f-actin via its rodlike tail . As cells apply tensile forces to newly formed adhesion sites , proteins that are part of such force-bearing networks get stretched and might change their structure and thus function . One of many proteins that are recruited to newly formed adhesions is vinculin , and vinculin recruitment is upregulated by tensile mechanical force—but how ? Since talin's vinculin binding sites are buried in its native structure , we used steered molecular dynamics here to derive a high resolution structural model of how tensile mechanical forces might activate talin's vinculin binding sites . Once tensile forces break up the talin rod into helix subbundles , an event that we find here to constitute the main energy barrier , we propose how the strain-induced gradual exposure of the vinculin-binding helices finally allows for their activation and enables helix swapping with the vinculin head .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biophysics",
"cell",
"biology",
"mammals",
"computational",
"biology"
] |
2008
|
How Force Might Activate Talin's Vinculin Binding Sites: SMD Reveals a Structural Mechanism
|
Most biological processes accelerate with temperature , for example cell division . In contrast , the circadian rhythm period is robust to temperature fluctuation , termed temperature compensation . Temperature compensation is peculiar because a system-level property ( i . e . , the circadian period ) is stable under varying temperature while individual components of the system ( i . e . , biochemical reactions ) are usually temperature-sensitive . To understand the mechanism for period stability , we measured the time series of circadian clock transcripts in cultured C6 glioma cells . The amplitudes of Cry1 and Dbp circadian expression increased significantly with temperature . In contrast , other clock transcripts demonstrated no significant change in amplitude . To understand these experimental results , we analyzed mathematical models with different network topologies . It was found that the geometric mean amplitude of gene expression must increase to maintain a stable period with increasing temperatures and reaction speeds for all models studied . To investigate the generality of this temperature–amplitude coupling mechanism for period stability , we revisited data on the yeast metabolic cycle ( YMC ) period , which is also stable under temperature variation . We confirmed that the YMC amplitude increased at higher temperatures , suggesting temperature-amplitude coupling as a common mechanism shared by circadian and 4 h-metabolic rhythms .
Many physiological processes are sensitive to temperature . At the biochemical level , the speed of the reactions tends to increase two- to three-fold with a 10°C temperature rise [1] . At the system level , cell growth accelerates with temperature , and the cell cycle period of NIH3T3 cells decreases to one-third as the temperature rises by 10°C [2] . In contrast , the circadian rhythm period is robust to temperature [1 , 3] . Notably , this property , so-called temperature compensation , is observed in both species without and with strong thermal homeostasis ( poikilotherms and homeotherms ) . The definition of temperature compensation is that the oscillator period is constant at different but constant temperatures ( 0 . 85<Q10<1 . 15 ) [4] . For example , the circadian period of mammalian NIH3T3 cells remains roughly unchanged over a wide range of temperatures [2] . Similarly , the period of the yeast metabolic cycle ( YMC ) , approximately 4 h under certain experimental conditions , is also robust to temperature [5] . How can the period of biological rhythms be stable under varying temperature [6–21] ? Proposed hypothesis for this temperature compensation include the ( i ) "critical reaction hypothesis , " ( ii ) "balance hypothesis , " and ( iii ) "temperature–amplitude coupling hypothesis . " The critical reaction hypothesis ( i ) posits that there are critical reactions governing the circadian period , and that the period can be stable if these critical reactions are insensitive to temperature [11] . Alternatively , the balance between period-lengthening and -shortening reactions ( ii ) could lead to a stable period [8 , 9] . Finally , temperature-sensitivity of circadian oscillation amplitude ( iii ) may stabilize the period [7] . The critical reaction hypothesis ( i ) is supported by experimental evidence from cyanobacteria [11] and mammals [14] . However , it still remains unclear how the circadian period can be maintained against temperature in which not only the critical reactions but also the other reactions in the network should affect the period . As for the temperature–amplitude coupling hypothesis ( iii ) , there is yet no strong experimental evidence . For example , the amplitude of bioluminescence in the Gonyaulax circadian clock does decrease with temperature [6] . While the abovementioned hypothetical mechanisms for temperature compensation are not mutually exclusive , they may raise the question of whether any of abovementioned hypotheses is actually the mechanism of temperature compensation for circadian rhythms . In Drosophila , Neurospora , mammals , and plants , diurnal activities are driven by transcriptional regulatory networks for genes and proteins [1] . Transcriptional regulatory network structures for circadian rhythms are similar between species in that there are both positive and also negative elements controlling gene-expression [22–24] . However , the topology of the regulatory network structures is not precisely the same across species , so the mechanisms of temperature compensation may also vary among species . Moreover , the mechanisms of temperature compensation for circadian rhythms are not necessarily the same as those for the yeast 4 h-metabolic cycle . In this study , we examined the mechanisms for temperature compensation . First , we demonstrate temperature-sensitivity ( or insensitivity ) of the circadian clock system in C6 rat glioma cells . We then applied theoretical models to explain these results . Specifically , we addressed the question of whether period stability to temperature depends on the structures of regulatory networks by comparing experimental results to models with different structures . Finally , we revisited the time-series of the YMC [5] and discuss the possibility that the mechanisms for temperature compensation are shared between circadian and yeast metabolic rhythms .
It has been reported that the expression time series of some genes in circadian clock system are sensitive to temperature whereas those of other genes are not [6 , 18] . To investigate the mechanisms for temperature compensation , we systematically examined the temperature sensitivities of clock gene expression levels in mammalian cultured cells . We compared the expression time series of clock genes for C6 glioma circadian rhythms at 35°C and 38°C . In preliminary studies , the amplitude of the first cycle after the stimulation was markedly higher than those of following cycles . This finding suggests that , immediately after the stimulation , the state variable jumped out of the attractor of the circadian limit cycle and that the first cycle exhibited a relaxation process that did not trace the trajectory of the circadian limit cycle . Therefore , we excluded the first cycle and evaluated the mRNA expression levels of clock genes rBmal1 , rCry1 , rDbp , rE4bp4 , rPer2 , rPer3 , and rRev-erbα mRNAs every 4 h from 32 to 96 h post-stimulation . During this period , the amplitude of the oscillation remained reasonably constant in the period after the temperature shift . ( Fig 1A ) . Analyses of period and amplitude using ARSER [25] revealed that all genes examined showed a circadian rhythm with periods within the circadian range , from 21 . 0 to 23 . 6 h . We plotted the state points in chronological order with rCry1 and rBmal1 mRNA expression levels on the x- and y-axis , respectively ( Fig 1B ) , and found that the oscillatory trajectory at 38°C was much larger than at 35°C . To examine whether temperature affects the period length and amplitude , we calculated temperature coefficient ( Q10 ) values , which expresses the reaction change over a 10°C increase in temperature of seven clock genes ( Fig 1C and 1D ) . Although the Q10 values of the period varied between 1 . 01 ( rDbp ) and 0 . 74 ( rBmal1 ) , this range is deemed as temperature-compensated [4] and suggests that temperature compensation is preserved in C6 glioma cells . Amplitudes of both rCry1 ( p = 0 . 0068 ) and rDbp ( p = 0 . 043 ) mRNA expression levels increased significantly ( Fig 1D ) with , Q10 values of 10 . 57 and 7 . 92 , respectively . All other genes also exhibited Q10 values of larger than unity . Because the Q10 values for biological systems are generally between 2 and 3 [4] , oscillation amplitudes of most if not all are affected by temperature change . Do the observed temperature-dependent circadian expression levels of some genes ( temperature–amplitude coupling ) account for temperature compensation ? To understand the experimental results , we analyzed a classical negative feedback model of circadian rhythms [26 , 27] , the so-called Goodwin model [9 , 28 , 29] , and investigated the conditions required for period stability . The regulatory networks of circadian rhythms are more complex than assumed by this model , so we compared the conditions for period stability between this simple model and a detailed model [30] . The dynamics of the negative feedback model are shown in Fig 2A . The first term on the right hand side of the equation ( Fig 2A ) denotes the expression of a clock gene that is downregulated by a nuclear clock protein ( P ) with cooperativity ( n ) . For certain parameter choices , limit cycle oscillations are produced by the model . In this study , we assume that all biochemical reaction speeds increase with temperature . In the present model , reaction speeds as represented by vS , vM , vD , kS , k1 , and k2 are assumed to increase with temperature while parameters such as Michaelis constants ( KI , KM , KD ) and the Hill constant ( n ) are fixed for simplicity . Using this negative feedback model , we numerically analyzed the conditions for period stability with changing temperature . We analyzed the conditions with multiple parameter sets as the results can be parameter dependent . The analysis conditions are as follows: ( 1 ) we randomly prepared “basal” parameter sets ( 500 sets ) for oscillations and then ( 2 ) we randomly increased all reaction speeds from the basal parameter sets ( 100 sets for each basal parameter set ) . The ratios of increased speeds to basal speeds were set to range uniformly from 1 . 5 to 2 . 5 . Thus , the ratios of increased speed to basal speed for the degradation of mRNA ( vM ) and that of proteins ( vD ) can differ ( for example , 2 . 1 and 1 . 9 , respectively ) . With the increased reaction speeds , the periods tend to be shorter than the original periods with basal parameter sets ( Fig 2B and 2C ) . Calculated oscillations are often out of the range of temperature compensation ( 0 . 85<Q10<1 . 15 ) . When oscillations are nearly temperature compensated and the periods are relatively unchanged from the original periods , the amplitudes of mRNA ( M ) oscillations always exceed the original amplitudes yielded by the basal sets ( Fig 2C and 2D ) . Moreover , when the periods are relatively unchanged , not only the mRNA amplitudes ( M ) but also the amplitudes of other variables ( R , P ) tend to increase simultaneously . To measure amplitude changes of all variables , we used the geometrical mean amplitude ratio while it is possible that other metrics could be better . The numerical results of the negative feedback model suggest that temperature–amplitude coupling of some clock genes ( Fig 1 ) can cause period stability with temperature . In contrast to the negative feedback model described above , circadian rhythms in mammals and Drosophila involve both positive and negative feedback loops [24 , 31] . To understand the mechanisms for period stability in networks with under both types of feedback , we developed a full model with both positive and negative feedback loops and then simplify it to a two-variable model ( S1 Text ) . The simplified dynamics are: εdX/dt= ( k+vXγ/ ( KVγ+Xγ ) ) / ( h+Y ) −aX ( 1 ) dY/dt=sXα/ ( KSα+Xα ) −dY ( 2 ) where X is mRNA and Y is inhibitor protein . The first term on the right-hand side of Eq ( 1 ) indicates negative feedback regulation by an inhibitor protein ( Y ) and indirect positive feedback regulation . We assume cooperativity ( γ ) for positive feedback regulation . Production of inhibitor is also a nonlinear function of mRNA concentration ( X ) and accumulation of inhibitor ( Y ) is assumed not to occur immediately after the transcription . Degradation rates of mRNA and protein are expressed by linear terms for simplicity . Again , we assume that all the reactions represented by rates in linear kinetics ( a , d ) and maximum rates in Michaelis–Menten kinetics ( v , k , s ) become faster with increasing temperature , while all other parameters are fixed . We then examined the role of oscillation amplitude in period stability under these conditions . Given that our model is simplified to a two variables , we can derive an approximate formula for the period by assuming that the dynamics of the activator ( X , mRNA ) is sufficiently faster than that of the inhibitor ( Y , protein ) . The period formula and the technical details of the mathematical derivations are presented in S1 Text [32] . Using the equation for period , period can be coupled with amplitude . If all the rate parameters are multiplied by a common constant , the period shortens by a factor equal to the inverse of that constant [9 , 33] , meaning that in general the period tends to shorten along with increasing temperature and reaction speed . In contrast , the positive feedback strength ( v ) always lengthens the period in this model . As positive feedback strength ( v ) increases , the peaks of X and Y rise and the nadir of X decreases: thus , the amplitudes of X and Y always become larger ( Fig 3A–3C , S1 Text ) . This increase in amplitude results in period-elongation . These results suggest that a period-increasing reaction can also be an amplitude-increasing reaction . If the temperature sensitivity of an amplitude-increasing reaction is stronger , and amplitude is larger at higher temperature , the period can be stable with temperature due to cancelation of the period-shortening effect ( Fig 3D ) . Further , we can consider the necessity of temperature-amplitude coupling for this simple model ( S1 Text ) . Using the period formula for a certain parameter condition ( i . e . ds/dT>dd/dT in which T is temperature ) , we can mathematically prove that it is impossible to maintain the period along with increasing temperature and reaction speeds if maximums of variables at higher temperature are smaller than those at lower temperature , and minimums of variables at higher temperature are larger than those at lower temperature . Therefore , larger maximum or smaller minimum at higher temperature is necessary for temperature compensation , qualitatively consistent with the numerical results of Goodwin model and experimental results . The regulatory network for the mammalian circadian clock is more complex than the models described thus far , so we examined if the amplitude of oscillation is also important for period maintenance in the more detailed mammalian circadian clock model proposed by Kim and Forger ( 2012 ) . Using the original parameter set [30] , we assume that all the reactions in the system ( with 70 parameters ) become faster as the temperature increases except for the ratio of nuclear to cytosolic compartment volume ( Fig 4A–4F ) . In the calculations , the rates were randomly increased 1 . 1–1 . 9 fold , generating 10 , 000 parameter sets , and the sensitivities of period and amplitude to temperature were assessed ( Fig 4E and 4F ) . Of the 10 , 000 parameter sets , oscillations are maintained for 2 , 495 sets . For other sets , the time series of the system converge to equilibria or do not converge to any periodic solutions after a certain calculation time . When oscillations are maintained with faster reactions , the period tends to shorten ( Fig 4E ) . When periods are relatively maintained , the geometric means of the amplitudes of 180 variables tend to increase simultaneously , in agreement with the theoretical results from the simpler models described in this article . It appears that relative periods ( new periods with faster reactions , divided by original periods ) are always smaller than unity . This result does not quantitatively correspond to our experimental results showing that Q10 values of circadian mRNA oscillation periods are close to or larger than unity ( Fig 1 ) . When the rates were randomly increased within a wider range ( i . e . 1 . 1- to 5-fold ) , the temperature-compensated oscillations could be reproduced in which amplitude of Cry1 and Per2 are larger at higher temperature . Example of time series is depicted in Fig 4A–4D . The calculated period at high and low temperatures are 24 . 4 and 23 . 9h , respectively . This result suggests the possibility that some reactions of mammalian circadian rhythms are highly temperature-sensitive . Among the parameter sets that can achieve relatively stable periods , some successfully yield time series of Per2 , Bmal1 , Rev-erbα , and Cry1 mRNA expression levels in good qualitative agreement with actual profiles from mammalian cells in which amplitude of Cry1 and Per2 are larger at higher temperature ( Figs 1 and 4A–4D ) . Thus , temperature–amplitude coupling of Cry1 oscillations as observed in our experiments ( Fig 1 ) is indeed one possible solution for period stability in the detailed model ( yielding a temperature-compensated period ) . From this calculation , we can also identify the essential reactions for temperature compensation in the mammalian circadian clock system . Among the 2 , 495 parameter sets that yielded sustained oscillations , the period is relatively stable , with relative period > 0 . 8 , for 176 parameter sets ( Fig 4F ) . Although many different parameter sets yield a relatively stable period , some general necessary conditions can be identified . In Fig 4F , we plot the parameter sets that yield relatively stable periods ( 176 sets ) . Notably , a stable period does not occur if degradation of Per2 mRNA or unbinding rate of BMAL–CLK complexes or NPAS2 proteins to Per1/2/Cry1 Ebox is greatly increased ( i . e . , due to high temperature sensitivity ) . If degradation of Per2 mRNA is weakly temperature sensitive and its synthesis is strongly temperature sensitive , the amplitude of Per2 mRNA oscillation should increase with temperature . Similarly , if the unbinding rate of BMAL–CLK complexes or NPAS2 proteins from Per1/2/Cry1 Ebox is weakly temperature sensitive but binding rate is strongly temperature sensitive , Per1 , Per2 , and Cry1 mRNA expression levels should increase with temperature , in good agreement with the data ( Fig 1 ) . From this analysis of the detailed model , we conclude that temperature–amplitude coupling is a plausible mechanism for period stability with temperature that holds not only for simple models but also for a more realistic detailed model . In yeast , glycolytic cycles with periods on the order of minutes are sustained under a constant environment , but both period and amplitude decrease with temperature [35] . Tu et al . ( 2005 ) reported that an autonomous YMC with period of approximately 4 h occurs after deprivation of nutrients , and this period is temperature compensated [5 , 36] . To validate the contribution of temperature–amplitude coupling in period stability with temperature , we revisited the time series of YMC , which is Fig 1D ( http://www . tandfonline . com/doi/abs/10 . 4161/cc . 6 . 23 . 5041 ) of Chen and McKnight ( 2007 ) [5] . It was shown that the oscillator period measured from a time series of dO2 is relatively unchanged over a temperature range from 25°C to 35°C . Notably , as temperature increases from 25°C to 35°C , the nadir of the oscillations decreases and the amplitude increases . Conversely , the amplitude decreases with temperature from 35°C to 25°C , suggesting that the high amplitude sensitivity to temperature is reproducible . Although the molecular mechanisms underlying YMC are unknown , the present experimental and theoretical results suggest that temperature sensitivity of gene expression amplitude accounts for the maintenance of period under temperature changes .
How can the period of circadian oscillations be maintained under temperature variation when the underlying reactions accelerate with temperature [6–21] ? By analyzing the sensitivities of circadian clock period and amplitude to temperature change in mammalian cultured cells , we found temperature–amplitude coupling of Cry1 and Dbp mRNA oscillations . We then addressed the question of whether this observed temperature–amplitude coupling can cause temperature compensation by testing various plausible circadian models with different network topologies . For all the models we studied , when the periods with faster reaction speeds are relatively unchanged from the original periods , the geometrical mean amplitudes always exceed the original amplitudes yielded by the basal sets . These numerical results indicate that larger amplitudes with faster reaction speeds are necessary for stable period for the parameters we tested . Moreover , using our two-variable model , we mathematically showed that larger maximum or smaller minimum at higher temperature is needed for stable period for a certain parameter condition . These theoretical results suggest that temperature–amplitude coupling , observed experimentally ( Fig 1 ) should be necessary for temperature compensation . Our results are consistent with the result of Lakin-Thomas et al . ( 1991 ) , who showed that temperature–amplitude coupling can lead to a stable period in one-variable model with delay [7] . Previously , temperature compensation of circadian rhythms has been studied by quantifying emission generated by a bioluminescence reporter [37] , melatonin secreted in a culture medium [38] , and neural firing [39] as well as gene expression [2 , 37 , 40 , 41] . While the range of temperature compensation has been defined as 0 . 85<Q10<1 . 15 , longer period at higher temperature was reported by many literatures , so called "temperature overcompensation" [2 , 37 , 38 , 40 , 41] . Temperature overcompensation is within the range of temperature compensation . In consistent with those previous reports , quantification of gene expression in the present study showed that the circadian period of C6 glioma cells is longer at higher temperature ( Fig 1 ) . We also confirmed this temperature overcompensation by using a cell line expressing Bmal1::luc ( Q10 = 0 . 86 . S1 Fig ) . In contrast , numerical simulations in the present study showed that it is difficult to achieve perfect temperature compensation ( Q10 = 1 ) or temperature overcompensation ( Q10<1 ) when reactions speeds are assumed to increase 2–3 fold or 1 . 1–1 . 9 fold with 10°C temperature rise ( Figs 2C and 4E ) . When reactions speeds are assumed to increase within a wider range ( i . e . 1 . 1–5 fold ) , temperature compensation is more likely to occur ( Fig 4A–4D ) . This numerical result suggests the presence of strongly temperature sensitive reactions in circadian rhythms for temperature compensation and overcompensation . In some previous experimental studies , amplitude of circadian oscillations was reported to decrease with temperature [6 , 40 , 42] , which is inconsistent with the conclusion of the present study . We think that this discrepancy might be due to temperature sensitivity of bioluminescence although we cannot exclude the possibility that temperature-amplitude coupling does not hold for those rhythms . The intensity of bioluminescence increases at lower temperatures , peaks around 25°C , and drops sharply at higher temperatures [43] . This indicates that the bioluminescence level does not proportionate to the amount of the gene expression when temperature was being changed . In our present experiments , the amplitude of the bioluminescence level of Bmal1::luc decreased while the level of the gene expression of Bmal1 mRNA increased along with the rise of the temperature ( Fig 1 , S1 Fig ) . Notably , the expression amplitudes of some genes ( e . g . , Rev-erbα ) in mammalian cultured cells did not change with temperature while those of others ( e . g . , Cry1 and Dbp ) increased markedly with temperature . From the theoretical results of models with different network topologies ( S1 Text , S2 Fig ) , we can interpret temperature-dependent amplitudes of some transcripts ( e . g . , Cry1 and Dbp ) and temperature-independent amplitudes of the other transcripts ( e . g . , Rev-erbα ) as a consequence of a network structure of circadian rhythm with positive feedback loop ( s ) . Ferrell and coworkers found that models with both negative and positive feedbacks can yield a various period with fixed time series amplitudes depending on the specific parameter values [44 , 45] . These studies suggest that period stability to temperature is independent of amplitude when the network includes positive feedback . Unexpectedly , we found that the period tends to be maintained when the geometric mean amplitude increases with reaction speed for both models with and without positive feedback ( S1 Text; S2 Fig ) . We think that the discrepancy between our results and a previous theoretical study [45] is due to differences in model structure or our parameter settings . In our models , the effects of temperature are incorporated , all the reactions are assumed to increase in speed with temperature , and the parameters are often far from bifurcation points . In this study , we analyzed transcriptional–translational feedback loop models . In contrast , the temperature compensation of cyanobacteria circadian rhythms was suggested to be generated by a post-translational network or by a single key protein ( KaiC ) [11 , 46 , 47] . It is possible that these mechanisms also hold for temperature compensation of mammalian circadian rhythms [13 , 14] . Indeed , circadian oscillations are maintained even when transcription or translation is greatly suppressed in mammalian peripheral cells [13 , 48] , and previous works have shown the role of phosphorylation for temperature compensation [14 , 19 , 49] . Although temperature compensation can still occur in our regulatory network model described by Eqs ( 1 ) and ( 2 ) with partial inhibition of transcription or translation ( S3 Fig ) , we cannot exclude the possibility that temperature compensation in mammalian peripheral cells is generated by post-translational network or by a single key reaction . Notably , in the report of Nakajima et al . ( 2005 ) , the rhythm amplitude of cyanobacterial KaiC protein phosphorylation increased with temperature . We suggest that it is important to study the theoretical possibility that temperature–amplitude coupling can lead to period stability under temperature variation even in post-translational network models [21] . Is temperature–amplitude coupling a general mechanism for temperature compensation ? Although we do not have an answer to the question , we found that temperature–amplitude coupling stabilizes the period for all models studied despite differences in network topology . Furthermore , we confirmed that respiration rhythm amplitudes also increase with temperature in the ( temperature-compensated ) YMC [5] . Previously , Ruoff demonstrated that balance between period-increasing and -decreasing reactions can generate temperature compensation in any biochemical oscillators [8 , 50] which was supported by recent experiments of circadian rhythms in Neurospora [9] and mammals [19] . This balance theory and temperature–amplitude coupling hypothesis are not mutually exclusive , and we think that these are connected . In fact , mathematical analysis of our simpler model showed period-increasing reaction can be simultaneously amplitude-increasing reaction . This result suggests that balance between period-increasing and -decreasing reactions works through modulating oscillator amplitude for achieving temperature compensation . Our prediction of temperature–amplitude coupling can also be tested in a different way using phase response curve experiments [7] . Ute et al . ( 2010 ) reported that the range of entrainment increases as the ratio between zeitgeber strength and oscillator amplitude increases , suggesting that the magnitude of the phase response curve ( PRC ) to certain stimuli increases as oscillator amplitude decreases [51] . If oscillator amplitude is larger at higher temperature , then the magnitude of PRC in response to stimuli at higher temperature should be smaller than that at lower temperature . Indeed , the magnitude of the phase shift by light was smaller at higher temperature in Neurospora [52] . Recently , the magnitude of the phase shift by light was shown to be smaller ( type 1 ) at higher temperature than at lower temperature ( type 0 ) in Drosophila [53] . Measurements of PRC at different temperatures can also be conducted in a circadian mammalian culture system or yeast respiration rhythms system modulated by chemical stimuli ( such as forskolin or H2O2 , respectively ) [54] . Such experiments could reveal similarities or differences in the mechanisms underlying period stability to temperature between circadian rhythms and yeast respiration rhythms .
C6 glioma cells and C6-Bmal1::dluc cells ( kindly gifted by Dr . Kazuhiro Yagita ) were plated on 35 mm dishes and cultured in DMEM containing 10% FBS for several days at 37°C . After the cells had reached confluence , they were incubated in serum-free DMEM for a further 24 h and then separately stimulated by supplementation with 100 nM dexamethasone ( Dex , ICN ) . After 2 h of treatment with Dex , the culture medium was replaced with serum-free DMEM and the dishes were moved to the incubator or of 38°C or 35°C in a 5% CO2 atmosphere . For bioluminescence measurements ( S1 Fig ) , the dishes were moved to photomultiplier tube ( PMT ) , detector assemblies ( Kronos Dio , ATTO , Tokyo , Japan ) with same temperature and CO2 condition described above . To analyze period of bioluminescence , detrended traces computed by a software supplied with PMT , were fitted to an exponentially damped sine curve y = y0 + Ae ( −x/t0 ) sin[π ( x − xc ) /w] using software ( Origin8 . 1J; OriginLab , MA , USA ) [55] . To determine the significance of the differences , Student’s t-test was used ( *** p < 0 . 001 ) . From 32 h to 96 h after the treatment with Dex , cultured cells ( n = 3 ) were harvesting 800 μl nucleic acid purification lysis solution ( ABI ) every 4h and then total RNA was extracted with an ABI Prism 6100 Nucleic Acid PreStation ( ABI ) . An 0 . 5 μg aliquot of each total RNA preparation was then reverse transcribed using ReverTra Ace ( Toyobo ) and 2 . 5 μM oligo ( dT ) . TaqMan real-time PCR ( qPCR ) was next performed with an ABI PRISM 7700 Sequence Detector , in a total volume of 15 μl using Premix Ex Taq ( Perfect Real Time ) ( Takara ) , according to the supplier’s instructions . mRNA quantification was performed with two primers and a fluorescent probe as follows: rBmal1: forward primer , CTGAG CTGCCTCGTTGCA; reverse primer , CCCGTATTTCCCCGTTCACT; probe , TCGGGCGACTGCACTCACACATG; rCry1: forward primer , TTCGTCAGGAGGGCTGGAT; reverse primer , GCCGCGGGTCAGGAA; probe , CACCATCTAGCCCGACATGCAGTTGC; rDbp: forward primer , TGCCCTGTCAAGCATTCCA; reverse primer , AGGCTTCAATTCCTCCTCTGAGA; probe , TCGACATAAAGTCCGAACGAGCCCG; rE4bp4: forward primer , CAGGTGACGAACATTCAAGATTG; reverse primer , TTGCCGCCCAGTTCTTTG; probe , TCCCTCAGATCGGAACACTGGCATC; rPer2: forward primer , GCTCTCAGAGTTTGTGCGATGA; reverse primer , AAAAGACACAAGCAGTCACACAAATA; probe , TTGTTCATGCG CAAACCAAACGTACC; rPer3: forward primer , CCGGAAGGTCTCCTTCATCAT; reverse primer , TGGTGGCAAAAACATCTTCATT; probe , TCGACATAAAGTCCGAACGAGCCCG; and rRev-erbα ( Nr1d1 ) : forward primer , TGAAAAACGAGAACTGCTCCATT; reverse primer , CCAACGGAGAGACACTTCTTGAA; probe , TATCAATCGCAACCGCTGCCAGC; and . For the normalization of template concentrations , primers and a probe for glycaldehyde-3-phosphate dehydrogenase ( GAPDH ( ABI ) ) were used . The resulting threshold cycle ( Ct ) values from the cDNA amplifications were thus normalized to the Ct values for GAPDH . With these results , the period and amplitude of each gene expression were analyzed with ARSER [25] , which supplied the period and amplitude of the circadian rhythm . The experiment was repeated three times independently and , by applying ARSER , amplitude and period for each experiment were obtained and statistically analyzed for the effect of temperature change . To determine the significance of the differences , Student’s t-test was used ( * p < 0 . 05 and ** p < 0 . 01 ) . Ordinary differential equations are solved numerically by Runge-Kutta method with Δt = 0 . 01 except for a detailed mammalian circadian model . For the detailed model , we used the Euler method with Δt = 0 . 001 . To express temperature sensitivity in rate constants for Fig 2D , Fig 3D and S3 Fig , we used Arrhenius relation for each reaction rate . We obtained period and amplitude sensitivity to temperature or reaction rates ( Fig 3 , S3 Fig ) by the Runge-Kutta method with Δt = 0 . 01 using XPPAUT [34] . Relative amplitude in Figs 2 and 4 , and S2 and S4 Figs was defined as the new amplitude with increased reaction speeds divided by the basic period with basic parameter sets .
|
Circadian rhythms govern the timing of many physiological events . Mysteriously , the period of the rhythm is robust to temperature although the underlying biochemical reactions usually accelerate with temperature , a paradox that has remained unsolved for more than 60 years . Experiments conducted over the last few decades in insects , mammals , and plants have demonstrated that biological rhythms are governed by cyclical changes in gene expression . However , the topologies of the regulatory network structures for these rhythms differ between species , suggesting that the mechanisms for period stability with temperature are not conserved . But is it true ? We examined the mechanisms for period stability with temperature by combining computational models with distinct network structures and experimental observations from mammalian cell cultures . Unexpectedly , we found that temperature-sensitive amplitude of gene expression , which we call "temperature–amplitude coupling , " can stabilize the period with temperature in all models studied , despite differences in regulatory network structures . Thus , the mechanisms for circadian period stability may be shared across species .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"ecology",
"and",
"environmental",
"sciences",
"messenger",
"rna",
"electromagnetic",
"radiation",
"circadian",
"oscillators",
"chronobiology",
"ecological",
"metrics",
"gene",
"expression",
"luminescence",
"physics",
"circadian",
"rhythms",
"biochemistry",
"genetic",
"oscillators",
"rna",
"bioluminescence",
"ecology",
"nucleic",
"acids",
"q10",
"temperature",
"coefficient",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences"
] |
2017
|
Temperature–amplitude coupling for stable biological rhythms at different temperatures
|
Cdc7p-Dbf4p is a conserved protein kinase required for the initiation of DNA replication . The Dbf4p regulatory subunit binds Cdc7p and is essential for Cdc7p kinase activation , however , the N-terminal third of Dbf4p is dispensable for its essential replication activities . Here , we define a short N-terminal Dbf4p region that targets Cdc7p-Dbf4p kinase to Cdc5p , the single Polo kinase in budding yeast that regulates mitotic progression and cytokinesis . Dbf4p mediates an interaction with the Polo substrate-binding domain to inhibit its essential role during mitosis . Although Dbf4p does not inhibit Polo kinase activity , it nonetheless inhibits Polo-mediated activation of the mitotic exit network ( MEN ) , presumably by altering Polo substrate targeting . In addition , although dbf4 mutants defective for interaction with Polo transit S-phase normally , they aberrantly segregate chromosomes following nuclear misorientation . Therefore , Cdc7p-Dbf4p prevents inappropriate exit from mitosis by inhibiting Polo kinase and functions in the spindle position checkpoint .
Accurate ordering of cell cycle events is an important requirement for the viability of all eukaryotic organisms . Once cells commit to duplicate their genome they must restrain mitosis until replication is complete and then accurately coordinate mitosis with cytokinesis to ensure the faithful transmission of chromosomes to daughter cells [1] . Importantly , errors in cell cycle checkpoints that enforce this ordering can be deleterious for accurate chromosome transmission . For instance , DNA damage or replication fork arrest during S-phase elicits a reversible block to mitotic progression by the budding yeast Mec1p ( HsATR ) and Rad53p ( HsChk2 ) checkpoint kinases [2] , [3] . In the absence of Mec1p or Rad53p , replication fork arrest during S-phase is not sensed leading to premature mitotic events and cell death ( reviewed by [4] ) . Additionally , since daughter cell growth is highly polarized in the budding yeast , exit from mitosis is prevented until sister chromatids segregate through the bud neck and into the daughter cell [5]–[7] . This ensures that spindle disassembly and mitotic exit are not initiated until accurate chromosome partitioning between mother and daughter cells has occurred . Failure to block mitotic exit when nuclear division takes place within the mother cell results in polyploid and anucleate progeny [8] , [9] . It is not surprising therefore , that both entry into and exit from mitosis are delayed by cellular checkpoints that respond to replication stress , chromosome damage , or spindle disruption [1] . Errors in these mitotic checkpoints are catastrophic and result in ploidy defects and genetic alterations , which are frequently observed in human cancers ( reviewed by [10] ) . The Cdc7p-Dbf4p kinase is required to catalyze the initiation of DNA synthesis at the beginning of S-phase ( reviewed by [11] ) . Cdc7p kinase activity is tightly regulated during the cell cycle by binding the Dbf4p regulatory subunit , which is cyclically expressed . Dbf4p accumulates in late G1 , is present throughout S-phase and then is destroyed during mitosis and early G1 by anaphase promoting complex ( APC ) -dependent degradation [12]–[17] . Therefore , Cdc7p-Dbf4p kinase activity is low following exit from mitosis and entry into G1-phase until it is needed to initiate a new round of DNA synthesis in late G1-phase of the following cell cycle . Multiple lines of evidence suggest that Cdc7p-Dbf4p activates the MCM DNA helicase [18]–[20] that is assembled at origins of replication in early G1 in an inactive form ( reviewed in [21] , [22] ) . In addition to its essential role in replication initiation , several studies suggest that the Cdc7p-Dbf4p kinase responds to DNA damage or replication fork stalling but its precise role in these activities is unknown [17] , [23]–[25] . Dbf4p encodes a dispensable BRCT-like domain in the N-terminus that might target the kinase to stalled replication forks [26] , [27] . In fission yeast , the Cdc7p-Dbf4p ortholog Hsk1p-Dfp1p interacts with Swi1p ( budding yeast Tof1p ) , a component of replication forks required for fork stability and also promotes centromeric cohesion in early mitosis [28] , [29] . Rad53p also phosphorylates Dbf4p in response to replication stress and this regulation requires N-terminal Dbf4p sequences through which Rad53p physically interacts [17] , [25] , [30] . Interestingly , the absence of the BRCT-like domain results in a defect in late origin activation suggesting that this domain might alter Cdc7p-Dbf4p binding at early versus late replication origins [26] . Together , these data suggest that the Dbf4p N-terminus encodes non-essential regulatory functions that target the kinase to particular substrates . To identify proteins that interact with the Dbf4p N-terminus , we performed a yeast two-hybrid screen with an N-terminal region of Dbf4p and identified an interaction with the Cdc5p kinase , the only Polo ortholog in yeast . Budding yeast Polo , like Drosophila Polo and human Polo-like kinase 1 ( Plk1 ) , functions as a master regulator of mitotic progression and is also required for cytokinesis ( reviewed by [31] , [32] ) . Polo activity is regulated by several independent cellular mechanisms . Polo protein levels are controlled by APC-dependent degradation in mitosis/G1-phase and activation of Polo catalytic activity requires phosphorylation by Cdk1 kinase early in G2 [33]–[35] . In addition , Polo function is inhibited by cell cycle checkpoints that are induced following DNA or spindle damage [36]–[38] . A genetic and physical interaction between Dbf4 and Polo was described previously [39] , [40] , however the biological significance of this interaction was not known . Polo controls multiple mitotic events to ensure accurate chromosome segregation . After anaphase initiation , Polo is required to activate the FEAR ( Cdc14 early anaphase release ) and MEN ( mitotic exit network ) pathways that promote nucleolar release of Cdc14p phosphatase [37] , [41]–[43] . Limited Cdc14p release by the FEAR pathway promotes accurate rDNA and telomere segregation [44]–[46] . Subsequent full nucleolar release of Cdc14p by the MEN reverses Cdk substrate phosphorylation that leads to APC-Cdh1p activation , cyclin destruction and mitotic spindle disassembly ( reviewed by [47] ) . Activation of the MEN is promoted by Tem1p-GTP and antagonized by Bfa1p-Bub2p , a two-component GTPase activating protein ( GAP ) [48]–[50] . To promote mitotic exit , Polo phosphorylates Bfa1p-Bub2p to inhibit its GAP activity and is also required for activation of Dbf2p kinase activity , independently of Bfa1p-Bub2p [37] , [49] , [51] , [52] . The Polo requirement for Dbf2p kinase activation may reflect that Polo also promotes Cdc14p release in the FEAR pathway , which primes the MEN [43] . Therefore , Polo promotes accumulation of Tem1p-GTP and activation of the downstream MEN kinases Cdc15p and Dbf2p , which ultimately cause full release of Cdc14p from the nucleolus . In response to replication fork arrest , Rad53 inhibits MEN activation , which may or may not impact Polo activity since the molecular basis of this regulation is not understood [53] , [54] . Spindle position defects also counteract Polo activity by targeting Kin4p kinase to the spindle poles where it inhibits Polo-dependent Bfa1p phosphorylation [8] , [9] , [55] . Failure to execute the spindle position checkpoint ( SPOC ) results in premature exit from mitosis and nuclear partitioning defects . Here we define an N-terminal Dbf4p polo-box interaction region ( that we refer to as the “PIR” ) that binds directly to Polo and show that Dbf4p inhibits Polo and Dbf2p activity . Deletion of the PIR allows Cdc14p nucleolar release in a cdc5-1 mutant at the non-permissive temperature . In response to nuclear mispositioning , a dbf4 mutant lacking the PIR fails to arrest in mitosis and prematurely exits the cell cycle . Thus , Dbf4 protein is required for proper functioning of the spindle position checkpoint most likely by antagonizing the ability of Polo to promote Cdc14p release in either the FEAR or MEN pathways . Our work therefore reveals a previously unrecognized function for Dbf4p in the regulation of mitotic progression through a direct interaction with Polo .
We conducted a yeast two-hybrid screen to identify proteins that interact with the Dbf4p N-terminus ( residues 67–227 ) and recovered multiple clones encoding the polo-box domain ( PBD ) of Cdc5p . Polo kinase has two conserved domains; an N-terminal kinase domain and a C-terminal region called the polo-box domain ( PBD ) ( reviewed by [56] ) , which is a phospho-Ser/Thr binding module that targets the kinase to its mitotic substrates [57] , [58] . The crystallographic structure of the Plk1 PBD bound to a phospho-threonine peptide has been solved [59] . Since the Dbf4p BRCT-like region alone ( residues 110–227 ) failed to interact with the Polo PBD ( Figure 1A ) , this suggested that the PBD interaction was occurring through Dbf4p N-terminal sequences from 67–109 . Residues 67–109 were similarly required for the Polo interaction within the context of full length Dbf4p ( Figure 1B ) and were sufficient to interact with the Polo PBD ( Figure 1A ) . Dbf4p residues 67–109 with all serines/threonines changed to alanine still interacted with the PBD ( Figure 1A ) suggesting that the PBD can bind to this Dbf4p region independently of phosphorylation . Further deletion and point mutant analysis ( Y-C . C . and M . W . , unpublished data ) revealed that residues 82–88 are essential for the Dbf4p-Polo interaction ( Figure 1A , B ) . The PBD is composed of three conserved regions called the Polo-cap ( Pc ) , Polo-box 1 ( PB1 ) and Polo-box 2 ( PB2 ) that fold together to form a functional phosphopeptide-binding domain [59] . We deleted conserved residues within the PBD to test their requirement for interaction with Dbf4p ( Figure 1C ) . Deletion of residues preceding the PBD ( GAD-Polo454–705 ) had little effect on the Dbf4p-Polo interaction . However , elimination of the Pc ( GAD-Polo510–705 ) completely disrupted the interaction with Dbf4p . These data suggest that the structural integrity of the Polo PBD is required for Dbf4p binding . Dbf4p binds and activates the Cdc7p kinase subunit in yeast and has no known role apart from its interaction with Cdc7p [14] , [60] . To determine whether the interaction between Dbf4p and Polo occurred in the context of the full-length Cdc7p-Dbf4p kinase , Sf9 cells were co-infected with baculoviruses expressing Polo , wild type HA-Cdc7p-Dbf4p or wild type HA-Cdc7p with various Dbf4p deletion derivatives . HA-Cdc7p-Dbf4p kinase was immunoprecipitated using an antibody against the HA tag and examined for the presence of Polo . All the Dbf4p deletion derivatives we examined interact with Cdc7p and activate normal Cdc7p kinase activity ( [26] and data not shown ) . Whereas Polo interacted with full-length Cdc7p-Dbf4p and Cdc7p-Dbf4-NΔ65p , Polo did not interact with Cdc7p-Dbf4p complexes that lacked the Dbf4p N-terminal 109 residues required for the Polo two-hybrid interaction ( Figure 2A ) . These data indicate that full length Cdc7p-Dbf4p kinase interacts with Polo but that Dbf4p residues 65–109 are required for this interaction . Importantly , HA-Cdc7p-Dbf4p interacts with Cdc5p in yeast when the proteins are expressed at endogenous levels and this interaction also depends on the Dbf4p N-terminus ( Figure 2B ) . We next tested whether Polo bound directly to Dbf4p using purified proteins . GST-PBD and Sumo-Dbf4p67–109 fusion proteins purified from E . coli were mixed , pulled down using glutathione-Sepharose and analyzed by immunoblotting . Although Sumo alone did not interact with GST-PBD , Sumo-Dbf4p67–109 interacted with GST-PBD but not with GST alone ( Figure 2C ) . These data indicate that Dbf4p residues 67–109 ( that we refer to as the Dbf4p PIR ) are sufficient for a direct interaction with the Polo PBD . Cdc7p-Dbf4p is required to initiate DNA replication , but is present throughout S-phase and during the metaphase to anaphase transition . Dbf4p is subject to APC-Cdc20p dependent degradation [16] but some protein is still present in late mitotic mutants that have activated the Cdc20p but not the Cdh1p form of the APC [17] . We examined Dbf4 protein abundance relative to Pds1 protein in cells moving synchronously through the cell cycle , since Pds1p is degraded at the onset of anaphase by APC-Cdc20 [61] . We found that although the abundance of both proteins declines at the same time , Pds1p is absent during mitosis while some fraction of Dbf4p persists ( Figure S5 ) . In contrast , Dbf4p has very low abundance or is absent in cells arrested in G1-phase by mating-pheromone when the APC-Cdh1p is active [12] , [13] , [15] , [17] and Dbf4p is stabilized by inactivation of the APC in G1 or by removal of its N-terminal D-box [13] , [15]–[17] . Together these data suggest that Dbf4p degradation can occur via both APC-Cdc20p and APC-Cdh1p mediated ubiquitylation . Since budding yeast Polo is not required for DNA replication but promotes multiple mitotic activities , we reasoned that Cdc7p-Dbf4p might influence Polo activity during mitosis . The dbf4-NΔ109 mutant progresses normally through the cell cycle and does not exhibit any obvious growth defects or temperature sensitivity ( Figure 3A , B ) [26] . We therefore tested for genetic interactions between dbf4-NΔ109 and the cdc5-1 temperature sensitive ( ts ) mutant . At the restrictive temperature cdc5-1 cells arrest in late telophase with divided nuclei , elongated spindles and high Cdk1p-Clb2p levels indicating a failure to exit mitosis [34] , [35] , [62] . The cdc5-1 mutant is ts at 30°C on rich media , but we found that dbf4-NΔ109 suppressed the cdc5-1 temperature sensitivity up to 35°C indicating a strong suppression of its growth defect ( Figure 3A ) . The dbf4-Δ82–88 mutant defective for interaction with Polo also suppressed the cdc5-1 ts ( Figure 3A ) . Since the cdc5-1 ts suppression was reduced in heterozygous dbf4-NΔ109/DBF4 diploids compared to dbf4-NΔ109/dbf4-NΔ109 diploids , i . e . dbf4-NΔ109 was haploinsufficient ( Figure 3C ) , our data strongly suggests that Dbf4p is a Polo inhibitor and that loss of the Dbf4-Polo interaction leads to increased cdc5-1 activity . We confirmed that Dbf4p inhibited Polo activity using several independent genetic tests . An extra plasmid copy of wild type DBF4 but not dbf4-NΔ109 inhibited the growth of cdc5-1 cells ( Figure 3D ) . A dbf4-NΔ65 mutant that disrupts the D-box ( residues 62–70 ) resulting in elevated protein levels was synthetically sick or lethal in combination with cdc5-1 ( Figure 3E ) . Since the dbf4-NΔ65 mutant exhibits a wild type growth rate and normal S-phase entry and progression ( data not shown; [26] ) but binds to Polo , this suggests that elevated Dbf4p levels are deleterious to cdc5-1 activity . Finally , elevated expression of the Dbf4p N-terminus from the GAL1 promoter completely inhibited the growth of cdc5-1 cells but had no effect on the growth of wild type ( not shown ) or a mcm2-1 ts mutant . Mcm2 is a component of the MCM helicase , which is thought to be the physiological target of Cdc7p-Dbf4p during the initiation of DNA replication [11] . The inhibition of cdc5-1 growth depended on the Dbf4p-Polo interaction , since deletion of the PIR ( residues 66–109 ) or Dbf4p residues 82–88 abrogated the growth inhibition ( Figure 4A , B ) . Since the Dbf4p N-terminus interacts with the PBD , this data suggest that overexpression of Dbf4 N-terminal peptides interferes with essential Polo-substrate interactions by competitive inhibition . Together , these data indicate that the Dbf4p N-terminus inhibits Polo activity and that this inhibition requires residues 66–109 , which are also required for the Dbf4p-Polo physical interaction . We wanted to determine whether the Cdc7p kinase subunit is required to inhibit Polo in the FEAR or MEN pathways . This is not straightforward since Cdc7p is an essential protein kinase . Importantly , inhibiting Cdc7p activity would not only inhibit replication origin firing but would also likely induce the replication checkpoint that inhibits the metaphase to anaphase transition and MEN activation [4] . Inhibiting Cdc7p activity would thus interfere with the mitotic pathways we would like to measure . Therefore , we addressed this question indirectly by taking advantage of our observation that high copy dbf4-NΔ109 suppressed the cdc5-1 ts phenotype ( Figure 4C ) . Since Dbf4p residues required for interaction with Cdc7p map between residues 312–704 ( C . G . and M . W . unpublished data ) , Dbf4-NΔ109 protein ( expressed in high copy ) will compete with full length Dbf4p ( in single copy ) for Cdc7p binding . Therefore , our finding suggested that high copy expression of Dbf4-NΔ109p reduced the cellular concentration of wild type Cdc7p-Dbf4p , the likely Cdc5p inhibitor , which suppressed the cdc5-1 ts allele ( Fig 4C ) . High copy expression of Dbf4-Δ82–88p that does not interact with Polo also suppressed the cdc5-1 ts ( Fig 4C ) . Importantly , deleting Dbf4p C-terminal residues required for interaction with Cdc7p ( CΔ312–704 ) eliminated the ts suppression by high copy dbf4-NΔ109 and dbf4-Δ82–88 . These data are consistent with full-length Cdc7p-Dbf4p kinase acting as the physiological Polo inhibitor . Wild type Cdc7p-Dbf4p might inhibit Polo abundance or kinase activity during the cell cycle and thus explain our genetic data . However , we saw little difference in Polo protein levels , cell cycle expression or Polo kinase activity comparing wild type yeast with the dbf4-NΔ109 mutant ( Figure S2 ) . This suggests that Dbf4p inhibits Polo independently of altering its expression or kinase activity . This is consistent with our genetic data since loss the Dbf4p PIR suppresses the cdc5-1 allele yet the Cdc5-1 protein retains considerable protein abundance and kinase activity at the non-permissive temperature [63] . The mitotic exit defect associated with cdc5-1 is due to a single P511L amino acid substitution preceding polo-box 1 of the PBD [64] , strongly suggesting that the cdc5-1 growth defect is caused by a defect in substrate recognition . Since genetically DBF4 is a negative CDC5 regulator we hypothesized that Cdc7p-Dbf4p phosphorylates Polo to prevent its access to key substrates in the MEN . Consistent with this possibility , we found that purified Cdc7p-Dbf4p phosphorylated recombinant GST-PBD but not GST alone ( Figure 2D ) . The Cdc14p phosphatase is sequestered within the nucleolus during the cell cycle prior to FEAR and MEN pathway activation [50] , [65] . Activation of the FEAR pathway allows limited Cdc14p nucleolar release , which promotes rDNA and telomere segregation during early anaphase [42] , [44]–[46] . Cdc14p is then fully released by MEN activation and antagonizes Cdk activity to trigger exit from mitosis [66] . Since the cdc5-1 mutant fails to release Cdc14p at the restrictive temperature [66] , suppression of the cdc5-1 ts by deletion of the Dbf4p PIR ( Figure 3A ) suggested that Cdc14p release is likely restored in these cells at higher temperatures . We quantitated the nucleolar release of Cdc14-EGFP in wild type , dbf4-NΔ109 , cdc5-1 and cdc5-1 dbf4-NΔ109 cells at a restrictive temperature for cdc5-1 . Cells were arrested in G1-phase at the permissive temperature and then released into the cell cycle at 34°C . The dbf4-NΔ109 cells progressed through the cell cycle and released Cdc14p similarly to wild type cells ( Figure 5A ) . Consistent with previous reports , the cdc5-1 mutant failed to release Cdc14p from the nucleolus but a significant amount of Cdc14p was released from the nucleolus in the cdc5-1 dbf4-NΔ109 mutant ( Figure 5A ) . We noticed a delay in mitotic progression at 34°C in the cdc5-1 dbf4-NΔ109 cells evidenced by a somewhat longer duration of Cdc14p release compared to the wild type and delayed cytokinesis ( indicated by the delayed appearance of unbudded ( G1 ) cells ) . This data indicates that Dbf4p inhibits Polo activity to prevent Cdc14p release , which might be significant during a slowed S-phase or during periods of replication stress . Inactivation of Bfa1p-Bub2p is required to activate the MEN [37] . This signaling cascade is partially activated as a result of Bfa1p phosphorylation by Polo , which leads to activation of the Cdc15p and Dbf2p-Mob1p kinases ( reviewed by [47] ) . Dbf2p kinase activation requires a Bub2p-Bfa1p independent function of Polo as well [49] , indicating that Polo either directly promotes Dbf2p kinase activity or promotes a MEN-independent pathway that activates Dbf2p . We tested whether Dbf4p functions as a negative regulator of MEN activation by examining whether deletion of the Dbf4p PIR could suppress growth defects associated with additional ts mutants in the MEN ( Figure 5B ) . We examined the growth of double mutants of cdc5-1 , cdc15-2 , dbf2-1 , cdc14-1 , or cdc14-3 with dbf4-Δ109 . As with cdc5-1 , loss of the DBF4 PIR rescued the growth of dbf2-1 cells at the non-permissive temperature . In contrast , the dbf4-Δ109 mutant failed to suppress the ts phenotype of cdc15-2 , cdc15-4 ( not shown ) , cdc14-1 , or cdc14-3 mutants ( Figure 5B ) . This suggests that Dbf4p may specifically inhibit Polo activation of Dbf2p and not Polo inactivation of Bub2p-Bfa1p GAP activity . Taken together , our observations suggest that Dbf4p antagonizes Polo activation of some Bfa1p-Bub2p independent step in MEN activation . This interpretation is further supported by the following experiments . Deletion of BUB2 ( or BFA1 ) is sufficient to cause premature mitotic exit when cells are arrested in metaphase with the spindle poison nocodazole [67] , [68] . This causes large budded cells to exit mitosis and rebud in the absence of chromosome segregation or cytokinesis . Since Dbf4p is a negative regulator of Polo activity , we examined whether deletion of the Dbf4p PIR was sufficient to induce rebudding in the presence of spindle poisons . Cells were arrested in G1 , released into media containing nocodazole and quantitated for rebudding ( Figure 6A ) . In contrast to bub2Δ , dbf4-NΔ109 did not allow rebudding in a wild type background nor did this mutation advance rebudding in a bub2Δ background . This indicated that loss of the Dbf4p Polo interaction is not sufficient to cause mitotic exit during metaphase and suggested that Dbf4p Polo inhibition may act independently of Bub2p-Bfa1p . Null alleles of CDC5 fail to activate the MEN and arrest in telophase with unphosphorylated Bfa1 [37] . Although the cdc5-1 mutant is also defective in MEN activation it is proficient for Bfa1p phosphorylation and retains substantial Polo kinase activity at the non-permissive temperature ( NPT ) [37] , [63] . This suggests that the cdc5-1 mutant is defective in activating a Bfa1p-independent function of the MEN , perhaps in FEAR pathway activation or activating a downstream MEN target . In contrast to cdc5-1 , cdc5-2 cells neither phosphorylate Bfa1p nor activate the MEN at the NPT [37] . The cdc5-2 temperature sensitivity is partially rescued by deletion of either BFA1 or BUB2 [37] but not by dbf4-NΔ109 ( Figure 6B ) . However , in a bub2Δ the cdc5-2 temperature sensitivity was further suppressed by deletion of the Dbf4p PIR . So , eliminating the requirement for Bfa1p-Bub2p inactivation in cdc5-2 cells ( i . e . cdc5-2 bub2Δ ) , allowed dbf4-NΔ109 to further suppress the cdc5-2 ts and promote mitotic exit at 34°C ( Figure 6B ) . These data are consistent with the interpretation that Dbf4p primarily inhibits Polo activation of a Bub2p-Bfa1p independent step in MEN activation , e . g . Dbf2p activity , Cdc14p release in the FEAR pathway , or some unknown activity . In the budding yeast , KAR9 and DYN1 encode cytoplasmic microtubule-associated and motor proteins , respectively , that operate in two redundant pathways essential for the correct positioning of the nucleus during anaphase ( reviewed by [7] ) . Deletion of either gene results in a small percentage of cells with nuclear orientation defects but deletion of both genes is lethal . In response to nuclear misorientation , S . cerevisiae inhibits premature activation of the MEN via the spindle position checkpoint ( SPOC ) . When the SPOC is activated the Kin4p kinase localizes to spindle pole bodies ( SPB ) to counteract Polo Bfa1p phosphorylation [8] , [9] , [55] . Kin4p thus counteracts Polo inactivation of the Bfap1-Bub2p GAP and this inhibits MEN activation . Failure to adequately respond to nuclear mispositioning allows inappropriate nuclear division within the mother cell , leading to anueploidy and loss of viability . Given that Polo Bfa1p phosphorylation is prevented when the SPOC is activated , we hypothesized Dbf4p may also inhibit Polo to prevent mitotic exit in response to nuclear misorientation . To test whether DBF4 inhibited mitotic exit when nuclei were mispositioned , we examined wild type , dbf4-NΔ109 , kar9Δ and kar9Δ dbf4-NΔ109 strains for evidence of mitotic exit in the presence of mispositioned nuclei . Asynchronous cultures were grown at 25°C and shifted to 30°C for 4 or 24 hours prior to analysis to increase the penetrance of the nuclear mispositioning phenotype . Although wild type and dbf4 mutant cells did not misorient their nuclei ( Figure 7C ) , both kar9Δ and kar9Δ dbf4-NΔ109 strains had approximately equal number of cells with nuclear positioning defects . Importantly , deletion of the Dbf4p PIR resulted in a 2 to 3-fold increase ( to 6% ) in binucleate , anucleate and multinucleate cells at 4 hours ( Figure 7A , C , D ) , which was CDC5-dependent , suggesting that premature mitotic exit occurred . At 24 hours the number of aberrant chromosome segregation events due to loss of the Dbf4p PIR increased six-fold ( 12% ) relative to kar9Δ alone ( 2% ) . Since exit from mitosis causes spindle disassembly , we quantitated the spindle morphology in cells containing correctly segregated nuclei ( between the mother and daughter cells ) and in those cells where anaphase had initiated solely within the mother cell . Although kar9Δ and kar9Δ dbf4–NΔ109 cells had similar frequencies of intact or disassembled spindles when nuclear division proceeded normally , kar9Δ dbf4-NΔ109 cells showed a three-fold increase in spindle disassembly within the mother cell ( leading to a bi-nucleate mother cell ) compared to kar9Δ single mutants at 4 hours ( Figure 7D ) . These data indicate that mitotic exit occurs in these cells in the absence of the Dbf4p PIR . Similarly , we tested whether deletion of the Dbf4p PIR allowed premature mitotic exit in cells disrupted in the dynein pathway ( Figure 7B , E , F ) . The spindle-positioning defect associated with deletion of DYN1 is especially prominent at low temperatures . Although after 8 hours at 14°C only ∼3% of dyn1Δ cells had exited mitosis inappropriately , deletion of the Dbf4p PIR resulted in a 3- to 4-fold increase in mitotic exit as evidence by the appearance of multinucleate and anucleate cells ( Figure 7E ) and this frequency was increased at 24 hours . For comparison , a dyn1Δ bub2Δ strain had a similar but higher frequency of segregation defects ( Figure 7E , F ) . Therefore , the bypass of the SPOC following loss of the Dbf4p-Polo interaction is comparable to deletion of the MEN inhibitor BUB2 . As with the kar9Δ mutant , we observed that a higher percentage of dyn1Δ dbf4-NΔ109 cells compared to dyn1Δ single mutants that divided nuclei within the mother cell had disassembled their spindles ( Figure 7F ) . These observations indicate that deletion of the Dbf4p N-terminus ( including the PIR ) overrides the mitotic arrest normally activated by the SPOC .
The N-terminal third of Dbf4p encodes multiple functions: a destruction box ( residues 62–70 ) , two putative nuclear localization signals ( residues 55–61 , 251–257 ) and a BRCT-like domain ( residues ∼117–220 ) . Nonetheless the 265 N-terminal amino acids of Dbf4p are not essential as long as a nuclear localization signal is present [26] . Deletion of the Dbf4p N-terminus through the PIR has no observable effect on growth , viability , or cell cycle progression either under normal growth conditions or in the presence of replication or spindle poisons ( [26] and data not shown ) . Here , we discovered an interaction between Dbf4p and the Polo PBD that mapped to a short sequence of ∼40 amino acids preceding the BRCT-like domain . Although a two-hybrid interaction between Polo and Dbf4p was reported before [39] , the significance of this interaction was not determined . The Polo PBD functions as a module for binding phosphorylated proteins and thereby targets Polo to its cellular substrates [57] , [59] . The question naturally arises as to whether phosphorylation of the Dbf4p PIR is required for PBD binding . Currently , our observations suggest that phosphorylation is not required . A polo-box binding consensus sequence ( S ( pS/pT ) P/X ) is not present within this region of Dbf4p [59] and mutation of all putative serine and threonine residues within the Dbf4p PIR did not significantly diminish the interaction in the two-hybrid assay . These data suggest that phosphorylation of Dbf4p is not crucial for Polo binding . Our observation that the Dbf4p PIR purified from E . coli directly interacted with the Polo PBD also supports this notion . Thus phosphorylation of Dbf4p was not critical for binding to Polo in vitro , but we cannot exclude the possibility that phosphorylation contributes to Dbf4p-Polo binding in vivo when the two proteins are present at physiological concentrations . The finding that deletion of the Dbf4p PIR significantly suppressed the ts phenotype of cdc5-1 suggests that Cdc7p-Dbf4p inhibits Polo during the normal cell cycle and perhaps during periods of replication stress , when Cdc7p-Dbf4p is stabilized [17] . Our data clearly demonstrate a role for Dbf4p in inhibiting mitotic exit , since loss of the Dbf4p PIR suppressed both cdc5 and dbf2 ts mutants and allowed sustained Cdc14p phosphatase release and cytokinesis in the cdc5-1 mutant at the NPT . However , during an unperturbed cell cycle the absence of this regulation had little impact . This is likely attributable to the fact that the cell cycle regulation of Polo activity is complex and modulated by multiple cell cycle checkpoints . Since dbf4-NΔ109 bub2Δ double mutants were more sensitive to growth on spindle poisons than either mutant alone ( Figure S3 ) , the Dbf4p-Polo and Bfa1p-Bub2p pathways may work together to suppress premature activation of the mitotic exit network . It was shown very recently that increased expression of a non-destructible form of Dbf4p ( Dbf4-NΔ65p ) could delay rDNA segregation when Clb5p was also stabilized [16] . This raises the possibility that Dbf4p inhibits Cdc14p release via the FEAR pathway under some circumstances and is consistent with our data showing that Dbf4p is a Polo inhibitor . However , we found no evidence that the dbf4-NΔ109 allele promoted premature rDNA segregation ( a FEAR pathway event ) in the cdc5-1 mutant ( Figure S4 ) . Similarly , we found no evidence that dbf4-NΔ109 caused premature Cdc14p release in a strain deleted for FOB1 , which is thought help sequester Cdc14p in the nucleolus ( Figure S4 ) . These data suggest that Dbf4p is not specifically inhibiting the FEAR pathway . The budding yeast SPOC prevents premature exit from mitosis in part by inducing Kin4p phosphorylation of Bfa1p , which antagonizes the Polo-dependent inhibition of the Bfa1p-Bub2p GAP [8] , [9] , [55] . Premature exit from mitosis in dbf4-NΔ109 dyn1 and dbf4-NΔ109 kar9 double mutants suggests that Dbf4p regulation of Polo is critical for robust cell cycle control in response to nuclear mispositioning . What remains unclear however , is whether this Dbf4p activity is regulated following activation of the SPOC . For instance , Cdc7p-Dbf4p may inhibit Polo to buffer against premature release of Cdc14p during late S-phase or early M-phase whether or not the SPOC is activated . Therefore , in the absence of the Dbf4p-Polo regulation Polo may prematurely activate the MEN before a nuclear orientation defect is sensed by the SPOC . Following APC-Cdc20p and APC-Cdh1p activation during anaphase onset and exit , we suggest that degradation of Dbf4p provides a positive feedback loop for full Polo activation of the MEN and ultimately , cytokinesis . Dbf4p did not influence Polo protein levels or overall kinase activity . Dbf4p nonetheless inhibited Polo activity since dbf4 mutants unable to interact with Polo significantly suppressed the cdc5-1 temperature sensitivity . Since the cdc5-1 allele retains significant Polo protein expression , Bfa1p phosphorylation and overall Polo kinase activity at the non-permissive temperature [37] , [63] , the primary cdc5-1 MEN defect is in a Bfa1p-independent requirement for MEN pathway activation , perhaps Dbf2p activation or some other MEN-dependent step . Our observation that the dbf2-1 temperature sensitivity was also suppressed by loss of the Dbf4p PIR suggests that Dbf4p may specifically inhibit Polo activation of Dbf2p kinase independently of Bfa1p-Bub2p phosphorylation . Dbf2p was recently shown to promote cytoplasmic Cdc14p localization following MEN activation [69] . In addition , deletion of both BUB2 and the DBF4 PIR suppressed the cdc5-2 ts better than either single mutant alone . In other words , since dbf4-NΔ109 further suppressed the ts phenotype of a cdc5-2 bub2Δ strain , this supports the contention that Dbf4p regulates MEN activity independently of Bfa1p-Bub2p . Deletion of the Dbf4p PIR also did not allow rebudding in the presence of spindle poisons as seen in bub2Δ mutants , again suggesting that Dbf4p plays a minor or redundant role to inhibit Polo phosphorylation of Bfa1p-Bub2p . Cdc7p-Dbf4p kinase phosphorylated the Polo PBD in vitro suggesting that Cdc7p-Dbf4p may antagonize Polo substrate binding . This possibility is consistent with the requirement for the PBD for targeting Polo to sites of MEN activity . Polo , Cdc15p , Dbf2p and Bfa1p-Bub2p are localized to SPB prior to activation of the mitotic exit network [47] . Thus , we favor a model whereby Cdc7p-Dbf4p kinase inhibits precocious Polo binding to critical MEN substrate ( s ) by phosphorylating the Polo PBD . It will be interesting to investigate whether the Cdc7p-Dbf4p inhibition of Polo is regulated by cell cycle checkpoints and to determine the precise activity of Polo that is affected .
Strains and plasmids used in this study are listed in Tables S1 and S2 , and supplemental methods in the Text S1 file . PJ69-4a cells ( MATa trp1-901 leu2-3 , -112 ura3-52 his3-200 gal4Δ gal80Δ LYS2::GAL1-HIS3 GAL2-ADE2 met::GAL7-lacZ ) were used for two-hybrid experiments . All other strains were derivatives of W303 ( MATa ade2-1 trp1-1 can1-100 leu2-3 , -112 his3-11 , -15 ura3-1 ) . Construction of Dbf4p N-terminal truncation mutants was previously described [26] . Cdc14-EGFP was constructed as described [70] . bub2Δ strains were created by replacement of the BUB2 ORF via homologous recombination with SacI-ClaI bub2Δ::URA3 fragment from pTR24 ( A . Hoyt ) . The BAR1 ORF was deleted by homologous recombination with linearized pZV77 containing bar1::LEU2 ( B . Futcher ) . Construction of KAR9 and DYN1 deletions were previously described [8] . For yeast two-hybrid analyses , a Gal4 DNA binding domain ( GBD ) fusion to Dbf4p67–227 was constructed by PCR amplification of Dbf4p residues 67–227 ( NcoI-PstI ) and cloned into pGBKT7 ( Clonetech ) . Deletion of 71 bp within the ADH1 promoter sequence of pGBKT7 ( -647 to -717 from the ATG ) removed a Rap1p binding site and reduced the strength of GBD-Dbf4p67–227 expression ( which was otherwise lethal ) to give pCG60 . Point mutations and deletions were generated by site-directed mutagenesis using the QuikChange system ( Stratagene ) . For all mutations , the entire coding sequence was verified by DNA sequencing . Construction of baculovirus plasmids encoding WT , NΔ65Dbf4p , NΔ109Dbf4p , NΔ221Dbf4p and HA-Cdc7p was previously described [26] . The baculovirus transfer plasmid containing 3Myc-NΔ65Polo was constructed in pAcSG2 ( BD Biosciences ) . High-titer baculoviruses were generated by transfection of Sf9 cells using the BaculoGold kit ( BD Biosciences ) followed by plaque purification and virus amplification . For in vitro interaction assays , DNA encoding Polo amino acids 357–705 were PCR amplified with BamHI-XmaI linkers and cloned into pGEX-KG for expression of GST-PoloPBD in E . coli . The region encoding Dbf4p amino acids 66–109 was PCR amplified from pMW489 with BsaI-BamHI linkers and cloned into pSUMO ( LifeSensors Inc . ) for expression of Sumo-Dbf4p67–109 . Cells were cultured in YPD ( 1% yeast extract , 2% bacto peptone , 2% glucose ) . Synchronous G1 cultures were obtained after addition of 5 µg/ml ( 0 . 1 µg/ml in bar1Δ cells ) alpha-factor to cells for 3 hours . DNA content was analyzed by flow cytometry as previously described [17] . Drugs were added directly to plates immediately before pouring . PJ69-4a cells containing pCG60 were transformed with a S . cerevisiae two-hybrid library . Interacting clones were recovered on medium lacking tryptophan , leucine , and histidine but containing 2 mM 3-aminotriazole at 30°C . Positive interactors were streak-purified and also tested for ADE2 reporter activity . Prey plasmids that activated both HIS3 and ADE2 expression were confirmed by retransformation in PJ69-4a and then sequenced . To quantify two-hybrid interactions , co-transformed cells were spotted at ten-fold serial dilutions on selective media and grown for 2–3 days . Yeast protein extracts were prepared for Western blotting by trichloroacetic acid extraction [71] or for immunoprecipitation ( IP ) by bead-beating in NP-40 lysis buffer ( 20 mM Tris-HCl , 150 mM NaCl , 0 . 5% NP-40 and 1 mM EGTA ) . HA-tagged proteins were immunoprecipitated using anti-HA monoclonal antibody ( 12CA5 ) conjugated to protein A-Sepharose . Blots were probed in phosphate-buffered saline containing 0 . 1% Tween and 1% dried milk . 12CA5 ( 1∶1000 ) was used to detect HA-tagged proteins , 9E10 ( 1∶1000 ) to detect Myc-tagged proteins , and polyclonal sera against Cdc7p ( 1∶4000 ) and Dbf4p ( 1∶1000 ) were used to detect those proteins . Sf9 cells were co-infected with HA-Cdc7p , 3Myc-NΔ65Polo and Dbf4p derivatives and then immunoprecipitated as previously described [26] . Whole cell extracts and IPs were probed with polyclonal antibodies against Cdc7p ( 1∶4000 ) and Dbf4p ( 1∶1000 ) as described above . 3Myc-NΔ65Polo was probed with 9E10 monoclonal antibody against Myc ( 1∶1000 ) . Cdc7p-Dbf4p kinase was purified as described [17] . GST or GST-PBD was induced in BL21 cells for 3 hours at 37°C using 0 . 5 mM IPTG . Cells were sonicated in PBS containing 1% Triton X-100 and GST proteins were purified from soluble extracts by binding to glutathione-agarose ( Amersham ) , elution in ( 20 mM Tris-HCl , 150 mM NaCl , 1 mM EDTA and 10% glycerol ) containing 5 mM glutathione followed by dialysis against the same buffer . 6His-tagged Sumo and Sumo-Dbf4p were expressed in BL21 cells and extracted in HEPES extraction buffer ( 50 mM HEPES-KOH , pH 7 . 5 , 150 mM NaCl , 2 M Urea and 10% glycerol ) . Proteins were loaded onto a Ni++ column and washed ( 20 mM HEPES-KOH , pH 7 . 5 , 200 mM NaCl and 10% glycerol ) before elution using an imidazole gradient . For GST pull-downs , Sumo , Sumo-Dbf4p , GST and GST-Polo were incubated with glutathione-agarose in the presence of buffer ( 20 mM Tris-HCl pH 7 . 0 , 300 mM NaCl , 0 . 5% NP-40 and 1 mM EGTA ) for 1 hour at 4°C . The glutathione agarose beads were washed extensively and bound proteins separated on 12 . 5% SDS-PAGE gels . Blots were probed with polyclonal antisera raised against GST-Polo PBD ( 1∶1000 ) and yeast Smt3p ( Sumo ) ( 1∶1000 ) . For in vitro kinase assays , purified HACdc7p-Dbf4p ( 100 ng ) [17] was incubated with GST or GST-Polo PBD ( 300 ng ) at 30°C in kinase buffer ( 50 mM Tris-HCl pH 7 . 5 , 10 mM MgCl2 , 1 mM DTT , 100 µM ATP and 10 µCi of [γ-32P] ATP ) for 20 minutes . Proteins were separated on 10% SDS-PAGE and visualized by autoradiography . For direct fluorescence analysis of Cdc14-EGFP , cells were fixed in 3 . 7% formaldehyde at room temperature for 1 hour . DNA was stained using DAPI ( 1 µg/ml ) for 10 minutes at room temperature . For the experiments in Figure 5A , the absence of a distinct Cdc14-EGFP fluorescent signal ( but accompanied by diffuse nuclear and cytoplasmic fluorescence ) was scored as “released . ” Any cell that had a distinct Cdc14-EGFP nucleolar fluorescence was counted as sequestered . Spindle morphology was detected after spheroplasting cells and incubation in methanol/acetone prior to incubation with antibodies: rat anti-tubulin ( YOL1/34 Accurate Chemicals , 1∶10 ) and goat anti-rat FITC ( Jackson Immunoresearch , 1∶50 ) . Cells were imaged using a 60× objective .
|
Cdc7p-Dbf4p is a two-subunit enzyme required to copy the genetic material present on every chromosome in a process termed DNA replication . Dbf4p is an essential regulatory subunit of this enzyme that likely directs the Cdc7p subunit to its targets within the cell . We found that Dbf4p physically interacts with another protein called Polo that acts during mitosis , a later step in the cell cycle when the newly copied chromosomes are equally divided to mother and daughter cells . Polo is a master regulator of mitosis and impacts many other proteins required for cell division . We determined that Cdc7p-Dbf4p is a Polo inhibitor and , further , that Cdc7p-Dbf4p delayed or prevented chromosome segregation when errors occurred during the cell division process . Interestingly , Dbf4p may bind the Polo substrate-binding domain using a type of interaction not previously described . Thus , we have uncovered a new activity for Cdc7p-Dbf4p in the cell cycle to inhibit chromosome segregation , and these findings impact multiple fields that investigate how cells accurately copy and segregate their chromosomes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"molecular",
"biology/dna",
"replication",
"genetics",
"and",
"genomics/chromosome",
"biology",
"genetics",
"and",
"genomics/cancer",
"genetics",
"molecular",
"biology/chromatin",
"structure",
"molecular",
"biology/dna",
"repair"
] |
2009
|
Cdc7p-Dbf4p Regulates Mitotic Exit by Inhibiting Polo Kinase
|
Detecting the targets of adaptive natural selection from whole genome sequencing data is a central problem for population genetics . However , to date most methods have shown sub-optimal performance under realistic demographic scenarios . Moreover , over the past decade there has been a renewed interest in determining the importance of selection from standing variation in adaptation of natural populations , yet very few methods for inferring this model of adaptation at the genome scale have been introduced . Here we introduce a new method , S/HIC , which uses supervised machine learning to precisely infer the location of both hard and soft selective sweeps . We show that S/HIC has unrivaled accuracy for detecting sweeps under demographic histories that are relevant to human populations , and distinguishing sweeps from linked as well as neutrally evolving regions . Moreover , we show that S/HIC is uniquely robust among its competitors to model misspecification . Thus , even if the true demographic model of a population differs catastrophically from that specified by the user , S/HIC still retains impressive discriminatory power . Finally , we apply S/HIC to the case of resequencing data from human chromosome 18 in a European population sample , and demonstrate that we can reliably recover selective sweeps that have been identified earlier using less specific and sensitive methods .
The availability of population genomic data has empowered efforts to uncover the selective , demographic , and stochastic forces driving patterns of genetic variation within species . Chief among these are attempts to uncover the genetic basis of recent adaptation [1] . Indeed , recent advances in genotyping and sequencing technologies have been accompanied by a proliferation of statistical methods for identifying recent positive selection [see 2 for recent review] . Most methods for identifying positive selection search for the population genetic signature of a “selective sweep” [3] , wherein the rapid fixation of a new beneficial allele leaves a valley of diversity around the selected site [4–6] , about which every individual in the population exhibits the same haplotype ( i . e . the genetic background on which the beneficial mutation occurred ) . At greater genetic distances , polymorphism recovers as recombination frees linked neutral variants from the homogenizing force of the sweep [4] . This process also produces an excess of low- and high-frequency derived alleles [7 , 8] , and increased allelic association , or linkage disequilibrium ( LD ) , on either side of the sweep [9] , but not across the two flanks of the sweep [10 , 11] . Selective fixation of de novo beneficial mutations such as described by Maynard Smith and Haigh [5] are often referred to as “hard sweeps . ” More recently , population geneticists have begun to consider the impact of positive selection on previously standing genetic variants [12 , 13] . Under this model of adaptation , an allele initially evolves under drift for some time , until a change in the selective environment causes it to confer a fitness advantage and sweep to fixation . In contrast to the hard sweep model , the selected allele is present in multiple copies prior to the sweep . Thus , because of mutation and recombination events occurring near the selected site during the drift phase , the region containing this site may exhibit multiple haplotypes upon fixation [14] . The resulting reduction in diversity is therefore less pronounced than under the hard sweep model [12 , 15] . For this reason sweeps from standing genetic variation are often referred to as “soft sweeps . ” Soft sweeps will not skew the allele frequencies of linked neutral polymorphisms toward low and high frequencies to the same extent as hard sweeps [16] , and may even present an excess of intermediate frequencies [17] . This mode of selection will also have a different impact on linkage disequilibrium: LD will be highest at the target of selection rather than in flanking regions [18] . In very large populations , selection on mutations that are immediately beneficial may also produce patterns of soft sweeps rather than hard sweeps , as the adaptive allele may be introduced multiple times via recurrent mutation before the sweep completes [14 , 19] . While this model of a soft sweep is similar to that of selection on standing variation in that it will produce additional haplotypes carrying the selected allele , there are differences in the patterns of polymorphism produced by these two types of soft sweeps [18 , 20] . Here , we examine only the model of selection on a single standing variant . Adaptation could proceed primarily through selection on standing variation if the selective environment shifts frequently relative to the time scale of molecular evolution , and if there is enough standing variation segregating in the population on which selection may act following such a shift [12 , 21] . However , it is important to note that selection on standing variation may produce a hard sweep of only one haplotype containing the adaptive mutation if this allele is present at low enough frequency prior to sweep [16 , 22] . In other words , the observation of hard sweeps may be consistent with selection on standing variation as well as selection on de novo mutations . For these and other reasons , there is some controversy over whether adaptation will result in soft sweeps in nature [22] . This could be resolved by methods that can accurately discriminate between hard and soft sweeps . To this end , some recently devised methods for detecting population genetic signatures of positive selection consider both types of sweeps [23–25] . Unfortunately , it may often be difficult to distinguish soft sweeps from regions flanking hard sweeps due to the “soft shoulder” effect [18] . Here we present a method that is able to accurately distinguish between hard sweeps , soft sweeps on a single standing variant , regions linked to sweeps ( or the “shoulders” of sweeps ) , and regions evolving neutrally . This method incorporates spatial patterns of a variety of population genetic summary statistics across a large genomic window in order to infer the mode of evolution governing a focal region at the center of this window . We combine many statistics used to test for selection using an Extremely Randomized Trees classifier [26] , a powerful supervised machine learning classification technique . We refer to this method as Soft/Hard Inference through Classification ( S/HIC , pronounced “shick” ) . By incorporating multiple signals in this manner S/HIC achieves inferential power exceeding that of any individual test . Furthermore , by using spatial patterns of these statistics within a broad genomic region , S/HIC is able to distinguish selective sweeps not only from neutrality , but also from linked selection with much greater accuracy than other methods . Thus , S/HIC has the potential to identify more precise candidate regions around recent selective sweeps , thereby narrowing down searches for the target locus of selection . Further , S/HIC’s reliance on large-scale spatial patterns makes it more robust to non-equilibrium demography than previous methods , even if the demographic model is misspecified during training . This is vitally important , as the true demographic history of a population sample may be unknown . Finally , we demonstrate the utility of our approach by applying it to chromosome 18 in the CEU sample from the 1000 Genomes dataset [27] , recovering most of the sweeps identified previously in this population through other methods; we also highlight a compelling novel candidate sweep in this population .
We sought to devise a method that could not only accurately distinguish among hard sweeps , soft sweeps , and neutral evolution , but also among these modes of evolution and regions linked to hard and soft sweeps , respectively [18] . Such a method would not only be robust to the soft shoulder effect , but would also be able to more precisely delineate the region containing the target of selection by correctly classifying unselected but closely linked regions . In order to accomplish this , we sought to exploit the impact of positive selection on spatial patterns of several aspects of variation surrounding a sweep . Not only will a hard sweep create a valley of diversity centered around a sweep , but it will also create a skew toward high frequency derived alleles flanking the sweep and intermediate frequencies at further distances [7 , 8] , reduced haplotypic diversity at the sweep site [24] , and increased LD along the two flanks of the sweep but not between them [10] . For soft sweeps , these expected patterns may differ considerably [14 , 16 , 18] , but also depart from the neutral expectation . While some of these patterns of variation have been used individually for sweep detection [e . g . 10 , 28] , we reasoned that by combining spatial patterns of multiple facets of variation we would be able to do so more accurately . To this end , we designed a machine learning classifier that leverages spatial patterns of a variety of population genetic summary statistics in order to infer whether a large genomic window recently experienced a selective sweep at its center . We accomplished this by partitioning this large window into adjacent subwindows , measuring the values of each summary statistic in each subwindow , and normalizing by dividing the value for a given subwindow by the sum of values for this statistic across all subwindows within the same window to be classified . Thus , for a given summary statistic x , we used the following vector: [x1∑ixix2∑ixi…xn∑ixi] where the larger window has been divided into n subwindows , and xi is the value of the summary statistic x in the ith subwindow . Thus , this vector captures differences in the relative values of a statistic across space within a large genomic window , but does not include the actual values of the statistic . In other words , this vector captures only the shape of the curve of the statistic x across the large window that we wish to classify . Our goal is to then infer a genomic region’s mode of evolution based on whether the shapes of the curves of various statistics surrounding this region more closely resemble those observed around hard sweeps , soft sweeps , neutral regions , or loci linked to hard or soft sweeps . In addition to allowing for discrimination between sweeps and linked regions , this strategy was motivated by the need for accurate sweep detection in the face of a potentially unknown nonequilibrium demographic history , which may grossly affect values of these statistics but may skew their expected spatial patterns to a much lesser extent . In total , we constructed these vectors for each of π [29] , θ^w [30] , θ^H [8] , the number of distinct haplotypes , average haplotype homozygosity , H12 and H2/H1 [24 , 31] , ZnS [9] , and the maximum value of ω [10] . Thus , we represent each large genomic window by the following vector , to which we refer as the feature vector: [π1∑iπiπ2∑iπi…πn∑iπiθ^w1∑iθ^wiθ^w2∑iθ^wi…θ^wn∑iθ^wiθ^H1∑iθ^Hiθ^H2∑iθ^Hi…θ^Hn∑iθ^Hi………ω1∑iωiω2∑iωi…ωn∑iωi] We sought to discriminate between hard sweeps , regions linked to hard sweeps , soft sweeps , regions linked to soft sweeps , and neutrally evolving regions on the basis of the values of the vectors defined above . While Berg and Coop [20] recently derived approximations for the site frequency spectrum ( SFS ) for a soft sweep under equilibrium population size , and π , the joint probability distribution of the values all of the above statistics at varying distances from a sweep is unknown . Moreover expectations for the SFS surrounding sweeps ( both hard and soft ) under nonequilibrium demography remain analytically intractable . Thus rather than taking a likelihood approach , we opted to use a supervised machine learning framework , wherein a classifier is trained from simulations of regions known to belong to one of these five classes . We trained an Extra-Trees classifier ( aka extremely randomized forest; [26] ) from coalescent simulations ( described below ) in order to classify large genomic windows as experiencing a hard sweep in the central subwindow , a soft sweep in the central subwindow , being closely linked to a hard sweep , being closely linked to a soft sweep , or evolving neutrally according to the values of its feature vector ( Fig 1 ) . Briefly , the Extra-Trees classifier is an ensemble classification technique that harnesses a large number classifiers referred to as decision trees . A decision tree is a simple classification tool that uses the values of multiple features for a given data instance , and creates a branching tree structure where each node in the tree is assigned a threshold value for a given feature . If a given data point’s ( or instance’s ) value of the feature at this node is below the threshold , this instance takes the left branch , and otherwise it takes the right . At the next lowest level of the tree , the value of another feature is examined . When the data instance reaches the bottom of the tree , it is assigned a class inference based on which leaf it has landed [32] . Typically , a decision tree is built according to an algorithm designed to optimize its accuracy [32] . The Extra-Trees classifier , on the other hand , builds a specified number of semi-randomly generated decision trees . Classification is then performed by simply taking the class receiving the most “votes” from these trees [26] , building on the strategy of random forests [33] . While individual decision trees may be highly inaccurate , the practice of aggregating predictions from many semi-randomly generated decision trees has been proved to be quite powerful [34] . In the following sections we describe our methodology for training , testing , and applying our Extra-Trees classifier for identifying positive selection . We simulated data for training and testing of our classifier using our coalescent simulator , discoal_multipop ( https://github . com/kern-lab/discoal_multipop ) . As discussed in the Results , we simulated training sets with different demographic histories ( S1 Table ) , and , for positively selected training examples , different ranges of selection coefficients ( α = 2Ns , where s is the selective advantage and N is the population size ) . For each combination of demographic history and range of selection coefficients , we simulated large chromosomal windows that we later subdivided into 11 adjacent and equally sized subwindows . We then simulated training examples with a hard selective sweep whose selection coefficient was uniformly drawn from the specified range , U ( αlow , αhigh ) . We generated 11 , 000 sweeps: 1000 where the sweep occurred in the center of the leftmost of the 11 subwindows , 1000 where the sweep occurred in the second subwindow , and so on . We repeated this same process for soft sweeps at each location; these simulations had an additional parameter , the derived allele frequency , f , at which the mutation switches from evolving under drift to sweeping to fixation , which we drew from U ( 0 . 05 , 0 . 2 ) , U ( 2/2N , 0 . 05 ) , or U ( 2/2N , 0 . 2 ) as described in the Results . For our equilibrium demography scenario , we drew the fixation time of the selective sweep from U ( 0 , 0 . 2 ) ×N generations ago , while for non-equilibrium demography the sweeps completed more recently ( see below ) . We also simulated 1000 neutrally evolving regions . Unless otherwise noted , for each simulation the sample size was set to 100 chromosomes . For each combination of demographic scenario and selection coefficient , we combined our simulated data into 5 equally-sized training sets ( Fig 1 ) : a set of 1000 hard sweeps where the sweep occurs in the middle of the central subwindow ( i . e . all simulated hard sweeps ) ; a set of 1000 soft sweeps ( all simulated soft sweeps ) ; a set of 1000 windows where the central subwindow is linked to a hard sweep that occurred in one of the other 10 windows ( i . e . 1000 simulations drawn randomly from the set of 10000 simulations with a hard sweep occurring in a non-central window ) ; a set of 1000 windows where the central subwindow is linked to a soft sweep ( 1000 simulations drawn from the set of 10000 simulations with a flanking soft sweep ) ; and a set of 1000 neutrally evolving windows unlinked to a sweep . We then generated a replicate set of these simulations for use as an independent test set . We used the python scikit-learn package ( http://scikit-learn . org/ ) to train our Extra-Trees classifier and to perform classifications . Given a training set , we trained our classifier by performing a grid search of multiple values of each of the following parameters: max_features ( the maximum number of features that could be considered at each branching step of building the decision trees , which was set to 1 , 3 , n , or n , where n is the total number of features ) ; max_depth ( the maximum depth a decision tree can reach; set to 3 , 10 , or no limit ) , min_samples_split ( the minimum number of training instances that must follow each branch when adding a new split to the tree in order for the split to be retained; set to 1 , 3 , or 10 ) ; min_samples_leaf . ( the minimum number of training instances that must be present at each leaf in the decision tree in order for the split to be retained; set to 1 , 3 , or 10 ) ; bootstrap ( a binary parameter that governs whether or not a different bootstrap sample of training instances is selected prior to the creation of each decision tree in the classifier ) ; criterion ( the criterion used to assess the quality of a proposed split in the tree , which is set to either Gini impurity [35] or to information gain , i . e . the change in entropy [32] ) . The number of decision trees included in the forest was always set to 100 . After performing a grid-search with 10-fold cross validation in order to identify the optimal combination of these parameters , we used this set of parameters to train the final classifier . We used the scikit-learn package to assess the importance of each feature in our Extra-Trees classifiers . This is done by measuring the mean decrease in Gini impurity , multiplied by the average fraction of training samples that reach that feature across all decision trees in the classifier . The mean decrease in impurity for each feature is then divided by the sum across all features to give a relative importance score , which we show in S2 Table . We also show values of Extra-Trees classifier parameters resulting from grid searchers in S3 Table . We compared the performance of our classifier to that of various other methods . First , we examined two population genetic summary statistics: Tajima’s D [36] and Kim and Nielsen’s ωmax [10] ( which we refer to as ω for simplicity ) , calculating their values in each subwindow within each large simulated chromosome that we generated for testing ( see above ) . We also used Nielsen et al . ’s composite-likelihood ratio test , referred to as CLR or SweepFinder [28] , which searches for the spatial skew in allele frequencies expected surrounding a hard selective sweep . When testing SweepFinder’s ability to discriminate between modes of evolution within larger regions , we computed the composite-likelihood ratio between the sweep and neutral models at 200 sites across each of the 11 subwindows of our large simulated test regions , taking the maximum CLR value . The only training necessary for SweepFinder was to specify the neutral site frequency spectrum . Next , we used scikit-learn to implement Ronen et al . ’s [37] SFselect , a support vector machine classifier that discriminates between selection and neutrality on the basis of a region’s binned and weighted SFS . In our implementation we collapsed the SFS into 10 bins as suggested by Ronen et al . , and also added soft sweeps as a third class ( in addition to hard sweeps and neutrality ) , using Knerr et al . ’s [38] method for extending a binary classifier to perform multi-class classification . We trained this classifier from simulated data following the same demographic and selective scenarios used to train our own classifier , and with the same number of simulated training instances , but these simulations encapsulated much smaller regions ( equivalent to the size of one of our eleven subwindows ) . To avoid confusion with the original SFselect , which only handles hard sweeps , we refer to this implementation as SFselect+ . For further comparisons , we also trained a support vector machine using a vector of two statistics: the maximum values of the SweepFinder CLR statistic and ω ( a subset of the features in the Pavlidis et al . ’s SVM [39] ) . We refer to this method as CLR+ω , and trained it in the same manner as SFselect+ , except for the different feature vector . We also tested the performance of the evolBoosting [40] , an R package which uses an machine learning approach called boosting [41] to classify genomic windows as sweeps or neutral on the basis of several statistics , including Tajima’s D , Fay and Wu’s H [8] , integrated haplotype homozygosity ( iHH; [42] ) , and several others . Boosting was also recently used by Pybus et al . [43] to accurately detect hard and partial sweeps and make coarse inferences about sweep ages . Like S/HIC , this method uses a vector of the values of each of these statistics from several subwindows surrounding the region being classified . However , unlike S/HIC , this method does not take the relative values of these statistics in each subwindow divided by the sum across all subwindows , instead just taking the value of the statistic measured in that subwindow . As with SFselect , we extend this method to discriminate between hard sweeps , soft sweeps , and neutral windows . This was done by first training a classifier to distinguish between sweeps ( hard and soft , balanced in number within the training set ) from neutral windows and secondarily training a classifier to distinguish between hard and soft sweeps . Finally , we implemented a version of Garud et al . ’s [24] scan for hard and soft sweeps . Garud et al . ’s method uses an Approximate Bayesian Computation-like approach to calculate Bayes Factors to determine whether a given region is more similar to a hard sweep or a soft sweep by performing coalescent simulations . For this we performed simulations with the same parameters as we used to train SFselect+ , but generated 100 , 000 simulations of each scenario in order to ensure that there was enough data for rejection sampling . We then used two statistics to summarize haplotypic diversity within these simulated data: H12 and H2/H1 [31] . All simulated regions whose vector [H12 H2/H1] lies within a Euclidean distance of 0 . 1 away from the vector corresponding to the data instance to be classified are then counted [24] . The ratio of simulated hard sweeps to simulated soft sweeps within this distance cutoff is then taken as the Bayes Factor . Note that Garud et al . restricted their analysis of the D . melanogaster genome to only the strongest signals of positive selection , asking whether they more closely resembled hard or soft sweeps . Therefore when testing the ability of Garud et al . ’s method to distinguish selective sweeps from both linked and neutrally evolving regions , we used large simulated windows and simply examined the value of H12 within the subwindow that exhibited the largest value in an effort to mimic their strategy of using H12 peaks [24] . We summarized each method’s power using the receiver operating characteristic ( ROC ) curve , making these comparisons for the following binary classification problems: discriminating between hard sweeps and neutrality , between hard sweeps and soft sweeps , between selective sweeps ( hard or soft ) and neutrality , and between selective sweeps ( hard or soft ) and unselected regions ( including both neutrally evolving regions and regions linked to selective sweeps ) . For each of these comparisons we constructed a balanced test set with a total of 1000 simulated regions in each class , so that the expected accuracy of a completely random classifier was 50% , and the expected area under the ROC curve ( AUC ) was 0 . 5 . Whenever the task involved a class that was a composite of two or more modes of evolution , we ensured that the test set was comprised of equal parts of each subclass . For example , in the selected ( hard or soft ) versus unselected ( neutral or linked selection ) test , the selected class consisted of 500 hard sweeps and 500 soft sweeps , while the unselected class consisted of 333 neutrally evolving regions , 333 regions linked to hard sweeps , and 333 regions linked to soft sweeps ( and one additional simulated region from one of these test sets randomly selected , so that the total size of the unselected test set was 1000 instances ) . As with our training sets , we considered the true class of a simulated test region containing a hard ( soft ) sweep occurring in any but the central subwindow to be hard-linked ( soft-linked ) —even if the sweep occurred only one subwindow away from the center . The ROC curve is generated by measuring performance at increasingly lenient thresholds for discriminating between the two classes . We therefore required each method to output a real-valued measure proportional to its confidence that a particular data instance belongs the first of the two classes . For S/HIC , we used the posterior classification probability from the Extra-Trees classifier obtained using scikit-learn’s predict_proba method . For SFselect+ , we used the value of the SVM decision function . For SweepFinder , we used the composite likelihood ratio . For Garud et al . ’s method , we used the fraction of accepted simulations ( i . e . within a Euclidean distance of 0 . 1 from the test instance ) that were of the first class: for example , for hard vs . soft , this is the number of accepted simulations that were hard sweeps divided by the number of accepted simulations that were either hard sweeps or soft sweeps . For Tajima’s D [36] and Kim and Nielsen’s ω [10] , we simply used the values of these statistics . To examine the power and sensitivity of S/HIC under non-equilibrium demographic histories , we simulated training and test datasets from a few scenarios that might be relevant to researchers . Firstly we examined the power of our method under two complex population size histories that are relevant to humans . Secondly we examined the case of simple population bottlenecks , as might be common in populations that have recently colonized new locales , using two levels of bottleneck severity . We simulated training and test datasets from Tennessen et al . ’s [44] European demographic model ( S1 Table ) . This model parameterizes a population contraction associated with migration out of Africa , a second contraction followed by exponential population growth , and a more recent phase of even faster exponential growth . Values of θ and ρ = 4Nr were drawn from prior distributions ( S1 Table ) , allowing for variation within the training data , whose means were selected from recent estimates of human mutation [45] and recombination rates [46] , respectively . For simulations with selection , we drew values of α from U ( 5 . 0×103 , 5 . 0×105 ) , and drew the fixation time of the sweeping allele form U ( 0 , 51 , 000 ) years ago ( i . e . the sweep completed after the migration out of Africa ) . We also generated simulations of Tennessen et al . ’s African demographic model , which consists of exponential population growth beginning ~5 , 100 years ago ( S1 Table ) . We generated two sets of these simulations: one where α was drawn from U ( 5 . 0×104 , 5 . 0×105 ) , and one with α drawn from U ( 5 . 0×104 , 5 . 0×105 ) . The sample size of these simulated data sets was set to 100 chromosomes . These two sets were then combined into a single training set . For these simulations , the sweep was constrained to complete some time during the exponential growth phase ( no later than 5 , 100 years ago ) . Finally , we examined two models with a population size bottleneck . The first was taken from Thornton and Andolfatto [47] , and models the demographic history of a European population sample of D . melanogaster ( S1 Table ) . This model consists of a population size reduction 0 . 044×2N generations ago to 2 . 9% of the ancestral population size , and then 0 . 0084×2N generations ago the population recovers to its original size . The second bottleneck model we used was identical except the population contraction was less severe ( reduction to 29% of the ancestral population size ) . For sweep simulations under both of these bottleneck scenarios , we drew α from U ( 1 . 0×102 , 1 . 0×104 ) . For all of our non-equilibrium demographic histories , when simulating soft sweeps on a previously standing variant , we drew the derived allele frequency at the onset of positive selection from U ( 2/2N , 0 . 2 ) . For each demographic model in S1 Table , we show in S1 Fig the means and standard deviations of Tajima’s D across 11 windows at increasing distances from a selective sweep ( for one possible sweep scenario ) , as well as values from neutrally evolving windows for comparison . For each demographic model in S1 Table , we show in S1 Fig the means and standard deviations of Tajima’s D across 11 windows at increasing distances from a selective sweep ( for one possible sweep scenario ) , as well as values from neutrally evolving windows for comparison . We applied our method to chromosome 18 from the Phase I data release from the 1000 Genomes project [27] . We restricted this analysis to the CEU population sample ( individuals with European ancestry , sampled from Utah ) , and trained S/HIC using data from the European demographic model described above . After training this classifier , we prepared data from chromosome 18 in CEU for classification . Prior to constructing feature vectors , we first performed extensive filtering for data quality . First , we masked all sites flagged by the 1000 Genomes Project as being unfit for population genetic analyses due to having either limited or excessive read-depth or poor mapping quality ( according to the strictMask files for the Phase I data set which are available at http://www . 1000genomes . org/ ) . In order to remove additional sites lying within repetitive sequence wherein genotyping may be hindered , we eliminated sites with 50 bp read mappability scores less than one [48] and also sites masked by RepeatMasker ( http://www . repeatmasker . org ) . Finally , we attempted to infer the ancestral state at each remaining site , using the chimpanzee [49] and macaque [50] genomes as outgroups . For each site , if the chimpanzee and macaque genomes agreed , we used this nucleotide as our inferred ancestral state . If instead only the chimpanzee or the macaque genome had a nucleotide aligned to the site , we used this base as our inferred ancestral state . For sites that were SNPs , we also required that the inferred ancestral state matched one of the two human alleles . For all cases where these criteria were not met , we discarded the site . After data filtering , we calculated summary statistics within adjacent 200 kb windows across the entire chromosome . Importantly , we divided the values of each summary statistic by the number of sites in the window , ignoring sites filtered as discussed above ( i . e . π summarizes average nucleotide diversity per site rather than total diversity in the subwindow ) . Windows with >50% of sites removed during the filtering processes were omitted from our analysis . These two steps limited the effect of variation in the number of unfiltered sites from window to window our classification . For the remaining windows , we used a sliding window approach with a 2 . 2 Mb window and a 200 kb step size to calculate the feature vector in the same manner as for our simulated data , and then applied S/HIC to this feature vector to infer whether the central subwindow of this 2 . 2 Mb region contained a hard sweep , a soft sweep , was linked to a hard sweep , linked to a soft sweep , or evolving neutrally . Visualization of candidate regions was performed using the UCSC Genome Browser [51] . We used hg19 coordinates for all of our analyses using human data . Our classification tool is available at https://github . com/kern-lab/shIC , along with software for generating the feature vectors used in this paper ( either from simulated training data or from real data for classification ) .
The most basic task that a selection scan must be able to perform is to distinguish between hard sweeps and neutrally evolving regions , as the expected patterns of nucleotide diversity , haplotypic diversity , and linkage disequilibrium produced by these two modes of evolution differ dramatically [5 , 8 , 10 , 18 , 24 , 52] . We therefore begin by comparing S/HIC’s power to discriminate between hard sweeps and neutrality to that of several previously published methods: these include SweepFinder [aka CLR; 28] , SFselect [37] , Garud et al . ’s haplotype approach using the H12 and H2/H1 statistics [24] , Tajima’s D [36] , and Kim and Nielsen’s ω [10] , evolBoosting [40] , and a support vector machine implemented that uses CLR and ω statistics ( Methods ) . We extended SFselect and evolBoosting to allow for soft sweeps ( Methods ) , and therefore refer to this classifier as SFselect+ and evolBoosting+ in order to avoid confusion . We summarize the power of each of these approaches with the receiver operating characteristic ( ROC ) curve , which plots the method’s false positive rate on the x-axis and the true positive rate on the y-axis ( Methods ) . Powerful methods that are able to detect many true positives with very few false positives will thus have a large area under the curve ( AUC ) , while methods performing no better than random guessing are expected to have an AUC of 0 . 5 . We began by assessing the ability of these tests to detect selection in populations with constant population size and no population structure . First , we used test sets where the selection coefficient α = 2Ns was drawn uniformly from U ( 2 . 5×102 , 2 . 5×103 ) , finding that S/HIC achieved had perfect accuracy ( AUC = 1 . 0; S2A Fig ) , and that several other methods performed nearly as well . When drawing α from U ( 2 . 5×103 , 2 . 5×104 ) , every method had near perfect accuracy ( AUC>0 . 99 ) except H12 and ω ( S2B Fig ) . For weaker selection [α ~U ( 25 , 2 . 5×102 ) ] this classification task is more challenging , and the accuracies of most of the methods we tested dropped substantially . S/HIC , however , performed quite well , with an AUC of 0 . 9797 , slightly better than evolBoosting+ ( AUC = 0 . 9702 ) and SFselect+ ( AUC = 0 . 9683 ) , and substantially better than the remaining methods ( S2C Fig ) . Note that Garud et al . ’s H12 statistic performed quite poorly in these comparisons , especially in the case of weak selection . This is likely because the fixation times of the sweeps that we simulated ranged from 0 to 0 . 2×N generations ago , and the impact of selection on haplotype homozygosity decays quite rapidly after a sweep completes [18] . Indeed , H12 has been shown to have good power to detect recent sweeps [24] . For the above comparisons , our classifier , evolBoosting+ , and SFselect+ , and the SVM combining CLR and ω were trained with the same range of selection coefficients used in these test sets . Thus , these results may inflate the performance of these methods relative to other methods , which do not require training from simulated selective sweeps . If one does not know the strength of selection , one strategy is to train a classifier using a wide range of selection coefficients so that it may be able to detect sweeps of varying strengths [37] . We therefore combined the three training sets from the three different ranges of α described above into a larger training set consisting of sweeps of α ranging from as low as 25 to as high as 25 , 000 . This step was done not only for S/HIC , but also for SFselect+ , evolBoosting+ , and the CLR+ω SVM , and we use this approach for the remainder of the paper when using classifiers trained from constant population size data . When trained on a large range of selection coefficients , S/HIC still detected sweeps with α drawn from U ( 2 . 5×102 , 2 . 5×103 ) with perfect accuracy , as did SFselect+ ( Fig 2A ) . For stronger sweeps , we again had excellent accuracy ( AUC = 0 . 999; Fig 2B ) and outperformed all other methods except SFselect+ and SweepFinder ( AUC = 1 . 0 ) . For weaker sweeps our method had the highest accuracy ( AUC = 0 . 9772 for S/HIC , 0 . 9660 for SFselect+ , 0 . 9562 for evolBoosting+ , and lower for other methods; Fig 2C ) . Thus , S/HIC can distinguish hard selective sweeps of greatly varying strengths of selection from neutrally evolving regions as well as if not better than previous methods . In order to uncover the targets of recent selective sweeps and also determine which mode of positive selection was responsible , one must be able to detect the signatures of soft ( initial selected frequency f ~U ( 0 . 05 , 0 . 2 ) ) as well as hard sweeps and to distinguish between them . We therefore assessed the power of each method to distinguish sets of simulated selective sweeps consisting of equal numbers of hard and soft sweeps from neutral simulations , using the same training data ( for methods that require it ) as for the analysis in Fig 2 . For all ranges of selection coefficients , S/HIC has excellent power to distinguish hard and soft sweeps from neutrality; our AUC scores ranging from 0 . 9533 to 0 . 9862 , and are higher than every other method in each scenario ( S3 Fig ) . S/HIC also distinguishes hard sweeps from soft sweeps with accuracy similar to evolBoosting+ and SFselect+ , and these three methods perform better than each other method , except in the case of weak sweeps where SFselect+ and evolBoosting+ have slightly better power ( AUC = 0 . 0 . 7978 for S/HIC , versus 0 . 8066 and 0 . 8239 , for SFselect+ and evolBoosting+ , respectively; S4 Fig ) . The goal of genomic scans for selective sweeps is not merely to quantify the extent to which positive selection impacts patterns of variation , but also to identify the targets of selection in hopes of elucidating the molecular basis of adaption . Unfortunately , hitchhiking events can skew patterns of variation across large chromosomal stretches often encompassing many loci . Furthermore , this problem not only confounds selection scans by obscuring the true target of selection , it may also lead to falsely inferred soft sweeps as a result of the soft shoulder effect [18] . Our goal in designing S/HIC was to be able to accurately distinguish among positive selection , linked selection , and neutrality , thereby addressing both of these challenging problems . In order to assess the ability of our approach and other methods to perform this task , we repeated the test shown in S3 Fig , but this time we included regions linked to selective sweeps among the set of neutral test instances . Thus , we ask how well these methods distinguish genomic windows containing the targets of selective sweeps ( soft or hard ) from neutrally evolving windows or windows closely linked to sweeps . Encouragingly , we find that S/HIC is able confidently distinguish windows experiencing selective sweeps from linked as well as neutrally evolving regions ( Fig 3 ) —S/HIC achieves substantially higher accuracy than each other method ( AUC = 0 . 9593 or higher for all values of α , while no other method has AUC>0 . 91 for any α ) . As the selection coefficient increases , S/HIC’s performance increase relative to that of other methods is particularly pronounced ( Fig 3A and 3B ) , which is unsurprising because in these cases the impact of selection on variation within linked regions is much further reaching than for weak sweeps ( Fig 3C ) . While ROC curves provide useful information about power , a more complete view of our ability to distinguishing among hard sweeps , soft sweeps , linked selection , and neutrality can be obtained by asking how our classifier behaves at varying distances from selective sweeps . We directly compared our method’s ability to classify regions ranging from 5 subwindows upstream of a hard sweep to 5 subwindows downstream of a hard sweep to evolBoosting+ and SFselect+ which were the top performers among all other methods we had examined . For these simulations , each subwindow had a total recombination rate 4Nr = 80 , corresponding to 0 . 2 cM per subwindow when N = 10 , 000 . We then counted the fraction of simulations predicted to belong to each of our five classes ( hard , hard-linked , soft , soft-linked , and neutral ) or the three classes used for evolBoosting+ and SFselect+ . As shown in Fig 4 , we find that when α ranges from 250 to 2500 and there has been a hard sweep in the central window , all three methods recover the sweep with high frequency when examining the correct window ( >95% accuracy ) . However , as we move away from the selected site , a large number of windows are misclassified as hard sweeps by SFSelect+ and evolBoosting+ . For example , both methods misclassify nearly 50% of cases two windows away from the true sweep as hard sweeps , and most of the remaining examples as soft sweeps . In contrast , our method classifies <5% of these regions as sweeps , correctly classifying >93% of these windows as hard-linked instead . At a distance of 5 windows away from the sweep , SFselect+ classifies the majority of windows as soft sweeps and many others as hard sweeps , and evolBoosting+ exhibits a smaller but still prominent shoulder effect ( 21 . 2% of windows classified as soft and 12 . 7% as hard ) . Meanwhile , S/HIC classifies >95% of these windows as hard-linked , and <1% as sweeps of either mode . For soft sweeps , we have the highest sensitivity in the sweep window ( 78 . 6% for S/HIC versus 75 . 9% for SFselect+ and 60 . 8% for evolBoosting+ ) . We also narrow the target of selection down to a smaller region , as we classify the majority of flanking windows as soft-linked , while SFselect+ produces many soft sweep calls in these windows . The difference between S/HIC and these two other methods is amplified when testing these classifiers on stronger hard sweeps ( α ranging from 2 , 500 to 25 , 000 ) . Our classifier is better able to narrow down the selected region by classifying flanking windows as hard-linked , while SFselect+ and evolBoosting classifies the vast majority of simulations even 5 windows away from the target of selection as hard sweeps ( Fig 5 ) . SFselect+ and evolBoosting both have more sensitivity to detect hard sweeps when examining the correct window ( >99% versus 88 . 8% ) , as S/HIC misclassifies 10 . 9% of these stronger sweeps as hard-linked . On the other hand , S/HIC recover 91 . 8% of soft sweeps versus 87 . 7% for SFselect+ and 73 . 3% for evolBoosting+ , and correctly classifies the mode of selection more often than these methods . We also misclassify relatively few regions linked to soft sweeps as sweeps themselves ( ~16% when one window away , versus ~50% for SFselect+ and ~20% for evolBoosting+ ) . For weaker sweeps [α ~U ( 25 , 250 ) ] , the impact of selection on linked regions is reduced , and SFselect+ and evolBoosting+ call fewer false sweeps in linked regions than under stronger positive selection . However , S/HIC has greater sensitivity to both hard and soft sweeps at the correct window , and also misclassifies fewer flanking regions as sweeps ( S5 Fig ) . Across the entire range of selection coefficients , S/HIC mislabeled fewer neutral simulations as sweeps than SFselect+ , though evolBoosting+ had a slightly lower false positive rate . In summary , across all selection coefficients S/HIC has greater sensitivity than other methods to detect soft sweeps , and also for hard sweeps except when selection is very strong . Importantly , for both types of sweeps S/HIC will identify a smaller candidate region around the selective sweep than SFselect+ or evolBoosting+ . S/HIC is able to classify far fewer linked windows as selected because it has two classes for this purpose , hard-linked and soft-linked , that the other methods lack . Though SFselect+ could be improved by incorporating these classes , it may prove difficult to determine whether a window is selected or merely linked to a sweep on the basis of its SFS alone [18] , rather than examining larger scale spatial patterns of variation . evolBoosting+ fares better in this respect because it does incorporate spatial information . However , perhaps because it takes the true values of each statistic in each window rather than the relative values and also lacks “linked” classes , this method still experiences a much greater soft shoulder effect than S/HIC . Up until this point our model of selection on previously stranding variation specified an initial selected frequency , f , ranging from 0 . 05 to 0 . 2 . However , a large fraction of soft selective sweeps may begin the sweep phase at a lower frequency [13 , 16 , 20] . Therefore , in order to assess how our classifier performs when soft sweeps have a lower initial selected frequency , we repeated these analyses with f drawn from U ( 2/2N , 0 . 05 ) . Again , for all three ranges of the selection coefficient S/HIC has greater accuracy than any other method ( S6 Fig ) . When attempting to distinguish between hard sweeps and soft sweeps under this parameterization , performance was reduced considerably for all methods , and there was no clear winner across all strengths of selection . While S/HIC was not the top performer at this task , its AUC was within 5% of the highest score for each range of selection coefficients ( S7 Fig ) . Next , for S/HIC and each other method that requires training , we constructed a training set in the same manner as above but allowing f to range from U ( 2/2N , 0 . 2 ) , and we use this range of initial selected frequencies for all analyses presented below . While training S/HIC on these data , we ranked the importance of each feature’s contribution to our Extra-Trees classifier’s accuracy ( Methods ) , which we list in S2 Table . Generally , we find that features near the center of the window have a greater contribution , and that relative values of θ^w and π tend to have greater importance than other statistics . Non-equilibrium demographic histories have the potential to confound population genetic scans for selective sweeps [53 , 54] . We therefore sought to assess the power of S/HIC and other methods to detect selection occurring in populations experiencing dramatic changes in population size . To this end we trained and tested our classifiers under four demographic scenarios ( S1 Table ) : two simple population bottlenecks of varying severity ( one of which models European Drosophila ) , a model of recent exponential population size growth , and finally a more complex model that describes out-of-Africa populations of humans . The results from simulated data described above suggest that our method has the potential to identify selective sweeps and distinguish them from linked selection and neutrality with excellent accuracy . In order to demonstrate our method’s practical utility , we used it to perform a scan for positive selection in humans . In particular , we searched the 1000 Genomes Project’s CEU population sample ( European individuals from Utah ) for selective sweeps occurring after the migration out of Africa . We focused this search on chromosome 18 , where several putative selective sweeps have been identified in Europeans [57] . The steps we took to train our classifier and filter the 1000 Genomes data prior to conducting our scan are described in the Methods . In total , we examined 344 windows , each 200 kb in length . We classified 34 windows ( 9 . 9% ) as centered around a hard sweep , 22 ( 6 . 4% ) as linked to a hard sweep , 48 ( 14 . 0% ) as centered around a soft sweep , 89 ( 25 . 9% ) as linked to a soft sweep , and 151 ( 43 . 9% ) as neutral . Surprisingly , we infer that over 56% of windows lie within regions whose patterns of variation are affected by sweeps either within the window or in linked regions . This may imply that , given the genomic landscape of recombination in humans , even if selective events are somewhat rare [58] , they may nonetheless impact variation across large stretches of the genome . However , we cannot firmly draw this conclusion given the difficulty of distinguishing between linked selection and neutrality under the European demographic model ( Fig 7 ) . Encouragingly , our scan recovered 4 of the 5 putative sweeps on chromosome 18 in Europeans identified by Williamson et al . [57] using SweepFinder . These include CCDC178 ( which we classify as a hard sweep ) , DTNA ( which we classify as soft ) , CCDC102B ( hard ) , and the region spanning portions of CD226 and RTTN ( hard ) . In each of these loci , the windows that we predicted to contain the sweep overlapped regions of elevated composite likelihood ratio ( CLR ) values from SweepFinder [visualized using data from 59] . Although the CLR statistic is not completely orthogonal to the summary statistics we examine to perform our classifications , the close overlap that we observe between these two methods underscores our ability to precisely detect the targets of recent positive selection . We also identify a novel candidate hard sweep within L3MBTL4 , an apparent tumor suppressor gene that is often mutated , downregulated , or deleted in breast tumors [60] . As shown in Fig 8 , π , Tajima’s D , ZnS , ω , and the CLR statistic all show patterns strongly suggestive of a selective sweep within this gene . The complete set of coordinates of putative sweeps from this scan is listed in S4 Table . Next , we asked whether S/HIC recovered evidence of positive selection on the LCT ( lactase ) locus . Previous studies have found evidence for very recent and strong selection on this gene in the form population differentiation and long-range haplotype homozygosity [61–63] . Moreover , several variants in this region are associated with lactase persistence . Nielsen et al . ’s CLR has also identified this region [28] , but not consistently: Williamson et al . ’s [57] CLR scan did not detect a sweep at this locus , nor does a recent scan using the 1000 Genomes data [data from 59] . This may be expected , as the selection on lactase persistence alleles appears to have not yet produced completed sweeps . Overall , there is very strong evidence of recent and perhaps ongoing selection for lactase persistence in human populations relying on diary for nutrition . Like the SweepFinder CLR , S/HIC in its current form is also designed to detect completed sweeps . Nonetheless , we applied S/HIC to a 4 Mb region on chromosome 2 spanning LCT and neighboring loci . Consistent with previous studies , we find evidence of a selective sweep in a region upstream of LCT ( S13 Fig ) that contains a mutation associated with lactase persistence in Europeans [64] , though unlike Peter et al . [25] , we classify this sweep as soft . We also find evidence of a hard selective sweep upstream of LCT , suggesting that there may be additional targets of selection in this region of chromosome 2 . Consistent with this is the observation that our candidate window also overlaps a region identified by Green et al . [65] as having an excess of derived alleles in the human genome relative to the number observed in Neanderthal . The computational speed of this scan is largely governed by the amount of time spent calculating summary statistics to generate the feature vectors ( as well as simulating training data , if absent ) , as the training and classification tasks are relatively inexpensive ( typically requiring several minutes for the former and only seconds for the later ) . The approximate runtime for calculating our set of summary statistics within the 4 Mb region encompassing LCT is ~30 minutes ( using code from https://github . com/kern-lab/shIC ) . Thus , if a compute cluster is available , one can subdivide the genome into segments of this size and perform these calculations in parallel , and classify every window in the human genome in under an hour .
Detecting the genetic targets of recent adaptation and the mode of positive selection acting on them—selection on de novo mutations versus previously standing variants—remains an important challenge in population genetics . The majority of efforts to this end have relied on population genetic summary statistics designed to uncover loci where patterns of allele frequency [e . g . 8 , 36 , 66] or linkage disequilibrium [e . g . 9 , 10] depart from the neutral expectation . Recently , powerful machine learning techniques have begun to be applied to this problem , showing great promise [18 , 37 , 39 , 40 , 43] . Here we have adopted a machine learning approach to develop S/HIC , a method designed to not only uncover selective sweeps , but to distinguish them from regions linked to sweeps as well as neutrally evolving regions , and to identify the mode of selection . This is achieved by examining spatial patterns of a variety of population genetic summary statistics that capture different facets of variation across a large-scale genomic region . Currently , this method examines the values of nine statistics across eleven different windows in infer the mode of evolution in the central window—this makes for a total of 99 different values considered by the classifier . By leveraging all of this information jointly , our Extra-Trees classifier is able to detect selection with accuracy unattainable by methods examining a single statistic , underscoring the potential of the machine learning paradigm for population genetic inference . Indeed , on simulated datasets with constant population size , S/HIC has power matching or exceeding previous methods when linked selection is not considered ( i . e . the sweep site is known a priori ) , and vastly outperforms them under the more realistic scenario where positive selection must be distinguished from linked selection as well as neutrality . We argue that the task of discriminating between the targets of positive selection and linked but unselected regions is an extremely important and underappreciated problem that must be solved if we hope to identify the genetic underpinnings of recent adaptation in practice . This is especially so in organisms where the impact of positive selection is pervasive , and therefore much of the genome may be linked to recent selective sweeps [e . g . 67] . A method that can discriminate between sweeps and linked selection would have three important benefits . First , it will reduce the number of spurious sweep calls in flanking regions , thereby mitigating the soft shoulder problem [18] . Second , such a method would have the potential to narrow down the candidate genomic region of adaptation . Third , such a method would be able to find those regions least affected by linked selection , which themselves might act as excellent neutral proxies for inference into demography or mutation . We have shown that S/HIC is able to distinguish among selection , linked selection , and neutrality with remarkable power , granting it the ability to localize selective sweeps with unrivaled accuracy and precision , demonstrating its practical utility . While S/HIC performs favorably to other approaches under the ideal scenario where the true demographic history of the population is known , in practice this may not always be the case . However , because our method relies on spatial patterns of variation , we are especially robust to demography: if the demographic model is misspecified , the disparity in accuracy between S/HIC and other methods is even more dramatic . For example , if we train S/HIC with simulated datasets with constant population size , but test it on simulated population samples experiencing recent exponential growth ( e . g . the African model from ref . [44] ) , we still identify sweeps with impressive accuracy , and vastly outperform other methods . We also tested S/HIC on a more challenging model with two population contractions followed by slow exponential growth , and more recent accelerated growth ( the European model from ref . [44] ) , obtaining qualitatively similar results . S/HIC therefore seems well suited for inference on populations with unknown demographic histories , though in such scenarios power could perhaps be improved by quickly fitting a relatively simple non-equilibrium demographic model prior to training . Even if oversimplified , simulations under such a model might better approximate patterns of variation around sweeps and within unselected regions than simulations under equilibrium , though we have not explored this possibility here . Though S/HIC performs far better than other tests for selection when tested on non-equilibrium populations , power for all methods is far lower than under constant population size , even if the demographic model is properly specified during training . Similar results are obtained under a severe population bottleneck . The reason for this is somewhat disconcerting: under these demographic models , the impact of selective sweeps on genetic diversity is blunted , making it far more difficult for any method to identify selection and discriminate between hard and soft sweeps . This underscores a problem that could prove especially difficult to overcome . That is , for some demographic histories all but the strongest selective sweeps may produce almost no impact on diversity for selection scans to exploit . A second and related confounding effect of misspecified demography is that following population contraction and recovery/expansion , much of the genome may depart from the neutral expectation , even if selective sweeps are rare . By examining the relative levels of various summaries of variation across a large region , rather than the actual values of these statistics , we are quite robust to this problem ( Fig 7 and S10 Fig ) . In other words , while non-equilibrium demography may reduce S/HIC’s sensitivity to selection and its ability to discriminate between hard and soft sweeps , we still classify relatively few neutral or even linked regions as selected . Thus , although inferring the mode of positive selection with high confidence may remain extremely difficult in some populations , our method appears to be particularly well suited for detecting selection in populations with non-equilibrium demographic histories whose parameters are uncertain . Indeed , applying our approach to chromosome 18 in a European human population , we detect most of the putative sweeps previously reported by Williamson et al . [57] . An additional advantage of machine learning approaches such as ours is the relative ease with which the classifier can be extended to incorporate more features , potentially adding information complementary to current features that could further improve classification power . For example , our examination of linkage disequilibrium is limited to within each subwindow; including features measuring the degree of LD between subwindows could also add valuable information . In addition , we could add statistics currently omitted which capture patterns of genealogical tree imbalance ( e . g . the maximum frequency of derived alleles [68] ) , or star-like sub-trees within genealogies ( e . g . iHS [42] , nSL [23] ) , both symptoms of various types of positive selection . Indeed , all tests for selective sweeps can be seen as methods to detect the distortions in the shapes of genealogies surrounding selected sites . Thus , if one could directly examine the ancestral recombination graph ( ARG ) surrounding a focal region , more powerful inference could be possible . It is now possible to estimate ARGs from sequence data [69] , and summaries of these estimated trees could be incorporated as features to identify sweeps and classify their mode . These are just some of a multitude of possible features that one can use to make inferences about natural selection . The success of S/HIC , evolBoosting [40] , and SFselect [37] in our tests relative to more conventional methods shows that machine learning approaches leveraging many different types of information have the potential to make far more powerful inferences than methods relying on an individual statistic . In summary , we have devised a machine learning-based scan for positive selection that possesses not only unparalleled accuracy , but is also exceptionally robust to the non-equilibrium demographic models we have examined here . This finding is extremely encouraging , though we can’t be certain it generalizes to every possible demographic scenario . Adjustments to the feature space can easily be made to better suit a particular study population . For example , if haplotypic phase is unknown , one can replace measures of gametic LD with zygotic LD . Additional classes could also be incorporated into the classifier ( e . g . “partial” or incomplete sweeps , balancing selection , or background selection ) , as long as they can be simulated to generate training data . Thus , our approach is practical and flexible . As additional population genetic summary statistics and tests for selection are devised , they can be incorporated into our feature space , thereby strengthening an already powerful method which has the potential to illuminate the impact of selection on genomic variation with unprecedented detail .
|
The genetic basis of recent adaptation can be uncovered from genomic patterns of variation , which are perturbed in predictable ways when a beneficial mutation “sweeps” through a population . However , the detection of such “selective sweeps” is complicated by demographic events , such as population expansion , which can produce similar skews in genetic diversity . Here , we present a method for detecting selective sweeps that is remarkably powerful and robust to potentially confounding demographic histories . This method , called S/HIC , operates using a machine learning paradigm to combine many different features of population genetic variation , and examine their values across a large genomic region in order to infer whether a selective sweep has recently occurred near its center . S/HIC is also able to accurately distinguish between selection acting on de novo beneficial mutations ( “hard sweeps” ) and selection on previously standing variants ( “soft sweeps” ) . We demonstrate S/HIC’s power on a variety of simulated datasets as well as human population data wherein we recover several previously discovered targets of recent adaptation as well as a novel selective sweep .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"cognitive",
"science",
"engineering",
"and",
"technology",
"demography",
"population",
"genetics",
"neuroscience",
"simulation",
"and",
"modeling",
"decision",
"analysis",
"management",
"engineering",
"artificial",
"intelligence",
"population",
"biology",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"decision",
"trees",
"people",
"and",
"places",
"population",
"metrics",
"population",
"size",
"natural",
"selection",
"evolutionary",
"processes",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"evolutionary",
"biology",
"genomics",
"statistics",
"computational",
"biology",
"machine",
"learning"
] |
2016
|
S/HIC: Robust Identification of Soft and Hard Sweeps Using Machine Learning
|
The complex life cycle of the parasitic nematode Strongyloides stercoralis leads to either developmental arrest of infectious third-stage larvae ( iL3 ) or growth to reproductive adults . In the free-living nematode Caenorhabditis elegans , analogous determination between dauer arrest and reproductive growth is governed by dafachronic acids ( DAs ) , a class of steroid hormones that are ligands for the nuclear hormone receptor DAF-12 . Biosynthesis of DAs requires the cytochrome P450 ( CYP ) DAF-9 . We tested the hypothesis that DAs also regulate S . stercoralis development via DAF-12 signaling at three points . First , we found that 1 μM Δ7-DA stimulated 100% of post-parasitic first-stage larvae ( L1s ) to develop to free-living adults instead of iL3 at 37°C , while 69 . 4±12 . 0% ( SD ) of post-parasitic L1s developed to iL3 in controls . Second , we found that 1 μM Δ7-DA prevented post-free-living iL3 arrest and stimulated 85 . 2±16 . 9% of larvae to develop to free-living rhabditiform third- and fourth-stages , compared to 0% in the control . This induction required 24–48 hours of Δ7-DA exposure . Third , we found that the CYP inhibitor ketoconazole prevented iL3 feeding in host-like conditions , with only 5 . 6±2 . 9% of iL3 feeding in 40 μM ketoconazole , compared to 98 . 8±0 . 4% in the positive control . This inhibition was partially rescued by Δ7-DA , with 71 . 2±16 . 4% of iL3 feeding in 400 nM Δ7-DA and 35 μM ketoconazole , providing the first evidence of endogenous DA production in S . stercoralis . We then characterized the 26 CYP-encoding genes in S . stercoralis and identified a homolog with sequence and developmental regulation similar to DAF-9 . Overall , these data demonstrate that DAF-12 signaling regulates S . stercoralis development , showing that in the post-parasitic generation , loss of DAF-12 signaling favors iL3 arrest , while increased DAF-12 signaling favors reproductive development; that in the post-free-living generation , absence of DAF-12 signaling is crucial for iL3 arrest; and that endogenous DA production regulates iL3 activation .
Strongyloides stercoralis is a parasitic nematode that infects both humans and dogs and is the causative agent of strongyloidiasis , which predominately afflicts socio-economically disadvantaged people in developing countries [1–3] . While chronic strongyloidiasis is often asymptomatic or accompanied by low-grade gastrointestinal symptoms , S . stercoralis infection in immunocompromised or corticosteroid-treated patients can progress to hyperinfection and disseminated strongyloidiasis , which can be fatal [4 , 5] . Understanding the mechanisms regulating the development of S . stercoralis may lead to improved diagnostic , control , and treatment strategies . Similar to many nematodes , including the free-living nematode Caenorhabditis elegans , S . stercoralis has crucial points in its life cycle , where the organism is either fated towards reproductive adulthood or developmental arrest ( Fig 1 ) . Female post-parasitic first-stage larvae excreted in the feces of an infected host can undertake two possible routes of development: a homogonic route leading directly to developmentally arrested infectious third-stage larvae ( iL3 ) or a heterogonic route leading to free-living adults [6] . This developmental switch is analogous to the switch between dauer arrest and reproductive development made by first-stage C . elegans larvae [7] , with dauer arrest favored at high temperatures , low food abundance , and high population density , which is signaled by rising titers of constitutively-produced ascaroside pheromones [7–9] . S . stercoralis also exercises strict developmental controls in the post-free-living generation , where larvae invariably mature into non-feeding iL3; however , upon entering a permissive host , third-stage larvae resume feeding and development , eventually maturing into parthenogenetic parasitic females in the intestinal lumen [3 , 10 , 11] . Similarly , C . elegans dauer larvae resume feeding and develop to reproductive adults when environmental conditions improve [7] . Parallels between S . stercoralis iL3 and C . elegans dauer larvae extend beyond these functional similarities and include shared morphological features , including a long , radially constricted filariform pharynx , a plugged buccal cavity , and a stress-resistant cuticle [12–14] . While pathways regulating dauer arrest , and to a lesser extent dauer exit , have been well-studied in C . elegans [15] , the developmental controls regulating S . stercoralis iL3 formation and activation have only recently been examined [16–20] . The "dauer hypothesis" predicts that similar mechanisms govern iL3 and dauer development [14 , 21]; however , given that C . elegans and S . stercoralis are members of two different nematode clades [22] , with parasitism thought to have arisen independently in each [23] , it is entirely plausible that different signaling mechanisms could regulate formation of S . stercoralis iL3 and C . elegans dauer larvae [14 , 24] . In C . elegans , one of the primary mechanisms regulating the determination between reproductive development and dauer arrest involves a class of endogenous steroid hormones known as dafachronic acids ( DAs ) [25] . Under conditions promoting reproductive growth and development , DAs are abundant and bind the nuclear hormone receptor Ce-DAF-12 , which controls a network of genes that carry out these functions . Conversely , conditions favoring dauer arrest lead to a paucity of DAs and to Ce-DAF-12 functioning as a co-repressor , thereby instituting a genetic program for developmental arrest [26] . This mutually exclusive developmental switch occurs in the first-stage larvae ( L1 ) and must be reinforced to prevent development of worms with both dauer and adult attributes , which would be detrimental to the organism . When an L1 encounters favorable conditions , environmental cues—transduced by upstream cyclic guanosine monophosphate ( cGMP ) signaling , and then by parallel insulin/insulin-like growth factor ( IIS ) and dauer transforming growth factor β ( TGFβ ) signaling pathways [8 , 15]—trigger DA production in neuroendocrine XXX cells in the head of the developing larva [27] . The initial small quantity of DAs produced by XXX cells promotes further DA production throughout the hypodermis via a Ce-DAF-12-mediated positive feedback loop of DA synthesis , thereby ensuring the worm commits to reproductive development [28–30] . In C . elegans , the key enzyme in endogenous DA biosynthesis is the cytochrome P450 Ce-DAF-9 [28 , 31 , 32] , which is reflected in a significant increase in Ce-daf-9 transcripts during reproductive development and dauer exit [29 , 33] . Similar to steroids in other animals , C . elegans DAs are derived from cholesterol [25] , which is first modified by the Rieske-like oxygenase Ce-DAF-36 [34] , and subsequently by the 3-hydroxysteroid dehydrogenase Ce-DHS-16 [35] , before the final redox reaction is carried out by the hydroxylase Ce-DAF-9 in partnership with the NADPH-cytochrome P450 reductase Ce-EMB-8 [35] . Careful biochemical work originally described DAs as the Ce-DAF-12 ligands , the most potent of which is Δ7-DA [25] . More recent work examining metabolites of DAs has described Δ1 , 7-DA and 3α-OH-Δ7-DA as additional ligands of Ce-DAF-12 [36] . When DAs are absent , the co-repressor Ce-DIN-1 blocks Ce-DAF-12 activity [25 , 37] . When DAs are present , they bind Ce-DAF-12 and reverse Ce-DIN-1 repression [25] , allowing Ce-DAF-12 to simultaneously increase the transcription of Ce-let-7 microRNA family members that block dauer-formation pathways [38–41] and initiate a reproductive developmental program that promotes the aerobic catabolism of fatty acids for growth [42] . In parasitic nematodes , the mechanisms controlling iL3 arrest and activation as well as homogonic and heterogonic development in Strongyloides spp . are less well-understood [43] . The developmental checkpoint regulating homogonic versus heterogonic development in female post-parasitic L1 is regulated by both strain genetics and temperature , with commitment occurring early in L1 development in S . stercoralis [44] and the closely related Strongyloides ratti [45] . In the S . stercoralis UPD strain used in this study , >95% of larvae develop via the heterogonic route; other isolates range from fully homogonic to fully heterogonic in their development [46 , 47] . Similarly , frequencies of homogonic and heterogonic development vary among geographical isolates and over the course of infection in S . ratti [48] . Temperature also regulates the developmental switch between homogonic and heterogonic pathways in Strongyloides spp . In S . stercoralis , temperatures below 34°C signal post-parasitic L1 to take the heterogonic pathway , while temperatures similar to that of the host , 34°C or above , result in development directly to iL3 [44] . Moreover , the two amphidial neurons ASF and ASI regulate the developmental switch between these two routes; when both neurons are inactivated , the vast majority of post-parasitic L1 develop homogonically even at temperatures below 34°C [49] . This is similar to the regulation of the developmental switch in C . elegans L1 by the analogous ADF and ASI amphidial neurons [50] . However , no specific cellular signal transduction pathway has been implicated in regulating this developmental switch in Strongyloides spp . In S . stercoralis and closely related parasitic nematodes , including S . ratti and Strongyloides papillosus , progeny of the single generation of free-living male and female adults invariably form developmentally arrested iL3 , which are all genetically female—thus leading to a strictly female parasitic generation [6] . However , this post-free-living developmental fate is not shared by all members of the Strongyloides genus , as Strongyloides planiceps can produce a limited number of free-living generations of males and females [51] , and the evolutionarily more distant Parastrongyloides trichosuri can produce apparently unlimited generations of free-living males and females [52] . While the formation of P . trichosuri iL3 is mediated by a constitutively secreted pheromone [53] , similar to C . elegans dauer pheromone [54 , 55] , this does not appear to be the case for S . stercoralis because iL3 form in the post-free-living generation regardless of population density . However , S . stercoralis iL3 arrest does require reduced IIS [17] . Interestingly , application of exogenous Δ7-DA to the post-free-living generation of S . stercoralis and S . papillosus prevents iL3 arrest , resulting in rhabditiform L3 and L4 in S . stercoralis [56] and a second generation of fecund free-living females in S . papillosus [57] , thus suggesting that iL3 arrest may be the result , in part , of diminished DAF-12 signaling . However , the duration of exposure to DA necessary to induce these phenotypes has been unknown . Of the developmental checkpoints , factors regulating iL3 activation in S . stercoralis and other parasitic nematodes are perhaps the best studied . S . stercoralis iL3 exhibit positive chemotaxis and thermotaxis towards a variety of molecules indicative of a host [58] , including carbon dioxide [59] , sodium chloride [60] , urocanic acid [61] , and host body temperature [62] , with many of these responses mediated by amphidial neurons . Upon entering a permissive host , iL3 quickly resume feeding and development , a process that is mediated in part by ASJ amphidial neurons [63] . Resumption of feeding is accompanied by modulation of insulin-like peptide transcripts [19 , 20] , while inhibition of IIS prevents iL3 feeding [18] . Furthermore , increases in cGMP signaling and DA signaling , by exogenous application of these compounds , trigger iL3 feeding [20 , 56] . However , to our knowledge , it remains unknown whether S . stercoralis , or any other parasitic nematodes , produces endogenous ligands for DAF-12 . In this study using S . stercoralis , we demonstrate that DA modulates the post-parasitic switch regulating the decision between reproductive development and iL3 arrest , with increased DAF-12 signaling favoring reproductive development . Furthermore , we demonstrate that in the post-free-living generation , exposure to DA also effects a shift from iL3 arrest , favoring formation of reproductively developing larvae . In the majority of worms , this commitment to reproductive development occurs within 24–48 hours of DA exposure . We also provide the first evidence for endogenous biosynthesis of Ss-DAF-12 ligand , as a blockade of iL3 activation by a chemical inhibitor of cytochrome P450s is partially overridden by administration of DA .
S . stercoralis was maintained in purpose-bred , prednisone-treated mix breed dogs and in purpose-bred Mongolian gerbils according to protocols 802593 and 804883 approved by the University of Pennsylvania Institutional Animal Care and Use Committee ( IACUC ) . All IACUC protocols , as well as routine husbandry care of the animals , were conducted in strict accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . Post-parasitic L1 of S . stercoralis that were unexposed to environmental cues affecting development were harvested from the intestines of experimentally infected gerbils at necropsy as follows . Tum/Mon strain Meriones unguiculatus ( Mongolian gerbils ) were experimentally infected with 3 , 000 iL3 and euthanized 21 days later by CO2 asphyxiation in accordance with standards established by the American Veterinary Medical Association . The intestines from individual animals were placed in Dulbecco's Modified Eagle's Medium ( DMEM ) , supplemented with 1 mg/ml gentamicin sulfate , with either 1 μM ( 25S ) -Δ7-DA ( CAS 949004-12-0 ) or ethanol carrier ( 0 . 1% ethanol ) at 37°C . Post-parasitic L1 were picked from the intestinal debris using a stereomicroscope with a 37°C heated stage . Post-parasitic L1 harvested in DMEM with Δ7-DA were transferred to a 35 x 10 mm nematode growth medium ( NGM ) plate containing 3 ml of agar and spotted with 300 μl of E . coli OP50 containing 10 μM Δ7-DA , resulting in a final concentration of 1 μM Δ7-DA ( 0 . 1% ethanol ) , and incubated at 37°C . Post-parasitic L1 harvested in DMEM with ethanol carrier were transferred to NGM plates spotted with E . coli OP50 containing ethanol carrier and incubated at either 37°C or 22°C . The developmental stage of the developing larvae was recorded for all three conditions at both 24 and 48 hours post-plating . Five biological replicates were performed , and the mean percentages of larvae developing to either filariform iL3 or rhabditiform L3-Adult forms , with the standard deviation , were calculated . Proportions of worms in each developmental class as functions of temperature and presence of Δ7-DA were analyzed by 2-way ANOVA with post hoc comparisons of selected frequencies by the Bonferroni test . Petri dishes ( 35 x 10 mm ) containing 3 ml of NGM agar were seeded with 300 μl of a suspension of E . coli OP50 in LB broth containing Δ7-DA at concentrations ranging from 331 nM to 10 μM . Assuming uniform dispersal of the compound in the agar , this resulted in NGM/OP50 plates containing 33 . 1 nM to 1 μM Δ7-DA . Control plates containing 0 . 1% ethanol , the maximal concentration of Δ7-DA carrier to which the worms were exposed , were made by spotting plates with OP50 suspensions containing 1% ethanol . Semi-synchronous populations of S . stercoralis eggs were prepared on experimental and control plates by transferring 20–30 gravid free-living S . stercoralis females to each plate and allowing them to oviposit for three hours at room temperature . At the end of this interval , egg-laying worms were removed from the plates , which then contained cohorts of 50–200 eggs . Plates with eggs were then sealed with Parafilm and incubated for 72 hours at 22°C . Following this incubation , developing worms were classed as rhabditiform L1-L2 , filariform iL3 , or rhabditiform L3-L4 . Mean percentages of worms in each developmental class from three experimental replicates , with standard deviations , were plotted as a function of Δ7-DA concentration . The EC50 for induction of rhabditiform L3-L4 forms was calculated by non-linear regression of frequency in the L3-L4 developmental class on log-transformed Δ7-DA concentrations . We established triplicate semi-synchronous cultures of post-free-living S . stercoralis larvae in the same range of Δ7-DA concentrations as described above for the dose-response assessment . At intervals of 24 and 48 hours in culture at 22°C , worms from one of the triplicate cultures at each Δ7-DA concentration were washed off the plate with M9 buffer and then subjected to two additional washes in 10 volumes of M9 buffer; subsequently , the worms were re-plated on non-DA-treated plates and cultured for the balance of the standard 72-hour culture period at 22°C . Worms cultured for 72 hours in the presence of Δ7-DA constituted the continuously exposed controls . At the end of the 72-hour incubation , developing S . stercoralis larvae were classed as rhabditiform L1-L2 , filariform iL3 , or rhabditiform L3-L4 as before . Mean percentages , and standard deviations , of worms in the L3-L4 class in each concentration of Δ7-DA were calculated for three experimental replicates . Effects of Δ7-DA exposure duration and concentration were analyzed by 2-way ANOVA with post hoc pairwise comparisons of frequencies by the Bonferroni test . In vitro activation of S . stercoralis iL3 was performed as previously described [18 , 20 , 63] with the following adaptations . All conditions were supplemented with antibiotics ( final concentrations: 100 U/ml penicillin , 10 μg/ml streptomycin , and 12 . 5 μg/ml tetracycline ) . iL3 were isolated from seven-day-old charcoal coprocultures ( incubated at 25°C ) by the Baermann technique at 27–29°C . iL3 were subsequently washed twice in deionized water and incubated in M9 buffer [66] , supplemented with antibiotics , for three hours at room temperature before distribution amongst the different conditions . Experiments examining inhibition of iL3 activation with ketoconazole ( CAS 65277-42-1; Sigma ) were carried out in DMEM supplemented with L-glutamine , 4 . 5 g/L glucose , and sodium pyruvate ( Corning ) . This medium with the indicated supplements supports resumption of feeding by a majority of S . stercoralis iL3 at 37°C without additional host-like factors [20] . A 10 mM stock solution of ketoconazole in dimethyl sulfoxide ( DMSO ) was used for the experimental conditions , which included varying concentrations of ketoconazole ( 10 μM , 20 μM , 30 μM , 40 μM , and 60 μM ) in DMEM , each with 0 . 8% DMSO . The negative control was M9 buffer , and the positive control was DMEM , each with 0 . 8% DMSO . Experiments to ascertain rescue of ketoconazole-mediated inhibition of iL3 activation by Δ7-DA were carried out in DMEM . A 1 mM stock solution of Δ7-DA in ethanol was used for the experimental conditions , which included varying concentrations of Δ7-DA ( 50 nM , 200 nM , 400 nM , and 800 nM ) in DMEM with 35 μM ketoconazole; each condition contained 0 . 35% DMSO and 0 . 2% ethanol . The positive controls were DMEM and DMEM supplemented with 800 nM Δ7-DA , while the negative control was M9 buffer; each condition contained 0 . 35% DMSO and 0 . 2% ethanol . Each condition consisted of three wells in a 96-well plate , with approximately 100 iL3 in 100 μl total volume in each well . iL3 were incubated at 37°C in 5% CO2 in air for 21 hours; subsequently , 2 μl of fluorescein isothiocyanate ( FITC ) ( CAS 3326-32-7; Sigma ) dissolved in N , N-dimethylformamide ( DMF ) ( CAS 68-12-2; Sigma ) at 20 mg/ml and incubated for ≥ one month was added to each well , and the cultures were incubated an additional three hours at 37°C and 5% CO2 in air ( 24 hours total ) . Pre-incubation of FITC solutions was empirically determined to decrease binding of FITC to iL3 cuticles relative to that seen with fresh dilutions of the dye in DMF . iL3 for each condition were pooled and washed five times in 14 ml of M9 buffer , with centrifugation at 75 x g for five minutes at 20°C . iL3 were then mounted on glass slides with grease-edged cover-slips and viewed by fluorescence microscopy using an SZX12 stereomicroscope ( Olympus ) equipped with an X-Cite 120LED illuminator ( Lumen Dynamics ) . Only live iL3 ( indicated by movement ) with FITC in the pharynx were scored as "positive" for feeding . Apart from loss of the buccal plug and resumption of feeding , no further development of the iL3 occurs in this system . Dead worms , indicated by lack of movement and/or whole-body FITC staining , were excluded from the feeding analysis , and the percentage of dead worms was determined from the total number of worms present . At least 200 iL3 were counted for each condition . Four biological replicates were performed , and the mean percentage of iL3 feeding in each condition , with the standard deviation , was plotted . Coefficients of correlations between ketoconazole concentration and iL3 feeding and between Δ7-DA concentration and iL3 feeding in the presence of 35 μM ketoconazole were computed by the nonparametric Spearman method . IC50 for ketoconazole inhibition of iL3 feeding and EC50 for Δ7-DA rescue of iL3 feeding in the presence of 35 μM ketoconazole were calculated by non-linear regression of iL3 feeding frequency on concentrations of ketoconazole and Δ7-DA , respectively . To identify S . stercoralis homologs of cytochrome P450-encoding genes , we performed reciprocal BLAST searches of the S . stercoralis genome v . 2 . 0 . 4 ( available: ftp://ftp . sanger . ac . uk/pub/project/pathogens/HGI/ ) with C . elegans cytochrome P450 protein sequences ( WormBase release WS245 ) using Geneious v . 6 . 1 . 8 ( Biomatters Ltd . ) ; BLAST hits were manually annotated using RNAseq data viewed with the Integrative Genomics Viewer v . 2 . 0 . 34 ( available: www . broadinstitute . org/igv/ ) and Geneious [19 , 20] . Using mapped and de novo assembled RNAseq data , coding sequences were manually corrected to derive full-length coding sequences ( S1 Data ) and putative protein sequences ( S2 Data ) . Putative cytochrome P450-encoding genes were named , or renamed from those previously described [19] , using the standard convention [67] . These genes were identified and related using a combination of reverse-BLAST searches , a protein-identity matrix , and a ClustalW-generated protein alignment ( S3 Data ) and neighbor-joining phylogenetic tree , with 1000 iterations of boot-strapping , utilizing metazoan cytochrome P450 protein sequences , including C . elegans , Caenorhabditis briggsae , and Homo sapiens . All of these analyses were performed with Geneious . Transcript abundances for each of the S . stercoralis cytochrome P450-encoding genes were determined as previously described [19 , 20] , with the following adaptations . RNAseq raw reads , derived from polyadenylated RNA libraries , for iL3 ( ArrayExpress accession number E-MTAB-2192; http://www . ebi . ac . uk/arrayexpress/experiments/E-MTAB-2192/ ) [20] as well as in vivo activated L3 ( L3+ ) , parasitic females , post-parasitic L1 , post-parasitic L3 enriched for females , free-living females , and post-free-living L1 ( ArrayExpress accession number E-MTAB-1164; http://www . ebi . ac . uk/arrayexpress/experiments/E-MTAB-1164 ) [19] , were mapped to the S . stercoralis v . 2 . 0 . 4 genome contigs using Tophat2 v . 2 . 0 . 13 [68] , which utilized Bowtie2 v2 . 2 . 4 [69] and Samtools v0 . 1 . 18 [70] , and previously established parameters [20] . Normalized transcript abundances for cytochrome P450-encoding genes ( S4 Data ) , calculated as fragments per kilobase of coding exon per million fragments mapped ( FPKM ) with paired-end reads counting as single sampling events , and 95% confidence intervals were determined using Cuffdiff v . 2 . 0 . 2 [71] . For all experiments , statistical analyses were carried out and plots were created with Prism version 5 . 03 ( GraphPad Software , Inc . ) . Statistical probabilities P < 0 . 05 were considered significant .
Based on its role in promoting continuous reproductive development in C . elegans [26] , we hypothesized that Δ7-DA or a similar Ss-DAF-12 ligand promotes heterogonic development by post-parasitic larvae of S . stercoralis ( Fig 1 ) . Conversely , we hypothesized that down-regulation of Ss-DAF-12 ligand ( s ) should drive direct development of post-parasitic L1 to iL3 . We tested these hypotheses by administering Δ7-DA to cultures of post-parasitic L1 developing at 37°C , where direct development predominates . As expected , control larvae reared at 22°C developed almost exclusively to free-living adults . However , a mean of 69 . 4 ± 12 . 0% ( standard deviation ) of control larvae reared at 37°C developed directly to iL3 , with the remaining minority developing to free-living adults . In contrast to controls reared at 37°C , larvae reared at this temperature in the presence of 1 μM Δ7-DA developed exclusively to free-living adults ( Fig 2 ) . These results demonstrate that Ss-DAF-12 signaling regulates switching between homogonic and heterogonic developmental alternatives in post-parasitic larvae of S . stercoralis . Moreover , the ability of Δ7-DA to override the high temperature signal to post-parasitic L1 indicates that Ss-DAF-12 signaling operates downstream of mechanisms transducing temperature cues to the larvae from the environment . Under normal circumstances , the progeny of free-living male and female S . stercoralis develop exclusively to iL3 [3] . Just as absence of DA synthesis promotes dauer arrest in C . elegans [25] , we hypothesized that the invariant pattern of iL3 formation that takes place in the post-free-living generation of S . stercoralis results from absence of Ss-DAF-12 ligands ( Fig 1 ) . In support , we previously demonstrated that exogenous Δ7-DA can induce a portion of post-free-living S . stercoralis larvae to bypass iL3 developmental arrest , developing instead to rhabditiform L3-L4 [56] . Moreover , similar concentrations of Δ7-DA are known to induce post-free-living larvae of S . papillosus to develop to reproductively competent second-generation free-living females [57] . Thus , we sought to achieve a more precise dose-response profile of Δ7-DA induction of development to advanced rhabditiform larvae or adults by post-free living S . stercoralis larvae; additionally , we sought to ascertain whether this induction occurs in a discrete time interval , or conversely , whether Δ7-DA is required throughout the period of post free-living larval development in order to effect this switch in developmental fate . We first cultured semi-synchronous populations of post-free-living S . stercoralis larvae for 72 hours at 22°C on NGM agar plates with lawns of E . coli OP50 and fecal bacteria and with Δ7-DA at concentrations ranging from 0 to 1 μM and then assessed the degree of larval development . In the range of 125 nM to 1 μM , Δ7-DA brought about a dose-dependent increase in the frequency of larvae developing to rhabditiform L3 and L4 ( Fig 3A ) . The EC50 for this response was 318 nM , which is comparable to the EC50 of 147 nM for activation of the Ss-DAF-12 ligand-binding domain by Δ7-DA in a cell-based reporter assay [56] . As proportions of rhabditiform L3 and L4 increased in response to increasing Δ7-DA concentration , proportions of larvae developing to the iL3 declined ( Fig 3B ) . Proportions of larvae remaining as L1 and L2 remained roughly constant at all concentrations of Δ7-DA ( S1 Fig ) . To ascertain discrete developmental triggering by Δ7-DA , we initiated semi-synchronous plate cultures of post-free-living S . stercoralis larvae in the presence of increasing concentrations of Δ7-DA . We then washed cohorts of larvae out of the compound at 24 and 48 hours of culture , re-plated them on non-DA-treated plates , and continued culture at 22°C for the balance of the 72-hour culture period . At 72 hours of culture , we compared the frequency of development to rhabditiform L3-L4 in these transiently exposed worms to that of larvae developing for 72 hours at 22°C under continuous exposure to Δ7-DA . None of the larvae exposed to Δ7-DA for the first 24 hours of culture developed to rhabditiform L3-L4 ( Fig 4 ) . However , a significant proportion of larvae exposed to Δ7-DA for the first 48 hours developed to rhabditiform L3 and L4 . Proportions of larvae undergoing the developmental switch increased with increasing concentration of Δ7-DA . Thus , it appears that between 24 and 48 hours of development at 22°C , a significant proportion of larvae exposed to exogenous Δ7-DA commit to bypassing the iL3 and developing instead to rhabditiform L3-L4 . Despite these kinetic refinements , post-free-living S . stercoralis larvae failed to develop to sexually mature free-living adults , as S . papillosus larvae do when treated with similar levels of Δ7-DA [57] . If we consider the iL3 of S . stercoralis to be the developmental equivalent of C . elegans dauer larvae , then resumption of development by these infectious larvae during the infective process would be the equivalent of dauer recovery [14] . Based on the role of Δ7-DA in promoting continuous development in C . elegans [25 , 72] , we hypothesized that biosynthesis of this or other Ss-DAF-12 ligand ( s ) is necessary for resumption of development by S . stercoralis iL3 at the time of infection ( Fig 1 ) . In C . elegans , the cytochrome P450 encoded by Ce-daf-9 is required for biosynthesis of Δ7-DA from its precursor lathosterone [25 , 35] . Extending this analogy to S . stercoralis , we hypothesized that cytochrome P450 activity is required for endogenous biosynthesis of DA ( s ) or related Ss-DAF-12 ligand ( s ) and that inhibition of this activity would block resumption of development by iL3 at the time of infection . To test this , we asked whether the cytochrome P450 inhibitor ketoconazole could inhibit resumption of development by S . stercoralis iL3 under host-like in vitro culture conditions . Ketoconazole suppressed resumption of feeding by S . stercoralis iL3 cultured under permissive conditions . The positive control ( DMEM ) cultures , which reflect host-like biochemical conditions , supported resumption of feeding by a mean of 98 . 8 ± 0 . 4% ( mean and standard deviation ) of the iL3 ( Fig 5A ) . The inhibitory effect of ketoconazole administered to iL3 in DMEM cultures was dose-dependent and maximal at 40 μM and higher , with a mean of 5 . 6 ± 2 . 9% of iL3 feeding at 40 μM . In the M9 buffer negative control , 0 . 1 ± 0 . 3% of iL3 were feeding . Ketoconazole broadly inhibits cytochrome P450s [73] , and so its suppression of feeding by developing S . stercoralis iL3 could result from inhibition of multiple such enzymes , including the ortholog of C . elegans DAF-9 . To address this uncertainty , we asked what proportion of the observed suppression of feeding by ketoconazole could be attributable to depletion of Δ7-DA or a similar developmental regulatory steroid from the worms . To this end , we cultured iL3 in permissive medium ( DMEM ) containing either 35 μM ketoconazole alone or in 35 μM ketoconazole supplemented with increasing concentrations of Δ7-DA . In DMEM with 35 μM ketoconazole , Δ7-DA restored feeding responses among cultured iL3 in a dose-dependent fashion to a maximum of 71 . 2 ± 16 . 4% feeding in 400 nM Δ7-DA , compared to 17 . 1 ± 3 . 8% feeding in DMEM cultures with 35 μM ketoconazole alone ( Fig 5B ) . This represents a 4-fold increase in feeding over worms cultured in 35 μM ketoconazole alone and accounts for approximately two-thirds of the feeding response seen in non-ketoconazole-treated controls . Mortality , as reflected by the proportion of non-motile ( scored as dead ) worms , increased as a function of Δ7-DA concentration in the presence of 35 μM ketoconazole ( Fig 5B ) . At concentrations of 200 nM Δ7-DA and higher , the frequency of this mortality was higher than that observed at any concentration of Δ7-DA alone or at comparable levels of ketoconazole ( Fig 5A ) , suggesting a synergistic toxic interaction between ketoconazole and Δ7-DA in this range of concentrations . Our findings that the CYP inhibitor ketoconazole suppresses resumption of feeding by iL3 under host-like culture conditions and that this effect can be partially rescued by Δ7-DA suggest that biosynthesis of a steroidal ligand of Ss-DAF-12 is necessary for this developmental step . The key enzyme for DA biosynthesis in C . elegans is the cytochrome P450 DAF-9 , which is homologous to the human CYP27A1 that produces bile acids [25] . In order to identify potential homologs of cytochrome P450-encoding genes in S . stercoralis , including a potential DAF-9 ortholog , we performed reciprocal BLAST searches , followed by manual annotation and correction of the hits to derive putative cytochrome P450 protein sequences ( S2 Data ) . This resulted in the determination of 26 cytochrome P450-encoding genes in S . stercoralis , which were grouped by family and subfamily by standard cytochrome P450 nomenclature [67] . Of the 26 cytochrome P450s , there were 2 subfamilies with several members which , due to their spatial proximity in the genome and sequence similarity , likely arose as a result of tandem gene duplications . These are the CYP3A subfamily , which contains seven genes , and the CYP29A family , which consists of nine . Since C . elegans daf-9 transcripts are up-regulated during reproductive growth and development as well as during dauer exit [28 , 31 , 33] , we examined the transcript abundance regulation for each of the cytochrome P450 homologs during the S . stercoralis life cycle ( S5 Data and S2 Fig ) . Congruent with the hypothesis that iL3 activation is mediated by a cytochrome P450 , we identified several genes that had an increase in transcript abundance from iL3 to in vivo activated L3 –members of the 3A subfamily: Ss-cyp3a23 and Ss-cyp3a27; members of the 29A subfamily: Ss-cyp29a6 , Ss-cyp29a13 , Ss-cyp29a25 , and Ss-cyp29a26; Ss-cyp33e2; and Ss-cyp22a9 . We identified Ss-cyp22a9 , previously identified as Ss-cyp-9 [19] , as the homolog with the closest similarity to Ce-daf-9 , based on both its phylogenetic relation to Ce-DAF-9 ( S3 Fig ) and its increased transcript abundance in developing larvae and during iL3 activation ( S2 Fig ) . Confirmation of Ss-CYP22A9 as the ortholog of Ce-DAF-9 awaits future functional studies .
In this study , we hypothesized that DA signaling in S . stercoralis , through the nuclear hormone receptor DAF-12 , would stimulate reproductive growth and development , while decreased DAF-12 activity , resulting from a reduction in DA , would cause developmental arrest . We interrogated several developmental checkpoints in S . stercoralis to determine whether DA signaling regulates the parasite's developmental program . Our data suggest that DA regulation of DAF-12 signaling plays an important role in the development of iL3 in this pathogen . While hookworm , filarial worm , and ascarid larvae all constitutively develop to iL3 , similar to dauer constitutive ( daf-c ) mutants in C . elegans [13] , post-parasitic larvae from Strongyloides spp . can form a non-obligatory free-living generation of male and female adult worms . However , the molecular mechanisms regulating the developmental switch in the post-parasitic L1 that controls homogonic and heterogonic development have remained elusive . The strain of S . stercoralis used in this study , the UPD strain , almost exclusively develops via the heterogonic route , whereby post-parasitic female larvae develop to a single free-living generation of adult worms . Other S . stercoralis isolates have post-parasitic females that develop predominantly via the homogonic route directly to iL3 , which are always female , or via a mix of heterogonic and homogonic development [46] . In this study , we demonstrated that developmentally uncommitted female post-parasitic L1 maturing at an elevated temperature , where iL3 arrest normally predominates , may instead be stimulated with exogenous Δ7-DA to develop to free-living adults via the heterogonic pathway ( Fig 2 ) . These data not only establish a role for Ss-DAF-12 in regulating the post-parasitic L1 checkpoint , but also demonstrate that signaling through this nuclear receptor lies downstream of the pathway sensing thermal cues . This is consistent with findings in C . elegans , where AFD thermosensitive neurons transduce information through cGMP signaling [74] , which lies upstream of DAF-12 signaling [26] . Future studies comparing the effect of DA on post-parasitic L1 development in other S . stercoralis strains that are genetically predisposed to homogonic development may shed additional light on the requirement for DAF-12 signaling in heterogonic development and the influence of additional signaling pathways on this checkpoint . S . stercoralis post-free-living larvae invariably undergo developmental arrest as iL3 in physiological conditions; however , another Strongyloides spp . , S . planiceps , can form several successive generations of free-living adults [43 , 51] . We hypothesized that addition of DA would stimulate S . stercoralis post-free-living L1 to complete a second free-living generation , similar to that observed with S . papillosus [57] , which also has a single free-living generation in physiological conditions . We found that rearing S . stercoralis post-free-living larvae in the presence of Δ7-DA suppressed iL3 formation and favored development of rhabditiform L3-L4 ( Fig 3 ) , which were morphologically similar to worms undergoing free-living development . While increasing concentrations of Δ7-DA increased the proportion of larvae developing to rhabditiform L3-L4 , no free-living adult females were observed . We hypothesize that this difference from the result observed in S . papillosus may be due to a requirement in S . stercoralis for additional stimulatory factors or the possibility that Δ7-DA is not the endogenous ligand for S . stercoralis DAF-12 . We also sought to determine the length of exposure to Δ7-DA required to stimulate S . stercoralis post-free-living larvae to develop into rhabditiform L3 and L4 . Based on work with C . elegans , where a small quantity of endogenously-produced ligand results in a DAF-12-mediated amplification loop [28] , we expected that only a brief pulse of Δ7-DA would be required to prevent iL3 arrest . However , we found that post-free-living larvae required 24–48 hours of exposure to Δ7-DA ( Fig 4 ) . This suggests that repressive mechanisms during iL3 arrest actively inhibit DAF-12 function , a phenomenon that could result from: production of DAF-12 antagonists; a decrease in DA precursors by the action of an S . stercoralis strm-1 homolog [75] , which is supported by an increase in Ss-strm-1 transcripts in iL3 [19]; metabolism of DAs into inactive compounds , potentially by other cytochrome P450s that are up-regulated in iL3 ( S2 Fig ) ; or the possibility that Δ7-DA is only a weak agonist for Ss-DAF-12 . We assume that such repressive mechanisms or pharmacokinetic differences are not found in S . papillosus , where comparable levels of exogenously applied Δ7-DA elicited formation of reproductively competent second-generation free-living females [57] . Δ7-DA causes S . stercoralis iL3 to resume feeding , which is a hallmark of activation , and initiates a developmental program similar to activation in a permissive host [20 , 56] . Consequently , we hypothesized that S . stercoralis endogenously produces DAs or similar Ss-DAF-12 ligands during iL3 activation . In C . elegans , the final biosynthetic step in the production of Δ7-DA is performed by the cytochrome P450 Ce-DAF-9 , and we hypothesized that a similar enzyme in S . stercoralis produces endogenous Ss-DAF-12 ligands . In order to test this hypothesis , we used ketoconazole , which is a cytochrome P450 CYP3A family inhibitor at nanomolar concentrations and a broad-spectrum P450 inhibitor at micromolar concentrations [73] , to broadly inhibit cytochrome P450 function in S . stercoralis iL3 . We found that ketoconazole does indeed inhibit feeding in S . stercoralis iL3 in a dose-dependent fashion in the micromolar range ( Fig 5A ) , consistent with the synthesis of an endogenous steroid hormone during activation . Inhibition of iL3 feeding by ketoconazole in hookworm species suggests that DA signaling may also be important in iL3 activation in Clade V parasitic nematodes [56 , 76] , which are more closely related to C . elegans [23] . However , in S . stercoralis , this ketoconazole-mediated inhibition may be due to a block in the function of a DAF-9 homolog or of one/several of the other cytochrome P450s in the worm . Thus , we sought to determine the extent to which this phenotype could be attributable to the loss of DA by attempting to rescue the ketoconazole-inhibited iL3 by adding back Δ7-DA . We found that approximately two-thirds of the iL3 feeding could be restored by Δ7-DA ( Fig 5B ) , providing evidence that ketoconzaole-mediated iL3 inhibition is due to suppressed production of DA or related Ss-DAF-12 ligand ( s ) . Together , these data strongly suggest that S . stercoralis synthesizes Ss-DAF-12 ligands that promote the developmental activation of iL3 . When performing the Δ7-DA rescue experiments , we also noted an increase in iL3 mortality that corresponded with increasing concentrations of Δ7-DA in the presence of ketoconazole ( Fig 5B ) . Since ketoconazole and carrier solution concentrations remained constant , we could only attribute this mortality to an interaction between Δ7-DA and ketoconazole . This synergism suggests ketoconazole may also be blocking the metabolism of exogenous Δ7-DA by inhibition of other cytochrome P450s , allowing toxic levels of the compound or partially metabolized intermediates to accumulate in the worms . This phenomenon might be comparable to the synergism of pyrethrin insecticides when paired with the cytochrome P450 inhibitor piperonyl butoxide [77] . This synergistic effect might be exploited in the potential development of DAF-12 ligands as anthelmintics . The data we report here support the hypothesis that endogenous production of DA in S . stercoralis promotes free-living development and iL3 activation , while repression of DAF-12 signaling promotes and maintains iL3 arrest . However , formal proof of this hypothesis awaits discovery and characterization of the natural ligands of Ss-DAF-12 . Since S . stercoralis and C . elegans are phylogenetically distant , as members of separate clades where parasitism is thought to have evolved independently [23] , the biosynthesis of steroid hormones may be different in these species . Biochemical- and in vitro-based studies , similar to those used to identify Δ7-DA in C . elegans [25 , 56] , are called for to identify endogenous DAs in S . stercoralis . Furthermore , additional studies to elucidate the biosynthesis of natural Ss-DAF-12 ligands are a significant priority , as this biosynthetic pathway may constitute a chemotherapeutic target in S . stercoralis and other parasitic nematodes . Our phylogenetic study of the cytochrome P450s in S . stercoralis ( S3 Fig ) and of regulation of their transcripts throughout the life cycle ( S2 Fig ) provide a logical starting point for such biosynthetic experiments . Based upon these findings , we hypothesize that Ss-cyp22a9 is the functional homolog of Ce-daf-9 in the parasite . Provided constructs encoding this gene can be optimized for expression in mammalian cells , this hypothesis should be testable using modifications of proven cell-based assay methods [56] . In conclusion , our demonstration that a developmental blockade in S . stercoralis iL3 by ketoconazole can be rescued by Δ7-DA ( Fig 2 ) constitutes the first functional evidence of DAF-12 signaling stimulated by endogenous synthesis of its ligand in a parasitic nematode . Furthermore , our findings that induction of rhabditiform L3 and L4 in the post-free-living generation by Δ7-DA ( Figs 3 , 4 and 5 ) and that homogonic to heterogonic development in post-parasitic L1 is shifted by administration of Δ7-DA ( Fig 5 ) support the hypotheses we frame in this paper about ligation states of Ss-DAF-12 and the overall function of Ss-DAF-12 signaling during the S . stercoralis life cycle ( Fig 1 ) . The fact that this crucial developmental regulatory signaling pathway can be manipulated by exogenous administration of a steroid ligand , in this case a heterologous one from C . elegans [25] , raises the possibility that Ss-DAF-12 signaling may represent a new chemotherapeutic target in S . stercoralis . This potential for a new class of anthelmintics likely extends to a diverse array of parasitic nematodes , as DAF-12 is conserved in all members of the Strongyloididae investigated to date [57 , 78] and in hookworms [56 , 79] . Regarding the latter group of parasitic nematodes , which are phylogenetically diverged from Strongyloides spp , the biological activity of the DAs has also been confirmed in Ancylostoma caninum [56] . Given the wide range of existing drugs targeting nuclear receptors [80] , we propose that the potential of DAF-12 signaling in parasitic nematodes be actively investigated as a novel chemotherapeutic target .
|
Strongyloides stercoralis is a parasitic nematode that infects hundreds of millions of people worldwide . The infectious form of S . stercoralis is a developmentally arrested third-stage larva ( iL3 ) ; once inside the host , the iL3 activates and develops into an adult parasitic female . First-stage larvae ( L1 ) excreted in the host feces have two routes of development: either directly to iL3 or indirectly to free-living adults . The molecular mechanisms controlling iL3 developmental arrest and activation , and the switch regulating post-parasitic L1 development , are poorly understood . The free-living nematode Caenorhabditis elegans has a developmentally arrested stage , morphologically similar to iL3 , called dauer . Dauer formation is prevented by endogenous production of a class of steroid hormones called dafachronic acids ( DAs ) , which are synthesized by a cytochrome P450 . We demonstrated that in S . stercoralis , administering DA can both stimulate post-parasitic L1 to develop to free-living adults instead of iL3 as well as prevent iL3 developmental arrest . Additionally , blocking cytochrome P450 function prevents iL3 activation in a host-like environment , suggesting endogenous DA production . We also characterized the developmental expression of cytochrome P450s present in the S . stercoralis genome . Together , our data demonstrate that DA regulates S . stercoralis iL3 arrest and activation and the post-parasitic developmental switch .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2016
|
Regulation of Life Cycle Checkpoints and Developmental Activation of Infective Larvae in Strongyloides stercoralis by Dafachronic Acid
|
Memories are stored and consolidated as a result of a dialogue between the hippocampus and cortex during sleep . Neurons active during behavior reactivate in both structures during sleep , in conjunction with characteristic brain oscillations that may form the neural substrate of memory consolidation . In the hippocampus , replay occurs within sharp wave-ripples: short bouts of high-frequency activity in area CA1 caused by excitatory activation from area CA3 . In this work , we develop a computational model of ripple generation , motivated by in vivo rat data showing that ripples have a broad frequency distribution , exponential inter-arrival times and yet highly non-variable durations . Our study predicts that ripples are not persistent oscillations but result from a transient network behavior , induced by input from CA3 , in which the high frequency synchronous firing of perisomatic interneurons does not depend on the time scale of synaptic inhibition . We found that noise-induced loss of synchrony among CA1 interneurons dynamically constrains individual ripple duration . Our study proposes a novel mechanism of hippocampal ripple generation consistent with a broad range of experimental data , and highlights the role of noise in regulating the duration of input-driven oscillatory spiking in an inhibitory network .
Sleep , which consumes about a third of our lives , is thought to play a critical role in memory consolidation . Specifically , sleep influences unconscious post-encoding processes that result in long term memory consolidation and reconsolidation . Behavioral studies show that performance in various memory tasks improves after sleep compared to a similar period of wake [1 , 2] , and such improvement was observed in declarative , procedural and emotional memory tasks [3–7] . Cortical and hippocampal circuits show characteristic oscillatory activities at different sleep stages [1] . During slow-wave sleep ( SWS ) , cortex is synchronized by low-frequency slow oscillations ( 0 . 2–1 Hz ) between Down states–in which most cells are hyperpolarized–and Up states , in which firing activity is intense and cells are depolarized [8] . The hippocampus generates sharp wave-ripple complexes ( SWR ) , in which a strong excitatory input from CA3 pyramidal cells leads to broadly distributed postsynaptic potentials ( the sharp waves ) in CA1 stratum radiatum , while the pyramidal layer shows a quick bout of high frequency LFP activity ( the ripple ) [9–11] . Ripples exist both in a quiet awake state and during slow-wave sleep , and disruption of ripple activity is known to impair memory [12 , 13] . In the rat , one of the mechanisms thought to underlie memory consolidation is place cell replay: a phenomenon in which the pattern of relative firing of hippocampal pyramidal cells that code for position ( place cells ) re-occurs during post-task sleep [14 , 15] . Importantly , hippocampal replay has been shown to take place during CA1 ripples , on very short time scales during which synaptic plasticity is likely to arise . Interestingly , SWRs are more likely to occur during cortical Up states [16] and may potentially influence the spatio-temporal pattern of Up state generation . Thus , understanding the process of ripple generation is a crucial step towards identifying the mechanism of brain-wide sleep-dependent memory consolidation . In this work , we propose a novel mechanism of CA1 ripple generation during sleep . Our in vivo data show that ripples have a broad frequency distribution , exponential inter-arrival times and a highly non-variable duration . In our model , high-frequency firing in perisomatic interneurons is caused by input from area CA3 , and mediates high-frequency local field potential ( LFP ) oscillations in CA1 pyramidal neurons . The main novelty of this model , compared to ones already proposed in the literature [17–19] ( see [20] for a review ) , is the prediction of the ability of CA1 to self-time ripple durations , and hence limit the extent of replay in a dynamic fashion , from ripple to ripple . We propose that phase-dispersion ( loss of synchrony ) induced by noise on the oscillatory dynamics constrains the duration of a ripple event . This minimal model is not only able to explain experimental data regarding basic ripple properties , but is also consistent with recent data on ripples and ripple-like activity triggered by optogenetic stimulations in vivo [21] . This paper is organized as follows: we first introduce the experimental results , then describe our computational model and show that it can produce ripple-like oscillations . We then use those observations to inform predictions that can be made by the full model . Next we study the role of interneurons in setting inhibitory transient mechanisms underlying ripple oscillations in the model , and the role of pyramidal cells in the overall ripple structure . We conclude showing that selective input from CA3 can induce sequential reactivation of CA1 pyramidal cells during ripples in our model .
Data shows that ripples are local events [22] , and that a given CA1 pyramidal cell rarely spikes more than once per ripple oscillation [10 , 23 , 24] . Many pyramidal cells are not recruited by ripples recorded across different sessions , while some are recruited by almost all events . On average 10% of pyramidal cells are active in any given ripple , and their spikes are locked to the trough of each the oscillations within a ripple event [25 , 26] . On the other hand , inhibitory interneurons in the pyramidal layer spike across the ripple , with a firing rate consistent with the ripple frequency [23 , 24] . This suggests a predominant role for pyramidal layer interneurons in organizing CA1 ripple firing . Moreover , spiking in CA3 is not locked to CA1 ripples [27 , 28] , which further advances the idea that ripples are an intrinsic CA1 rhythm , which can be initiated by sufficiently strong incoming inputs . We started investigating the nature of ripple oscillations by studying LFP recordings in hippocampal CA1 in rats . Representative examples of the wide-band and band-passed recordings are shown in Fig 1A . We focused on a few salient characteristics of ripple waves , such as frequency , duration and inter-arrival time ( defined as the time between a ripple event and the next ) . Fig 1 shows that ripple frequency , defined as the inverse of the average inter-peak interval during a ripple event , was normally distributed around 163 . 5 ( ± 20 . 6 ) Hz ( Fig 1B ) , their inter-arrival times were approximately exponential , with fitted rate 1 . 7748 Hz ( Fig 1C ) , and their duration ( Fig 1C ) was centered about 51 ( ± 9 . 4 ) ms showing a high-kurtosis distribution ( K = 20 . 1952 , where for a normal distribution K = 3 ) [28] . Note that ripple frequency is representative of the peak-to-peak time within a given ripple ( see S1 Fig for a representation ) , while the count of ripple events in a given time interval would be called ripple density , and can be found as the inverse of the inter-arrival times in such interval . Ripples are events specific to the pyramidal layer of CA1 , and ripples simultaneously recorded across different tetrodes appear to have amplitudes that vary independently [22] . This suggests that ripples are local events within the CA1 pyramidal layer [22] . Furthermore , data show that ripple events turn to epileptic activity when GABAA is blocked in CA1 slices [29] suggesting that interneurons limit the extent and sculpt the frequency content of these events . The population of pyramidal cells is most active close to the peak of a ripple event , which is defined as the time when the filtered LFP reaches its maximum amplitude ( often the biggest trough ) . Moreover , only a few pyramidal cells are recruited by ripples in CA1 . On the other hand , most basket cell interneurons spike in ripples , and across the event duration [23 , 24] . Our data revealed that the length of time between successive ripples is not dependent on any feature of the current ripple ( frequency or duration ) . In fact , scatter plots of ripple frequency vs time-to-next ripple , and ripple duration vs time-to-next ripple , did not show any particular correlation ( see S2 Fig in supporting information ) . Poincaré return maps show that both ripple frequency and duration are not dependent on the frequency or duration of the preceding ripples ( S2 Fig ) . This implies that we can model each ripple in CA1 as an event directly triggered by an incoming CA3 input volley . Hence , we looked for a ripple mechanism that can generate Gaussian-distributed frequencies for a fairly constrained time interval . We reasoned that our model needed to represent a small patch of CA1 , reached by strong incoming excitation from CA3 , so that the activity of all cells we modeled would be picked up by a single electrode . Given that ripples are measured in the pyramidal layer , we chose to model only pyramidal cells and parvalbumin positive basket cells . In fact , other interneuron types that might be active during sharp waves impinge on pyramidal cells at different layers [30 , 31] , which do not show ripple frequency oscillations . Hence , they might have a modulatory influence on the overall ripple appearance but are not in a position to set the pace of ripple frequency . The network consisted of 800 pyramidal cells and 160 interneurons , a ratio in agreement with CA1 anatomy [32] , and we used all-to-all connectivity , with the exception of the synapses between pyramidal cells , which were few and much weaker than all others [33] , consistent with CA1 anatomy [34] . Fig 2 shows a network representation ( Fig 2A ) , example traces from a pyramidal cell and an interneuron ( Fig 2B ) , and the distributions of synaptic weights ( Fig 2C ) in the model . To model each neuron , we used the Adaptive Exponential Integrate-and-fire model [35] , as it is simple ( only two variables ) and has been shown to reproduce many different spiking behaviors [36] , because of the expressed essential non-linearities [37] . To account for heterogeneity , each neuron received a different independent Ornstein-Uhlenbeck ( OU ) noise and a mean DC current to set baseline excitability ( see Materials and Methods for details ) . The noise term represents the in vivo state of the voltage in each cell , which is likely receiving a much higher barrage of synaptic inputs than the one provided by the network spiking activity in our model . The OU process , which can be thought of as a filtered white noise process , is used in dynamic-clamp experiments to mimic in vivo state in hippocampal slice recordings [38] . We assigned fast time scales to the synapses , in agreement with recent in vitro estimates [39] . The average synaptic strengths values were chosen to induce post-synaptic potentials of less than 1mV . To represent the integrated input from CA3 localized in time , we delivered input current to all cells , with different magnitudes for pyramidal cells and interneurons , due to the lower input resistance of pyramidal cells [40 , 41] ( see the first panel of Fig 2D; details in Materials and Methods ) . In the text below we will refer to this current as CA3 input to the CA1 network . All details of model implementation and justifications for specific parameter choices are reported in Materials and Methods . The rastergram and spike probability curves in Fig 2D show that when inputs from CA3 reached a patch of CA1 , high-frequency firing was triggered in the interneuron population , which self-organized in oscillations . Firing in pyramidal cells increased as well , but the probability of firing for the pyramidal cell population was much smaller than for interneurons ( 0 . 2 vs 0 . 005% ) . To compare our model to experimentally recorded ripples , we approximated the LFP in the pyramidal layer using the average net synaptic input ( from both excitatory and inhibitory cells ) received by all pyramidal cells , and derived a wideband LFP signal ( Fig 2D ) . The model generated ripple-like oscillations that could be detected by band-passing our LFP estimate , shown in the bottom panel of Fig 2D . Ripples produced by our computational model have properties consistent with the ones recorded in vivo . Fig 3 shows that the mean frequency for a given input intensity is 162 . 4 ± 12 . 5 Hz ( Fig 3A ) , and ripple duration is 57 . 2 ± 3 . 1 ms ( Fig 3B ) . Also , most pyramidal cells do not spike during a ripple , in fact on average a ripple shows spikes from 14 . 76% of the pyramidal cell population ( Fig 3D ) , and those that do will not spike across the ripple duration , but only once ( Fig 3C ) in agreement with previous experimental observations [42] . Furthermore , the spiking activity of pyramidal cells is known to precede the spiking of interneurons within each ripple wave [43] . We tested this property in our model by computing the cross-correlation between the filtered LFP and the firing probability of each cell population , averaged across 40 ripples . Fig 3E shows that peaks in the correlation between pyramidal cell spiking and LFP preceded those for interneuron activity across ripple waves . For completeness , we also verified that there was no inherent rhythmic activity in the network background state that could be inducing this relationship within ripples beyond the mechanistic phenomena we report ( S3 Fig ) . This model is consistent with CA1 activity during non-REM sleep in vivo , when ongoing theta-gamma activity is not present in the background . We next studied the main properties of the dynamics of ripple oscillations in our model , looking for intrinsic CA1 properties that played a role in shaping ripple spiking . We found that CA1 properties determine ripple duration , while ripple frequency is not controlled by the time scale of inhibitory synapses . Also , the amount of pyramidal cells spiking during ripples is determined by competing forces: the excitatory drive they receive from CA3 and the amount of local inhibition they receive from the CA1 inhibitory population . Since the input current ( representing the sum of synchronized spiking in CA3 ) caused ripples to initiate , we asked if ripples would continue oscillating for as long as the input was present . Fig 4A shows that the ripple LFP duration stayed un-varied independently of the different CA3 input durations we tested . The band-passed LFPs for 40 ripples across different input durations are shown in gray , while the black line is their average . The graph shows that even if spiking was still enhanced for the duration of CA3 inputs , the organized oscillatory activity was lost after about 60 ms for all cases considered . This emphasizes that CA1 can control ripple duration , even if it cannot control their initiation . Pyramidal cell spiking within a ripple is responsible for carrying information to downstream areas . Hence , it is important to understand what factors regulate the overall recruitment of CA1 pyramidal cells to a given ripple in our model . The ratio of CA1 pyramidal cells recruited in a given ripple in our model is modulated by both excitatory drive from CA3 on this population and inhibitory currents within CA1 [44] . In fact , increasing CA3 input to this population raised the percentage of pyramidal cells spiking on every ripple ( Fig 4B , doubling the peak value takes the recruitment percentage from 14% to above 90% ) , and increasing the time scale of inhibition onto pyramidal cells reduced their recruitment ( Fig 4C , setting τ = 6ms results in an average 4 . 4% of pyramidal cells spiking in any given ripple ) . Fig 4C also shows that increasing inhibition onto interneurons resulted in higher pyramidal cell excitability , because the inhibitory population activity was reduced overall . Thus , a net change in the GABAA time scale has competing effects on pyramidal cells recruitment to ripple activity , leaving the fine regulatory function to highly selective CA3 input . The predominance of CA3 input over local inhibition in choosing which pyramidal cells are recruited to a specific ripple is consistent with the known synaptic plasticity at Shaffer collateral synapses between CA3 and CA1 pyramidal cells [45] . Furthermore , we found that changing the time scale of inhibition did not drastically slow ripple frequency ( Fig 4D ) . To better understand mechanism underlying the dynamics of ripple oscillations , we next moved to studying a simplified system . Since pyramidal cells typically spike less and at much lower frequencies than inhibitory cells during ripples , we started from studying the role of interneurons in the ripple dynamics we observed in the full model . To do that , we considered a network of only inhibitory neurons receiving a step of DC current , and studied the resulting input-driven high-frequency population firing . The step of DC current delivered to all inhibitory interneurons amounted to the same value as the peak input current from CA3 to interneurons in the full model ( 700 pA ) . To construct the profile of spiking probability in response to the current step , we run 100 simulations for each parameter set , and built the cumulative histogram of probability of spiking as a function of time . Fig 5A shows a schematic of the reduced network , while Fig 5B shows that the common input step ( at time = 1s ) initially organized the network as indicated by the rhythmic oscillations of the population activity . The amplitude of these oscillations progressively decreased , indicating a transient nature of the high-frequency activity , which was de-synchronized by the intrinsic noise . Eventually , the population firing rate stopped oscillating and settled to a mean constant value , which depended on the size of the current step . This behavior of the isolated interneuron population is consistent with data [46]: in vitro optogenetic experiments ( albeit in area CA3 ) show that activating only parvalbumin positive interneurons with a step of light , of duration up to 50ms , results in an average oscillatory behavior in which the peaks are progressively attenuated in time . Once we found that synchronous inputs organize transient oscillatory patterns in our purely inhibitory network , we asked what role synaptic inhibition could play in setting the properties of this transient oscillation . We reasoned that the role of inhibitory synapses in network behavior could be twofold: they promote synchrony while enough synaptic inputs are aligned ( early after the initiation of the step of current ) , but synaptic currents switch to propagating de-synchronization if enough neurons are asynchronous . We numerically studied the role of inhibition in transient high-frequency synchronization by changing two main parameters: the decay time scale of the inhibitory synaptic conductance ( τ ) and the strength of synaptic inhibition . To address synaptic strengths while respecting the choice of a normal distribution of synaptic conductances around a mean , we introduced a non-dimensional scaling factor α , which we systematically varied . When α = 0 , no synapse was active , and when α = 2 all synapses were twice as strong as their baseline values . Fig 5B is provided to show the effect of changing the inhibitory signal: the overall impact of inhibition on the oscillation pattern depended on the decay time scale τ , or synaptic strength α , or both . As one can see in Fig 5 , changing strength of synaptic inhibition in the inhibitory network affects both the duration of transient synchronization ( Fig 5C ) ( defined as a time window when amplitude of oscillations is still high , see Materials and Methods , Analysis of inhibitory network model ) and its frequency ( Fig 5D ) ( defined as the inverse of the average peak-to-peak time delay , within transient duration ) . Because of that , the number of oscillatory peaks within a transient changed as well ( Fig 5E ) . Fig 5 shows that frequency remained within ripple range for a time scale of inhibition within a broad range of 2–6 ms as long as synaptic weights were limited to within about 200% of the baseline value . Also , in the condition of α = 0 , when the network is disconnected , one can still see the noise-induced de-synchronization; when the network has active and fast ( small τ ) synaptic connections the de-synchronization is delayed ( Fig 5C ) . The increase of inhibition ( through longer decay times or stronger conductances ) resulted in the reduction of both the transient oscillation frequency ( Fig 5D ) and its peak count ( Fig 5E ) . This is an important observation and it highlights the point that oscillations in this system are not a rhythm in the traditional sense of the word: they are not an oscillation that would persist in time as long as there are no changes to the CA1 model or the input , but rather a transient arising from a strong initial input capable of synchronizing CA1 neurons . The last point is critical: given that the stationary state of the network receiving a step current input is non-oscillatory , if the initial common step fails to synchronize enough neurons , then the transient oscillation would only last about 1 or 2 cycles , if at all . To further explore this point , we studied how the transients organize when the initial step of the input current is halved ( Fig 6A ) . In that case , much fewer interneurons were recruited to the initial synchronous population ( note the scale of firing probability on the y-axes of Fig 6A ) . Inhibitory currents still affected oscillations , but the transient lasted very few cycles ( Fig 6B ) and peaks were smaller . We concluded that if the initial current step failed to synchronize a large enough population of neurons , the resulting slower oscillation faded in only 2 to 3 cycles . Hence , this network shows an all-or-none property: a smaller step of input that could in principle recruit lower frequency oscillations cannot recruit a transient at all . In fact , it takes an input of sufficient size to generate a transient that lasts enough cycles and recruits enough neurons for a fast oscillation to be visible in the LFP . We next investigated the role of noise in setting the duration and frequency of oscillatory transients in the inhibitory network . We progressively reduced the noise coefficient using a scaling factor σ . Fig 6C shows the network behavior in the case of reduced noise amplitude , where spikes were far more synchronized around the network peaks compared to the 100% noise case shown in Fig 5B . Fig 6D shows that the duration of transient oscillations was controlled by the noise level ( larger noise resulted in shorter transients ) , while frequency was not . In summary , the network oscillations we observed in our purely inhibitory network do not emerge out of interaction among the neurons , but rather from receiving a common , synchronizing , input . If the neurons were identical , received identical inputs , had no noise , and had no CA1 synapses then the oscillations would be trivially persistent , since all of the interneurons would remain in the phase locked state . The transient nature of this oscillation was due to network heterogeneity , introduced in our model via a different direct current to each neuron , and independent noise ( where larger noise introduced more heterogeneity ) . Input strength , noise size , synaptic inhibition strength and duration are all factors that can control and tune the transient behavior reported in our model . Our findings in the purely inhibitory network have direct implications for the full hippocampal model: due to the underlying transient mechanism that is highlighted by the dynamics of the reduced inhibitory network , local CA1 properties such as the level of noise and the strength of inhibitory synapses will control ripple duration , while the degree of synchrony of CA3 inputs ( which is represented by the amplitude of the input current in our model ) will determine whether a ripple occurs . In fact , if CA3 spikes were not synchronous enough , the net sum of the post-synaptic currents impinging on CA1 cells would be small , and hence the size of the CA3 current input we deliver to the model would be small as well . Therefore , the size of our current input from CA3 is effectively a model for synchrony in CA3 pyramidal cells spiking . Ripple frequency depends on balancing the number of interneurons recruited by the ongoing activity with the lateral inhibition known to exist between interneurons . Since , in principle , we could find oscillations in reduced model in the case of α = 0 , where no synapses between the interneurons were present , we checked if removing I-to-I synapses in the full model would affect the properties of CA3-input driven ripples ( S5 Fig ) . We found that ripple frequency transients can indeed be elicited in the network , however duration of ripples was generally reduced and the likelihood for occasional very short transients increased . The inhibitory network also showed that an input step of current too small would result in very few ( 1 to 3 ) oscillatory peaks , effectively failing to induce the transient . In the full model , we verified that reduction of the input size ( to both pyramidal cells and inhibitory neurons ) resulted in decrease of the ripple amplitude and disruption of ripple events ( S4 Fig ) . This was particularly evident when the input was scaled to 30% of its magnitude . As a result , the ripple , and consequently a bout of memory consolidation , only occurs if activity in CA3 is sufficiently synchronous , and results in a strong enough input to CA1 . This ‘synchrony threshold’ ensures that uncorrelated CA3 inputs are essentially ignored by CA1 . Furthermore , this means that a transient oscillation is triggered only at high frequencies above the gamma ( 30–90 Hz ) range . As it is known that the interaction between pyramidal cells and basket cells in hippocampal CA1 region underlies gamma ( 30–90 Hz ) oscillations [47 , 48] , the above findings suggests the mechanism recruiting ripple oscillations can co-exist with , and be independent from , other slower rhythms that arise in the same region . It is also important to note that ripple characteristics ( frequency , duration ) in the full model receiving input from CA3 that was scaled in size saturated at around 80% of the baseline input and remained stable above this value ( S4 Fig ) . Thus , ripple properties predicted by our model to match experimental data are observed in the broad range of CA3 mediated input amplitudes , i . e . , structurally stable . Our results on the effects of reducing CA3 input on CA1 ripples are consistent with experimental observations made in simultaneous CA3-CA1 recordings [27] , in which sizeable ( hence synchronous enough ) CA3 activity was related to CA1 ripples , while smaller ( less synchronous ) CA3 activity did not induce ripples in CA1 . After investigating the role of inhibitory spikes in ripples , we moved to study the role of pyramidal cells activity in our model . In fact , recent optogenetics work [21] has raised the interesting idea that the minimal circuit to obtain ripples in CA1 may include local excitatory synapses on inhibitory interneurons . Specifically , CA1 in vivo recordings showed that when activating both parvalbumin positive interneurons and pyramidal cells in the pyramidal layer , high-frequency oscillations ( HFO ) in the LFP emerged . When only pyramidal cells were driven optogenetically with a step of light , HFO were also measured , although their duration was shorter than HFOs obtained driving both neuron populations . In contrast , if only fast-spiking parvalbumin positive inhibitory interneurons in the pyramidal layer were driven to fire , LFP oscillations were not found . We set out to see if our model could show results consistent with these experiments . We started by delivering input current only to interneurons ( Fig 7A ) : since simulations of a purely inhibitory network showed that current size controls synchrony among the interneurons ( Fig 6A ) , the size of the current step was first kept the same as for the full model . In Fig 7B we show examples of spiking probabilities for two stimulation conditions ( 50% or 100% of all interneurons were stimulated ) in the model when only interneurons received inputs . The inhibitory activity was still oscillatory , however all pyramidal cells were hyperpolarized , resulting in a field potential consisting of shunted currents . As a consequence , the LFP was much smaller than the one in the full model ( compare Fig 7B with Fig 2D ) ; LFP amplitude increased with the percent of interneurons recruited ( Fig 7C ) . Specifically , in the case where 50% of interneurons were stimulated ( which is likely much larger than any optogenetic stimulation in vivo ) , we found that the LFP amplitude was only about 10% of the LFP observed when both excitatory and inhibitory neurons received an input drive from CA3 . Increasing the amplitude of current stimulation in the model led to even stronger shunting of pyramidal neurons and even smaller LFP amplitude . In comparing with experimental results , we emphasize that the reduction of the LFP amplitude in our model depends on shunting effect of inhibition on pyramidal neurons , which in actual in vivo experiments would depend on the cell geometry , location of excitatory and inhibitory synapses , and other factors that our minimal cell model cannot explicitly capture . Even in this simplified setting , we are able to show a qualitative match with the strong reduction of the amplitude of oscillations in the LFP . Thus , our model reveals that if only interneurons are driven by a light source , LFP oscillations will be very small in amplitude , regardless of the stimulation amplitude , to the point that they will be experimentally negligible , in agreement with in vivo results [21] . We next tested whether the input delivered only to pyramidal cells ( Fig 7D ) could give rise to short bursts of oscillations . Note that to achieve an initial spiking rate capable of triggering fast network oscillations , pyramidal cells had to receive a current step size bigger than the one used for the full model . Since light-driven cells in optogenetics experiments are not receiving a fixed step of current that is identical across populations , we chose to adjust the current size to the measured behavior . In vivo , the LFP initially showed high frequency oscillations ( about 150 Hz ) , but only for about 25 ms [21] . Immediately after the initial high frequency response , the LFP amplitude in the high-frequency band decreased and the activity slowed down to about 80 Hz . In agreement with these data , in our model , we first achieved a high frequency firing for about 25 ms , but the pyramidal cell population was not able to sustain firing much longer ( Fig 7E ) . A variable controlling the ability of pyramidal cells to sustain firing is the strength of spike-frequency adaptation , which in the model equations is controlled by the size of the jump ( b ) imposed on the slow refractory variable ( w ) after each spike ( see Materials and Methods ) . Since change in adaptation can be biologically achieved by a variety of neuromodulatory phenomena , and the less adaptive the cell is , the longer it could sustain firing , we tested whether the current input to pyramidal cells could generate ripple-like oscillations in the networks with progressively less spike-frequency adaptation in the pyramidal cell population . To reduce adaptation in our neurons , we multiplied the b value by a scaling factor , in a range 10–100% . We found that even when pyramidal cell firing can be sustained ( Fig 7E ) , oscillations still showed the same profile of initial high frequency response that quickly slows down to a frequency below the ripple range as observed in vivo . Specifically , the amplitude of the LFP filtered in the high-frequency band ( to detect the HFO ) initially peaked but then dropped very quickly , which is consistent with what was shown in vivo ( Fig 7F ) . We concluded that our model , in which CA3 input to both pyramidal cells and interneurons is necessary to trigger a ripple in CA1 , is consistent with optogenetics data [21] . So far we have focused on a global mechanism of ripple generation , in which CA1 pyramidal cell timing within ripples is regulated by the ongoing rhythmic inhibition and incoming CA3 input . During sleep , CA1 pyramidal cells that spike in ripples are known to reactivate in firing order consistent with the one recorded during behavioral tasks [15 , 26] , and this property is thought to be a hallmark of a memory trace in the hippocampus . Memory trace reactivation has been measured in CA1 , but it is not clear whether the mechanism inducing reactivation within ripples is intrinsic to CA1 or due to input . We set out to see which biologically relevant properties could control ordered pyramidal cell reactivation across ripples in our model . It is known that neurons recruited by behavior to form a memory trace tend to have higher excitability [49–51] . We reasoned that cells in CA1 that reactivate during a ripple might have higher intrinsic excitability . Since , in our model , pyramidal cells are picking windows of opportunity to spike during ripples , cells with substantially higher DC levels will have an easier time finding a window of opportunity in which to spike ( because overcoming the incoming inhibition would be easier for them , compared to all other cells ) . In our model , we tested whether all it takes to be replayed in sequence for a set of pyramidal cells activated during the training phase ( and therefore known to have higher DC , and to spike in more ripples because of that ) is the fact that they are more depolarized . If that was the case , than we would know that the sequential reactivation in our model results from competition between intrinsic cell depolarization and ongoing inhibitory oscillations . It is also known that the Schaffer Collaterals ( the projections from CA3 pyramidal cells to CA1 neurons ) are plastic [52] , hence potentially target-selective . In other words , neurons in CA1 could receive selective inputs from CA3: higher and more intense inputs from some CA3 pyramidal cells ( presumably the ones correlated with the same behavior ) and lower and less intense from other CA3 pyramidal cells . The selectivity of CA3 input could be further modulated by hippocampal inhibitory neuron types that are not modeled in our network , which are hypothesized to gate Entorhinal cortex input and CA3 input on CA1 pyramidal neurons [18 , 53] . Hence we have two possible mechanisms with the potential of inducing reactivation: intrinsic CA1 excitability is a parameter of CA1 properties that could induce sequence reactivation , while CA3 selective input is a potential input-dependent mechanism for sequence replay in CA1 . To test these two properties , we randomly selected 10 pyramidal cells ( “sequence” cells ) to represent neurons that reactivate sequentially during ripples . First , we increased the constant current input that the 10 selected neurons received ( Fig 8A ) , which resulted in the appearance of a small peak at high values in the distribution of excitability of all pyramidal cells of the CA1 network ( Fig 8A ii ) . This was introduced as a mechanism to increase the likelihood of selected cells to spike during ripples . In a second case , we changed the time course of incoming current input , only for the selected sequence cells ( Fig 8B ) . As shown in Fig 8B–8I , each sequence cell received an input that had a peak in a narrow time window within the duration of a ripple . This peak represented the spiking of the sub-set of CA3 pyramidal neurons that were preferentially connected to the target CA1 cell; thus we assumed that there are CA3 neurons that would spike during a sequence-specific time window of the sharp wave event . In other words , in this case our model assumes that during a sharp wave-ripple there is an organized reactivation in CA3 , inducing selective inputs to CA1 , which results in sequential spike reactivation in CA1 . Finally , in a third case , we combined both manipulations on our selected cells: increased excitability and selective CA3 inputs . For each case , we show whether the inputs from CA3 are selective in panel i , and whether sequence cells received extra excitability in panel ii . In our model , sequence replay needed to lead to two properties: ( a ) specific spike sequences repeated more often than chance and ( b ) the temporal order of the spikes of CA1 cells within those sequences was consistent across ripples . For the replay properties to be satisfied , we checked if our sequence cells spiked in a greater fraction of ripples compared to all other cells ( panel iv ) , and spiked in a consistent order across ripples ( panel iii ) . To verify that , we computed the spike time difference between sequence cells , and averaged across all ripples in a 10 s simulation ( 41 ripples ) . If the order of cells was maintained across ripples , the average spike time differences will look like shifted versions of the same line . Fig 8 shows the effect of these manipulations . In Fig 8A , selected cells only received enhanced intrinsic excitability ( see a ii ) without selective temporal ordering in CA3 input ( see a i ) . Note that , in this case , there was no sequential spiking behavior during ripples ( see a iii ) , and sequence cells did not spike in more ripples than all other cells ( see a iv ) . In Fig 8B selected cells only received selective CA3 input ( see b i ) , without enhanced intrinsic excitability provided by constant direct current ( see b ii ) . Note that , in this case , the orderliness of spiking across ripples was overall preserved ( see b iii ) . Also note that the fraction of ripples visited ( b iv ) was higher for sequence cells compared to all other cells . In Fig 8C we show that CA1 neurons that received selective input from CA3 ( c i ) and higher intrinsic excitability ( c ii ) showed spiking in a greater fraction of ripples compared to all other cells ( c iv ) , and spiked in a consistent order across ripples ( c iii ) . Note that compared to condition b , in which the additional intrinsic excitability is not present , the consistency of ordered firing across ripples is improved . In summary , we found that being more depolarized served the selected cells well in terms of the number of ripples during which they spike , but not for spiking in the orderly fashion that is found experimentally . On the other hand , selective CA3 input seemed to be effective at inducing sequence reactivation . The necessary elements of the model to allow for mapping of the structure in CA3 input onto spiking in CA1 pyramidal cells were 1 ) generalized input lasting throughout the sharp wave event to most cells in CA1; 2 ) time-selective , ordered input from some CA3 cells to some pyramidal cells in CA1; 3 ) the selective input , when present , results in current delivered to its target ( sequence ) cells at higher magnitude than generalized input delivered to the non-sequence cells ( because of plasticity in CA3-to-CA1 synapses ) . Our model therefore predicts that replay in the hippocampus is generated in CA3 ( and possibly in the dentate gyrus ) and ordered reactivation in CA3 is required for reliable sequential spiking in CA1 , where it is packaged in a fast-rhythm at rates that are conducive to synaptic plasticity , to be transmitted to target regions , such as the cortex . Intrinsic CA1 properties , such as heightened excitability of selected cells , can enhance but not cause the ordered replay .
In general , there is an agreement in the field that parvalbumin positive interneurons are involved in LFP ripple oscillations , since they spike at ripple frequency across the event duration [31] and removing GABAA cancels ripple activity [29] . Differences arise in the interpretation of the specific mechanism underlying the oscillations . Two main hypotheses have been proposed . According to one idea , a ripple in CA1 is exactly the same phenomenon as its triggering CA3 excitatory event , simply propagating from one hippocampal sub-region to the next . In this setting , a ripple can be thought of as fundamentally one whole excitatory event reverberating across the hippocampus , much like throwing a small stone in still water [17 , 58] . Models that account for such behavior have to include non-standard excitatory mechanisms among CA1 pyramidal cells , such as more-than-linear dendritic summation [58] or electrotonic connections ( known as gap junctions ) [17 , 59 , 60] . However , recent in vivo recordings from a strain of gap-junction deficient mice still showed sharp wave-ripple events at frequencies similar to that of the wild-type [61 , 62] . This led to the idea [19 , 53 , 57] that it is appropriate to introduce models in which CA1 synapses can generate ripples . According to this approach ripples are seen as a local phenomenon in CA1 triggered by incoming CA3 excitation [18 , 19 , 57] , in which inhibitory synaptic connections are responsible for the oscillations in CA1 . Our model is consistent with the latter hypothesis . We propose that the minimal model capable of ripple generation is an inhibitory network receiving a brief wave of excitation . The crucial role of parvalbumin positive basket cells in organizing ripple oscillations has been previously shown by Schlingloff et al [46] , who used a network of only parvalbumin positive interneurons to study ripple frequency when a step of current was applied to the population . While in this earlier model the loss of reciprocate inhibitory synapses induced a loss of rhythmicity , we now show that oscillatory firing in such inhibitory networks is controlled by a synchronous and strong common input , which is characteristic of a transient oscillation . The strength of recurrent connectivity between inhibitory interneurons plays a critical role in determining the type of oscillatory activity the inhibitory network can produce . Indeed , if synaptic connections were strong , a purely inhibitory network with enough reciprocate connections would give rise to gamma oscillations ( 30–90 Hz ) , paced by the duration of inhibitory currents: a mechanism known as Interneuron-Network Gamma ( ING ) [63 , 64] . ING oscillations persist as long as cells are driven to fire . In contrast , the stationary behavior of our model is a disorganized firing state . Oscillatory persistent firing in a network of irregularly spiking inhibitory neurons depends on synaptic strengths and the size of noise [65]: our model belongs to the asynchronous stable state part of the bifurcation diagram ( Fig 5A in [65] ) , where the size of noise overcomes the ability of mutual inhibitory synapses to organize the firing rate in oscillations that would be below ripple frequency . Crucially , we assume synapses to be weak , meaning that inhibitory currents do not overcome the effect of intrinsic noise even when a lot of interneurons are spiking . Hence , interneurons are driven to fire by CA3 inputs and not slowed down by inhibition . As a consequence , the ING mechanism is not arising in the network , and inhibitory currents are not setting the firing frequency , but merely modulating it . Instead , upon input arrival , inhibitory interneurons become transiently synchronized , leading to high frequency LFP oscillations , with properties matching in vivo data; the synchronization then disappears after a characteristic duration in the presence of noise . While the minimal mechanism is identified in a purely inhibitory network , we emphasize that transient oscillations in that reduced model did not last longer than 40ms . This underscores the role that pyramidal cells and their interaction with interneurons still play in shaping a realistic ripple oscillation in the full model we present . Our model predicts that synchrony of interneuron firing should decrease over the ripple duration . Spontaneously occurring ripple-like activity in slices suggest that during a ripple excitatory and inhibitory currents onto pyramidal cells oscillate at ripple frequency and increase in size during the progression of a ripple [66] , which likely reflects the increasing number of cells involved in the slice spontaneous event , rather than a higher synchronicity in cell firing . On the other hand , optogenetically induced ripples in vivo and in vitro seem to have a number of similar properties , and in vivo light-triggered events are likely to account for spontaneous , physiological ripple oscillations . Hippocampal slices in which ripples are triggered with optogenetic drive show that synchrony among nearby inhibitory neurons decreases across a ripple event [46] . Beyond parvalbumin positive basket cells , different hippocampal interneuron types show different spiking behaviors during ripples [30 , 31 , 43 , 67] , and they might be involved in the fine timing of specific pyramidal cell spiking or recruitment , and in gating entorhinal and thalamic inputs on CA1 pyramidal cells . Furthermore , post-inhibitory rebound in pyramidal cells has been proposed as a mechanism for ripple initiation in CA3 [57] . Previous models have been trying to dissect potential separate roles for different interneuron types in gating entorhinal and CA3 input on CA1 and potentially contribute to fine selectivity of the pyramidal cells recruited by a given ripple [53] . The goal of our study was to find a minimal model capable to explain ripple oscillations frequency and duration in CA1 , therefore we did not model the multitude of interneurons beyond parvalbumin positive basket cells . We focused on the pyramidal layer and its main spiking actors during the events our model represents . We emphasize that in previous modeling approaches [17 , 19 , 58] ripple duration is established by the duration of excitatory propagation within CA3 . We show here that this does not have to be the case . In fact , we predict that the CA1 network is responsible for determining the duration of the organized firing during ripple oscillations . The origin of the excitatory CA3 event which initiates a ripple in CA1 is not yet clear [46] , and goes beyond the scope of this work . We predict that such CA3 events could show a broad range of durations , but they will still induce ripples of a fixed length , lasting about 50–80 ms in CA1 . This implies that the CA1 region can produce a standardized “package” of information with every ripple , which is projected to downstream regions . Within this package , pyramidal cells spiking is organized by selectivity of CA3 input , and possibly input directly from other brain areas . We also predict that pyramidal cell replay in CA1 is organized by pyramidal cell replay in CA3 , consistent with the fact that CA3 to CA1 connectivity is crucial for memory consolidation [54] . A common hypothesis is that synaptic connections between CA1 pyramidal cells and intrinsic cell properties regulate the order of spiking in a CA1 sequence [53] . In our study we tested this hypothesis first . Since in CA1 very few pyramid-to-pyramid synapses are found , we only had to check the effect of DC input . We found that preferential DC levels , even much stronger than average , were not enough to guarantee a robust repetition of the correct order of firing in our target cells . We then tested the potential role of the incoming CA3 input . Given the projection onto CA1 pyramidal cells are plastic , and that CA3 is the most typical set in which Hebbian learning is found in the brain , we formulated the hypothesis that CA3 is replaying its own cells , and projecting selectively to CA1 the activity that is relevant to behavior , because cells in CA3 and CA1 were active together during behavior ( and hence learning ) . We then found that a combination of the two factors was best at inducing hippocampal replay in CA1 in our model . Our study further predicts that when activating only a fraction of basket cells , the mechanism for ripple oscillations is present but not measurable in the LFP at the stimulation location . In fact , pyramidal cells would be shunted by an incoming barrage of inhibition and would not be sustained by excitation . This results in a shunted LFP in the CA1 pyramidal layer . While our minimal model does not show an exact quantitative match with experiments , our representation of the effect of optogenetically driving only interneurons results in a strong reduction of the magnitude of the LFP , which we believe is consistent with what has been found in published data [21] . Our model is also consistent with experimental results on optogenetic stimulation of only CA1 pyramidal cells . Both data and our model show the initial high frequency events lasting about 25 ms were quickly followed by oscillations at a slower frequency . This means that ripple generation in vivo has to rely on additional mechanisms other than excitation of pyramidal cells to sustain firing beyond 25–50 ms . We suggest that the CA3 input driving both the excitatory and the inhibitory populations in CA1 pyramidal layer raises the firing rate of basket cells to within ripple frequency range , inducing transient oscillations , while maintaining the excitatory population above shunting level , so that selected pyramidal cells can fire during ripples , and so that their spike timing stays modulated by ripple phase . In our model , input from CA3 activates both CA1 interneurons ( hence triggering ripple activity ) and CA1 pyramidal cells . The selection of which pyramidal cell is recruited to spike within a given ripple , and their potential sequential activation , is the result of a balance between CA3 drive , ( potentially filtered by different types of interneuron populations omitted in the model [18 , 53] ) and local feedback inhibition in CA1 . Hence , CA3 is seen as the place in which the initial reactivation of a memory can take place , while CA1 optimizes CA3 output for downstream transmission . In other words , ripples in CA1 prepare a time-bounded package of carefully selected pyramidal cell spikes , encoding information that can then be routed across neocortex . It is also possible that incoming excitatory input from entorhinal cortex could be responsible for initiating a CA1 ripple phenomena , and/or regulate the recruitment of CA1 pyramidal cells within a ripple . In fact , this computational model represents CA1 as a network which is ready to burst , just waiting for an incoming input to select a local group of interneurons to oscillate , so that they can organize the timing of pyramidal cells , recruited by the input combined with their initial excitability .
Data were recorded using extracellular tetrodes targeted to region CA1 of the hippocampus in 6–7 months old Brown Norway rats during natural sleep . All experiments were approved by the University of Arizona IACUC and followed NIH guidelines . The recorded LFP where band-passed between 0 . 1 and 500 Hz , and SWR complexes were found when the filtered LFP ( 100–300 Hz ) crossed a threshold of 2 standard deviations of the baseline . The center of the SWR was positioned at the peak values of the LFP envelope and SWR start and end were found as the first points around the peak where the envelope passed below a threshold of half the distance between the peak and the baseline , see S1 Fig for a schematic . Ripple duration was then computed as the time between SWR start and end . Note that discrepancies between the numbers reported for ripple duration in our work compared to other reports could be induced by different methods of finding ripple durations , beyond potential differences between rats and mice . In our method of detecting ripple starts and ends , we use as a threshold the half distance between baseline and peak . That is , every ripple has its own threshold , which depends on the size of the ripple amplitude . If one were to use one absolute threshold for all ripple envelopes to find their start and end time points , ripples that have larger peak amplitudes would stay longer above threshold , and the average value of ripple duration would be larger as a result . This difference in methods can contribute to the apparent difference in the data across the many ripple papers in the field . Ripple frequency was calculated as the average inter-peak interval of the filtered LFP within the SPW duration , see S1 Fig for a schematic . The Gaussian fit for the frequency distribution was found using MATLAB ( www . themathworks . com ) function fitdist . Ripple inter-arrival times are simply the time difference between a ripple onset and the next . The exponential fit for the distribution was obtained using function expfit in MATLAB . Data is presented as standard deviation from the mean . The CA1 computational model includes 160 interneurons and 800 pyramidal cells . For each neuron , the equations are Cv˙=−gL ( v−EL ) +gLΔexp ( ( v−Vt ) Δ ) −w+I ( t ) τww˙=a ( v−EL ) −w v ( t ) =Vthr⇒v ( t+dt ) =Vr , w ( t+dt ) =w ( t ) +b Parameter values are as follows . For all pyramidal cells: C = 200 pF; gL = 10 nS; EL = -58 mV; a = 2; b = 100 pA; Δ = 2 mV; τw = 120 ms; Vt = -50 mV; Vr = -46 mV , Vthr = 0 mV . For fast spiking inhibitory interneurons: C = 200 pF; gL = 10 nS; EL = -70 mV; a = 2; b = 10 pA; Δ = 2 mV; τw = 30 ms; Vt = -50 mV; Vr = -58 mV . Every cell is receiving a different input , all with the same structure: I ( t ) =IDC+βηt+Isyn ( t ) +Iinp ( t ) τdηt=−ηtdt+dWt Iinp ( t ) =Imax ( 1+exp ( −t−tonk ) ) −1 ( 1+exp ( t−toffk ) ) −1 Where IDC is a constant , different for each cell , selected from a normal distribution ( mean 40 pA for pyramidal cells and 180 pA for interneurons; standard deviation 10% of the mean for both populations ) . ηt is a stochastic process known as Ornstein-Uhlenbeck ( OU ) process , with a filter time scale imposed by the τ value , in our case 100Hz . The coefficients were β = 80 for pyramidal cells , β = 90 for interneurons . To computationally introduce a stochastic process that solves the equation for ηt above , we first generated its representation in the frequency space taking advantage of its known power spectral density [68] , and then computed its inverse Fourier transform ( ifft in MATLAB , www . themathworks . com ) . Synaptic currents are modeled with double exponential functions , for every cell n we have Isyn ( t ) =∑j=1160gj→nsj→n ( t ) ( vn−Ei ) +∑j=1800gj→nsj→n ( t ) ( vn−Ee ) sj→n ( t ) =∑spikesofcelljF ( eH ( −t−tkτD ) −eH ( −t−tkτR ) ) with F a normalization factor that ensures at every spike the double exponential peaks at one , H ( · ) is the Heaviside function . Every gj -> n is selected from a Gaussian distribution with a given mean and standard deviation 10% of the mean ( see Fig 2 ) ; values below 0 are rectified to 0 . Mean synaptic values are g¯Int→Int = 0 . 0234 nS , g¯Pyr→Int = 0 . 0083 nS , g¯Int→Pyr = 0 . 0521 nS , g¯Pyr→Pyr = 0 . 001 nS . The reversal potential for synaptic currents are Ei = -80 mV and Ee = 0 mV . The time scales of synaptic rise and decay are as follows: for excitatory synapses on pyramidal cells [34 , 39] we have τR = 0 . 5 ms and τD = 3 . 5 ms; while on interneurons we have τR = 0 . 9 ms and τD = 3 ms . For inhibitory synapses on interneurons: τR = 0 . 3 ms and τD = 2 ms , on pyramidal cells τR = 0 . 3 ms and τD = 3 . 5 ms . Iinp ( t ) represents the net effect of CA3 synaptic excitation on CA1 cells . ton and toff , in most simulations are kept 50ms apart . Imax = 210 pA for pyramidal cells and Imax = 700 pA . Network simulations were solved with a 1-step Euler algorithm ( Δt = 0 . 001 ms ) using MATLAB . ( www . mathworks . com ) In this section we introduce the rationale behind the choices of equations and parameters used to model CA1 ripple activity and CA3 input to CA1 . We model only the pyramidal layer of CA1 , because that is where ripples are usually measured . To reveal the basic mechanisms of ripple generation , we do not model the rich number of interneuron cell types known to exist in CA1 , but only the parvalbumin positive basket cells , which are active during ripples [32] and are a predominant interneuron type in the pyramidal layer . Ripple oscillations are only seen in the pyramidal layer [24] , so we had to include at least the populations known to be dominant in such layer: pyramidal cells and fast-spiking interneurons . While a rich number of interneuron cell types are known to exist in CA1 , we thought that they were going to have limited impact on ripples , for the following reasons . Bi-stratified cells are known to spike during ripples , hence they could potentially also contribute , but they project on the proximal dendrites of pyramidal cells , while basket cells project at the soma . This means that synaptic events due to basket cells spiking will have bigger representations and effect on pyramidal cells voltage , and ultimately spikes . Another cell type known to be important to hippocampal theta rhythm , the OLM cells [69] , might be relevant outside ( that is , before and after ) ripples . OLM cells spike only in about half the ripple episodes in naturally sleeping animals [23] , and transmit inhibition to pyramidal cells via slower IPSCs ( longer than 10ms time scale [70] ) . This inhibitory time scale , together with their intrinsic oscillatory properties , makes OLM cells amazing candidates in the participation to hippocampal theta rhythms [69 , 71–73]; however , ripples are fast oscillatory events caused by incoming CA3 signals: the role that OLM cells could reasonably play in such events ( given they do not spike during half of them ) is potentially modulate the amount of excitation it would take to a CA3 input to start a ripple . In other words , they could be contributing to switching from ongoing theta activity to ripple activity in awake state [74] . Since our is a model of what happens when the sharp wave activity from CA3 hits CA1 and therefore fast oscillations arise , we are not modeling the switch between theta and other rhythms . Therefore , we consider it acceptable to omit the modeling of OLM cells activity during ripples in our model . We emphasize that if this was a model of awake state , in which ongoing theta-gamma oscillations are interspersed with sharp wave-ripples , modeling of the activity of these interneuron types would be necessary . In our model , we only have parvalbumin positive basket cells to represent the overall disorganized background inhibition that matches and balances the ongoing excitation . Hence , when we set up the ratio of pyramidal cells to interneuron , we choose a ratio that encompasses the overall excitatory to inhibitory ratio in CA1 , rather than the fine count of basket cells . We also model pyramidal cells as one uniform population even if it is known that they are quite selective in their specific connectivity , because we are interested in representing the overall oscillatory phenomena more than the specific cell-to-cell variability . The proportion of interneurons and pyramidal cells in the model is in agreement with CA1 anatomy [32] . We choose to model both neuron types with Exponential Integrate and Fire equations , because of its simplicity ( only two variables ) combined with a formalism that can represent explicitly spike-triggered adaptation and intrinsic cell resonances ( essential to give CA1 pyramidal cells their characteristic bursting spiking profile [34] ) . In looking for appropriate parameter values for the model cells types , we build on work by Gerstner [36] , who has classified parameter sets for the Adaptive Exponential IF model that match known spiking behaviors such as bursting ( for pyramidal cells ) and regular fast spiking ( for basket cells ) . We chose parameters that guaranteed a time scale ( see C/gL ratio ) for membrane voltage of about 20ms for both cell populations , the ability of pyramidal cells to burst when driven to fire ( the value of Vr is crucial for that ) , a theta resonance in pyramidal cells given by τw , and the regular spiking behavior of basket cells . Given that all cells of a population are modeled by the same equation , we introduce heterogeneity in the network using input currents and variability in synaptic strengths . Special attention was made to design an input term I ( t ) that is composed of different parts . The IDC term variability produces variability in the excitability level of each cell , therefore introducing a first level of heterogeneity in the network . The noise term βηt represents the in vivo state of the voltage in each cell , which is likely receiving a much higher barrage of synaptic inputs than the one provided by the network spiking activity . ηt is an Ornstein Uhlenbeck ( OU ) process , which can be thought of as a filtered white noise process . This kind of noise does not introduce slow frequency forcing , which could alter the network behavior , or too high frequency voltage fluctuations , which are known to not be present [75] . This kind of noise is also used in dynamic-clamp experiments to mimic in vivo state in hippocampal slice recordings [38] . The β coefficients establishing noise size were chosen so that hyperpolarized cells show a voltage fluctuation of about 2mV in size [38 , 75] . Isyn is the term representing synaptic current in the input . With the chosen formalism for synaptic equations , the strength of synaptic connections gj→n scales the maximum peak size of a post-synaptic potential . Every gj→n is selected from a Gaussian distribution to introduce heterogeneity . The reversal potential for synaptic currents are chosen in agreement with many published hippocampal models [34] , and the time scales of synaptic rise and decay are estimated from literature [34 , 39] , in particular taking advantage of the fact that EPSP on interneurons are faster than on principal cells [76] and IPSPs from basket cells are slower on pyramidal cells than on other basket cells [39] . Average synaptic strength values in the model induced post-synaptic potentials of less than 1mV and have been tuned to induce a balanced average of excitatory and inhibitory currents in pyramidal cells . Thus we applied a common approach of normalizing the average synaptic weight by total number of cells in each population [77] , hence taking into account the fact that the total incoming excitatory connections in a given cell are more than the inhibitory ones ( note the magnitude of g¯Int→Int compared to g¯Pyr→Int ) . The synaptic values we chose resulted in small post-synaptic potentials , where a single or few synchronous incoming spikes are not enough to cause a spike or suppress one in the post-synaptic cell . This condition implies the network is weakly coupled [78] as a dynamical system . The incoming current input Iinp ( t ) is bell-shaped , gradually rising and falling , between ton and toff . Incoming input from CA3 reaches a maximum value Imax different across populations , chosen to obtain a fraction of pyramidal cell spiking during ripples consistent with experimental data [24 , 43] and a firing frequency for interneurons about 120 Hz [23] . When a net current is used to represent the sum of a barrage of incoming post-synaptic currents , such current has to take into account the scaling that cell properties will operate on the post-synaptic currents . Since cell membranes have capacitive and resistive properties , there will be a time scaling by a time constant which is the cell time constant . Since we add the current input to the right-hand side of the dv/dt equation , this filtering is operated by solving for voltage in time . The other filtering that cells can operate is due to their size . Input resistance is defined as the ratio between difference in voltage induced by a step of current and the size of the current step . As such , it interacts with cell size: more current is required to change the voltage of a larger cell . Our model cells are point cell , that is , they do not have a radius . We have not defined the overall cell parameters based on a per-area scaling , but rather on their dominant time-scale properties . As a result , the intuitive choice of giving to both pyramidal cells and inhibitory interneurons the same size of current step ( because the same CA3 cells are connected to them ) would be overlooking the well-known fact that pyramidal cells are much bigger cells that parvalbumin positive interneurons . Rather than re-scaling the model to introduce size properties , we re-scale the incoming current . While the exact nature of the LFP components likely includes both synaptic activity and spikes [79] , we here focus on the CA1 pyramidal layer , where perisomatic interneurons are known to synapse , and are interested in a phenomena in which pyramidal cells spikes are known to not contribute significantly . To obtain an LFP estimate , the average synaptic current input across all pyramidal cells was calculated , and then rescaled by 1 mS to represent a potential , such that 100 pA of synaptic current produce a 100 μV LFP change . In the model , SWR were found when the filtered LFP ( 50–350 Hz ) exceeded a threshold of 5 standard deviations of the mean computed in one SWR-free second of activity . To estimate the duration of synchrony in a purely inhibitory network receiving a current step input of size k at time t = 1s ( Fig 5 ) , we considered histograms of spike probability constructed averaging 200 trials . The asymptotic value ( lim ) of the histogram was the average firing probability between 1 . 5 s and 2 s . The heights of peaks in the histograms were decreasing in time . The size of the first peak ( p1 ) defined a threshold 0 . 2* ( p1-lim ) . The first peak that was far from lim less than threshold marked the end of the transient .
|
Our memories are consolidated while we sleep through a bidirectional exchange of information between two brain areas called cortex and hippocampus . Neurons that were active in behavioral tasks reactivate again in both structures during sleep in a process of linking and modifying memories from the short term storage of the hippocampus to permanent storage in the neocortex . This process occurs mainly during short oscillatory hippocampal electrical events called sharp wave-ripples . We propose a novel mechanism of ripple generation consistent with a wide range of experimental data , to explain how hippocampal network properties shape ripple frequency and duration . Understanding the neuronal mechanism underlying ripples is crucial to explaining how the interaction between hippocampus and cortex during sleep enables memory consolidation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"action",
"potentials",
"medicine",
"and",
"health",
"sciences",
"nervous",
"system",
"sleep",
"membrane",
"potential",
"brain",
"electrophysiology",
"neuroscience",
"learning",
"and",
"memory",
"physiological",
"processes",
"ganglion",
"cells",
"cognition",
"memory",
"interneurons",
"animal",
"cells",
"memory",
"consolidation",
"cellular",
"neuroscience",
"hippocampus",
"cell",
"biology",
"pyramidal",
"cells",
"anatomy",
"synapses",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"cognitive",
"science",
"neurophysiology"
] |
2016
|
Hippocampal CA1 Ripples as Inhibitory Transients
|
Glioblastoma ( GB ) is a highly invasive and lethal brain tumor due to its universal recurrence . Although it has been suggested that the electroneutral Na+-K+-Cl− cotransporter 1 ( NKCC1 ) can play a role in glioma cell migration , the precise mechanism by which this ion transporter contributes to GB aggressiveness remains poorly understood . Here , we focused on the role of NKCC1 in the invasion of human primary glioma cells in vitro and in vivo . NKCC1 expression levels were significantly higher in GB and anaplastic astrocytoma tissues than in grade II glioma and normal cortex . Pharmacological inhibition and shRNA-mediated knockdown of NKCC1 expression led to decreased cell migration and invasion in vitro and in vivo . Surprisingly , knockdown of NKCC1 in glioma cells resulted in the formation of significantly larger focal adhesions and cell traction forces that were approximately 40% lower than control cells . Epidermal growth factor ( EGF ) , which promotes migration of glioma cells , increased the phosphorylation of NKCC1 through a PI3K-dependant mechanism . This finding is potentially related to WNK kinases . Taken together , our findings suggest that NKCC1 modulates migration of glioma cells by two distinct mechanisms: ( 1 ) through the regulation of focal adhesion dynamics and cell contractility and ( 2 ) through regulation of cell volume through ion transport . Due to the ubiquitous expression of NKCC1 in mammalian tissues , its regulation by WNK kinases may serve as new therapeutic targets for GB aggressiveness and can be exploited by other highly invasive neoplasms .
Glioblastoma ( GB ) is the most common malignant primary brain tumor . GBs are aggressive and display key features of invasion and infiltration of healthy brain tissue [1] . Due to its invasive nature , GB is not curable through surgical resection [2] , [3] . The surgical and medical treatment for patients with this disease has evolved in the last 20 years , however the prognosis remains dismal due to tumor recurrence [4] . Thus , understanding the mechanisms that GB cells utilize during migration and invasion into normal brain tissue is paramount in the development of novel , effective therapies . Volume regulation , cytoskeletal rearrangements , and adhesion dynamics are major determinants of cell migration and are essential processes in invasion [5] , [6] . Migration of mammalian cells is accompanied by volume changes . For instance , neutrophils [7] and dendritic cells [8] undergo cell volume increases when exposed to signals leading to migratory responses . Indeed , it has been hypothesized that inhibition of cell volume regulation impairs cell migration [9] , [10] . NKCC1 , a transporter that belongs to the SLC12A family of cation-chloride cotransporters , is a fundamental transporter utilized in the regulation of intracellular volume and in the accumulation of intracellular Cl− [11] , [12] . NKCC1 mediates the movement of Na+ , K+ , and Cl− ions across the plasma membrane using the energy stored in the Na+ gradient , generated by the Na+/K+ ATPase . Recent work supports the notion that intracellular volume regulation by NKCC1 [13] , [14] , as well as aquaporin 4 ( AQP4 ) [15] , may indeed promote glioma cell invasion . However , whether cell volume regulation is the only or primary mechanism mediating NKCC1 effects is unclear . It is equally unclear if NKCC1 is differentially regulated in invasive cells . In addition to cell volume regulation , ion transporters can participate in anchoring the cytoskeleton to the plasma membrane by binding to ezrin-radixin-moesin ( ERM ) proteins [16] , [17] . ERM proteins associate directly with actin and integral membrane proteins , which connect the cytoskeleton to the plasma membrane [18] . Anion exchangers ( AE ) 1 , 2 , and 3 , Na+/H+ exchanger 1 ( NHE1 ) , and a Na+/Ca++ exchanger are all able to act as cytoskeletal anchors by interacting with ERM proteins [19] . It has been shown that ERM proteins bind to clusters of positive amino acids in the juxtamembranous domain of NHE1 , CD44 , CD43 , and ICAM-2 [16] , [20] and that these interactions regulate cell migration and contractility , as well as focal adhesion turnover [21] , [22] . The interaction between ion transporters , as integral membrane proteins , and the cytoskeleton mediates the transduction of contractile forces generated from within the cell to the extracellular matrix and promotes migration . However , the mechanistic action of NKCC1 on cell contractility and focal adhesion dynamics in the context of GB cell migration and invasion are entirely unknown . Activation of NKCC1 transport activity requires phosphorylation of key threonine residues in the NKCC1 N-terminal domain [23] . Phosphorylation of NKCC1 is mediated by at least three members of a novel family of unusual kinases that lack a key lysine in their catalytic domain , the WNK kinases ( With No K-lysine ) [24] , [25] . These kinases have been implicated in the pathogenesis of hypertension and epilepsy [26] , [27] . Of these , WNK3 is the most abundantly expressed in the brain [28] . Interestingly , WNK1 is a substrate for Akt-mediated phosphorylation [29] . Hence , it is possible that Akt may regulate NKCC1 activity through the regulation of the WNK kinases . Intracellular signaling pathways , such as phosphoinositide 3-kinase ( PI3K ) -Akt , are frequently altered in GBs [30] . Akt is able to regulate various cellular functions through phosphorylation of a conserved substrate sequence , and altered regulation of this pathway can lead to aberrant cell behavior , such as increased proliferation and migration [31]–[33] . Importantly , intracellular signaling pathways of promigratory factors such as epidermal growth factor ( EGF ) [34]–[36] , and integrin signaling pathways converge on Akt , modulating cell processes such as cell cycle , apoptosis , and migration [37] . PI3K , the activator of Akt , is thought to be critical in mediating both chemotactic and random cell migration [38] . Therefore , the regulation of NKCC1 by the interaction between Akt signaling and WNK kinases may be important in determining the invasive properties of GB cells . To further our understanding of the role of NKCC1 in GB cell migration and invasion , we investigated ( 1 ) whether the expression of NKCC1 in human tumors correlates with tumor grade , ( 2 ) whether NKCC1 affects cell contractility and migration , ( 3 ) whether NKCC1 can have an effect on the interaction between the cells and the cells' adhesion substratum , and ( 4 ) whether a signaling mechanism involved in the regulation of NKCC1 by promigratory factors exists in GB cells . We found that NKCC1 expression indeed correlates with in vivo glioma aggressiveness and that the transporter activity modulates migration speed and invasiveness of cells derived from various human GBs . Furthermore , we show that NKCC1 expression affects GB cell traction forces , possibly by regulating focal adhesion dynamics . Moreover , the regulation of NKCC and KCC transport by WNK3 may determine the invasive behavior of GB cells . Additionally , we show evidence of NKCC1 phosphorylation regulation by Akt through WNK3 phosphorylation upon stimulation with a promigratory factor , EGF . This suggests an important link between the activation of WNK3 by Akt as well as changes in the activity of ion transport systems in glioma cells . Taken together , these findings strongly suggest that ion transport regulation might be integrated into the control of glioma cell invasiveness in a complex fashion that extends beyond regulation of cell volume and involves the interplay between cell adhesion and growth factor signaling . The understanding of these complex interactions may assist in the design of novel therapeutic strategies .
Prior data implicating NKCC1 in GB invasiveness were based on established , model GB cell lines , rather than primary cells or tissues . We therefore first evaluated and characterized whether primary cells isolated from human GB indeed supported the role of NKCC1 in invasiveness as suggested previously [13] . We assayed invasiveness using the transwell invasion assay in the presence or absence of the NKCC1 inhibitor bumetanide [39] . Inhibition of NKCC1 transport in various primary human glioma cells exposed to 25 and 50 µM of bumetanide led to a dose-dependent decrease in the number of invasive cells ( Figure 1A and Figure S1A ) . Significant inhibition of invasion was seen in GB cells tested at a concentration of 50 µM ( Figure S1B ) , a concentration at which bumetanide does not exhibit considerable non-specific effects on other cation-chloride transporters [40] , [41] . To further examine whether the effect of bumetanide on cell invasion is due to inhibition of NKCC1 , we performed stable knockdown of NKCC1 using lentiviral particles carrying NKCC1 shRNA . Knockdown of NKCC1 in GB cells ( NS561 , NS567 , NS501 , and NS318 ) was successfully established in 4 GB cell lines and the efficiency of knockdown was assessed by immunoblot of whole cell lysates of these GB cells ( Figure S1C ) . We confirmed that , as previously shown by Haas and colleagues [13] , knockdown of NKCC1 significantly reduced the invasiveness of all these cells ( Figures 1B , S1D ) . Taken together , these data suggest that NKCC1 may indeed play a role in invasiveness of primary GB cells , supporting prior results [13] obtained in non-primary cell cultures . Since NKCC and KCC transporters work in a concerted inverse manner to regulate intracellular volume and intracellular chloride concentration ( [Cl−]i ) [42] , we tested whether inhibition of KCC transport , an important Cl− extrusion mechanism , might mimic NKCC1 overexpression and lead to increased invasion . To test this hypothesis we performed transwell invasion experiments in the presence of DIOA ( R ( + ) -Butylindazone ) , a potent K+-Cl− transport inhibitor that has no effect on NKCC transport activity . Consistent with this hypothesis , inhibition of KCC transport with DIOA resulted in increased cell invasion . This effect was statistically significant in two of the four cell lines tested ( NS221 and NS318 ) ( Figure 1C ) . DIOA is a non-specific inhibitor of KCC co-transporters , which may be the cause of a heterogeneous effect on GB cells observed . To avoid this confounding result we induced the genetic knockdown of KCC4 ( Figure S1E ) , a KCC family member implicated in cervical and ovarian cancer invasiveness [43] . KCC4 had similar expression levels in the cell lines used for the experiments ( Figure 2D ) . Knockdown of KCC4 in NS318 showed a significant increase in the number of invading cells ( Figure 1D ) . These data suggest that KCC transport inhibition could lead to an increase in [Cl−]i promoting invasive behavior of GB cells . It is thought that GB tumor stem cells may be the core component of the invasive cell population [44] . Therefore , in addition to invasiveness of primary GB cells in vitro , we explored the role of NKCC1 in the invasion of primary brain tumor stem cells ( BTSC ) in vivo . Tumor area and area of invasion in the corpus callosum were then quantified to evaluate differences in tumor size and invasive ability of BTSCs carrying the control shRNA as well as BTSCs carrying NKCC1 shRNA . We found that tumors generated after the implantation of BTSCs with control shRNA were significantly smaller than tumors generated with the NKCC1 shRNA harboring BTSC line ( Figure 1E ) . Consistent with previous results by Haas and colleagues using commercial GB cell lines [13] , the invaded area in the corpus callosum of mice that were implanted with BTSCs carrying the control shRNA was significantly larger than that of mice implanted with BTSCs carrying NKCC1 shRNA ( Figure 1F ) . NKCC1 knockdown did not affect the proliferative potential of the BTSCs injected in vivo ( Figure S2 ) . These results suggest that NKCC1 may be an important determinant of primary and GB stem cell invasiveness , in congruence with prior suggestions based on commercial GB cell lines [13] . To evaluate the potential clinical importance of NKCC1 in glioma invasion in vivo , we characterized NKCC1 expression in a large array of glioma tissue samples using a tissue microarray ( TMA ) containing several tumors of different grades ranging from World Health Organization ( WHO ) Grade II to WHO Grade IV ( Table S1 ) . The results revealed that NKCC1 protein expression was significantly higher in GB and anaplastic astrocytoma ( AA ) tissue samples compared with expression in Grade II astrocytomas and normal brain ( Figure 2A and 2B ) . Epithelial tissues included in the TMA were used as positive controls ( intestinal mucosa and tissue from the distal collecting duct in the kidney ) ( Figure 2B ) . As a corollary to this analysis and a complement to the results in Figure 1 , we characterized the expression levels of NKCC1 protein in multiple primary human GB cells and found that all cell lines tested showed substantial expression of NKCC1 ( Figure 2C ) . The data obtained from this set of samples showed that NKCC1 protein expression indeed correlates with glioma grade , in that tissues from GB and AA expressed higher NKCC1 protein levels than low-grade astrocytomas and normal brain . This correlation between NKCC1 expression with glioma grade suggests that NKCC1 may contribute to the increased invasiveness of high-grade tumors . Our results so far strongly suggest that NKCC1 may indeed be an important determinant of GB cell invasion . While all prior analyses attempted to link the role of NKCC1 in cell migration to its role as a cell volume regulator [11] , [13] , [14] , [45] , we examined whether NKCC1 plays an essential role in the regulation of polarization of cell morphology and migration of GB cells . We were particularly interested in whether NKCC1 might affect cell migration and how this migratory behavior may depend on the mechanical cues mimicking the extracellular matrix components . In this study we employed nanoscale grooves to analyze the migratory behavior of glioma cells . Our substrate mimics ECM features , such as myelinated fiber tracts , upon which brain cancer cells have been shown to migrate [46] , [47] . This model offers the advantage of allowing biased cell migration along the nano-ridges of the textured surface that can be quantified in terms of cell speed and migration ( Figure S3A–C ) . We found a significant reduction in the cell migration speed of human primary GB cells stably transduced with NKCC1 shRNA ( Figure 3A–B ) . Similarly , a significant decrease in migration speed was observed when GB cells were treated with bumetanide ( Figure S3D ) . Migration directionality was quantified by measuring the ratio of cell movements parallel to the ridges on the pattern versus those that were perpendicular to the pattern . This metric tended to correlate with the speed of migration , showing significant decreases in directionality for cells expressing NKCC1 shRNA ( Figure 3C–D ) . Overall , GB cells stably transduced with NKCC1 shRNA displayed a lower speed of migration and showed more random migration as demonstrated by the decrease in directionality . The cell migration data indicated that NKCC1 can directly or indirectly affect cell motility , but the mechanism of how an ion transporter can be involved in this process is not immediately apparent . It is therefore of interest to note that at least some ion transporters have been reported to associate with the ERM complex to anchor actin to the plasma membrane , affecting cell migration [16] , [19] . The ERM complex proteins bind to clusters of positive amino acids such as lysine ( K ) and arginine ( R ) in proteins that are known to bind ERM proteins and to serve as anchors for the actin cytoskeleton such as CD44 , CD43 , and ICAM-2 [20] . Also , NHE1 , a Na+-H+ exchanger , acts as an anchor for the cytoskeleton in migrating cells , through the interaction with ERM proteins [16] . Based on these data , we studied the sequence of the juxtamembrane carboxy-terminus domain of human NKCC1 and found clusters of positively charged amino acids identical to those found in other ERM binding proteins . These clusters of positive amino acids are conserved in the human , mouse , and rat NKCC1 sequences ( Figure 4A ) . These amino acids may be important in the interaction between ERM proteins and NKCC1 and may be similar to other ERM-integral membrane protein binding [20] . To assess the possibility that NKCC1 may affect GB cell migration through a mechanism other than cell volume regulation , we compared the size of focal adhesions formed by NKCC1 knockdown cells and cells transduced with the control shRNA . Focal adhesions were stained with an antibody against vinculin and paxillin , cytoskeletal proteins that are part of focal adhesions that also regulate mechanical coupling of the cytoskeleton to the extracellular matrix ( ECM ) . We observed small , thin , and elongated focal adhesions primarily in the extending processes in control virus shRNA cells , whereas in NKCC1 shRNA cells , focal adhesions were much larger ( Figure 4B–C and Figure S4 ) , indicative of focal adhesion maturation [22] , [48] . The area of focal adhesions was significantly larger in NKCC1 shRNA cells when compared to control virus cells ( Figure 4D ) . The increased focal adhesion area was also seen when we used paclitaxel ( a drug that stabilizes microtubules dynamics and disrupts focal adhesion formation ) as a positive control for this experiment [49] . These results suggest that NKCC1 expression not only regulates cell volume but may also be important in modulating focal adhesion dynamics and maturation . Cells exert traction forces on their environment during migration and invasion in response to different mechanical and chemical cues in the extracellular matrix . These forces are applied through points of cell adhesion via focal adhesion-mediated integrin-ECM connections [21] , [22] . To evaluate whether the increase in size of focal adhesions after NKCC1 depletion had a functional effect on the generation of contractile forces by GB cells , we quantified cell traction forces exerted by adherent living GB cells ( control virus versus NKCC1 shRNA ) . We found that NKCC1-deficient cells exerted significantly lower cell traction forces than control virus cells ( Figure 5A–B ) . Compared to control virus cells , NKCC1 shRNA cells exhibited approximately a 40% decrease ( NS501 , 44% decrease; NS561 , 37% decrease; p<0 . 002 , nested ANOVA ) in net contractile moments , which is a scalar measure of cell contractile strength ( Figure 5C–D , Figure S5 ) . No within-group differences existed between both tested cell lines . To further support the interaction of NKCC1 and ERM proteins , we immunoprecipated endogenously expressed NKCC1 and probed the immunoprecipated lysate with an antibody against Ezrin . We found that endogenous Ezrin associates with immunoprecipitated NKCC1 . These results strongly suggest that in primary human GB cells , Ezrin is an NKCC1 binding partner . As expected , actin , a binding partner of Ezrin , also co-immunoprecipitated with NKCC1 ( Figure 6A ) . We also performed the reverse experiment where Ezrin was immunoprecipitated and then probed for NKCC1 on the immunblot . In multiple primary human GB cell lines , we found that after performing immunoprecipitation of Ezrin , NKCC1 was also pulled down ( Figure 6B ) . As expected , actin was also co-immunoprecipitated . To assess whether the association of NKCC1 and Ezrin is important for the generation of cell traction forces , we mutated the two clusters of basic amino acids found in the juxtamembranous domain of NKCC1 ( putative Ezrin-binding sites ) and measured the functional consequences in GB cells . We found that the net contractile movements of cells expressing the Ezrin-binding null NKCC1 were significantly lower than cells expressing wild type-NKCC1 ( Figure 6C ) . Furthermore , cells expressing the Ezrin-binding null-NKCC1 had a lower projected cell area than cells expressing wild-type NKCC1 ( Figure 6D ) . Thus , the absence of NKCC1 expression may lead to the formation of more mature focal adhesions . In turn , these mature focal adhersions may further enhance the adhesion of cells to the substratum , which decreaces the migration speed of NKCC1 shRNA cells , as observed above . Mature focal adhesions do not participate in the generation of contractile forces in migrating cells; rather , they participate in anchoring cells to the substrate [21] . On the other hand , nascent adhesions apply forces to the substratum to drive cell movement [21] . Hence , our results suggest that NKCC1 affects cell-ECM interactions by stimulating higher traction force generation and lower substratum adhesion , thus enhancing cell motility in a synergistic fashion . The pronounced effect of NKCC1 expression on focal adhesion formation and cell migration suggests the importance of its partial intracellular localization . We initially approached this issue by performing immunocytochemistry experiments on multiple GB cells . We found that all GB cells had a polarized subcellular expression of NKCC1 . In cells that appeared to have a more stationary phenotype with multiple projections , the expression of NKCC1 was primarily correlated with these projections ( Figure 7A and Figure S6 ) . Specifically , NKCC1 was localized either to the apparent leading edge of a moving cell or to its rear , frequently in a mutually exclusive pattern ( Figure 7B and Figure S6 ) . Expression of NKCC1-EGFP fusion protein in GB cells supported that NKCC1-EGFP expression was mainly localized to the plasma membrane of the extending processes confirming the results obtained by immunofluorescence ( Figure 7B , Figure S7 , and Video S1 and Video S2 ) . In cells spreading on nano-structured substrata , NKCC1-EGFP localization oscillated between the two transiently existing edges , before a prominent single edge was formed . Furthermore , when using immunocytochemistry , we examined the sub-cellular localization of WNK3 , a serine/threonine kinase that regulates the transport activity of NKCC1 through phosphorylation [26] , [27] , [50] . We observed partial co-localization of WNK3 immunoreactivity with NKCC1 immunoreactivity in the edges of extending processes ( Figure 7A and Figure S6 ) . These findings suggest that the cellular localization of NKCC1 is spatially heterogeneous during GB cell migration . Although the localization patterns were diverse in different cell states , the overall pattern that emerged from this analysis was that NKCC1 is associated with extending processes of the cell . This finding correlates with the suggestion that NKCC1 is important in the formation of new focal adhesions and controlling existing focal adhesions and active cytoskeletal components . The aforementioned data suggest that NKCC1 transport activity is important for glioma cell migration and invasion , at least in part through direct regulation of the cytoskeletal and ECM-cell adhesion dependent processes . NKCC1 transport activity is known to be regulated through phosphorylation and de-phosphorylation events mediated by members of the novel serine/threonine kinase family WNKs [26] . NKCC transport is activated by stimulation with EGF in corneal epithelial cells [51] . It is well established that EGF promotes astrocytic [34] and glioma cell migration [36] , [52] , [53] . Thus , we examined the effect of EGF on the phosphorylation of NKCC1 as an indication of NKCC1-activation using an NKCC1 phospho-specific antibody [23] . After stimulating glioma cells with EGF , NKCC1 phosphorylation increased in a time-dependent and dose-dependent manner in NS318 and NS567 cells ( Figure 8A ) . To gain insight into the regulation of phosphorylation of NKCC1 in an unbiased cellular system , we stimulated HEK-293 cells with EGF in the presence or absence of wortmannin ( WM ) , a PI3K inhibitor . After exposure of HEK-293 cells to EGF , NKCC1 phosphorylation increased significantly . However , in the presence of WM , EGF-induced NKCC1 phosphorylation was blocked ( Figure 8B ) . These findings together demonstrate that the EGF-induced increase in phosphorylation of NKCC1 requires activation of the PI3K-Akt pathway . The activity of cation-chloride cotransporters is regulated in a coordinated manner by the novel family of serine-threonine kinases WNK [50] . WNK3 promotes phosphorylation and activation of NKCC1 transporters while promoting phosphorylation and inactivation of KCC transporters [50] . It has also been shown that WNK1 , another member of the WNK family , is phosphorylated and activated by Akt ( protein kinase B ) [29] , [54] . Therefore , we decided to test if Akt phosphorylates WNK3 after stimulation with EGF . We immunoprecipitated total WNK3 from HEK-293 cells exposed to serum-free media , and we stimulated with EGF , or EGF in the presence of the PI3K inhibitor WM . Samples were immunoblotted with an antibody that recognizes phosphorylated Akt substrates ( αPAS antibody ) ( Figure 8C ) . Data obtained from this experiment showed a basal phosphorylation level of WNK3 , which increases with exposure to EGF . This increase in phosphorylation is repressed by inhibition of PI3K with WM , suggesting that Akt phosphorylates WNK3 . The same samples were immunoblotted against phosphorylated Akt , and as expected , we found that WM also inhibited the phosphorylation of Akt ( Figure 8D ) . In silico analysis of the protein sequence of WNK3 revealed that two putative Akt phosphorylation motifs are present and are conserved in available WNK3 protein sequences of the human and rat , as previously found in WNK1 ( Figure 8E ) [54] . These findings indicate that NKCC1 may be activated by factors that stimulate migration of astrocytic or glioma cells , such as EGF via kinases of the WNK family , a family of kinases that have been shown to regulate the transport activity of multiple members of the SLC12A family of transporters .
The nearly universal recurrence of GB after surgical resection is largely due to invasion of glioma cells into healthy brain tissue and presents a major impediment in the improvement of GB patient survival . In this study , we tested the hypothesis that NKCC1 expression and transport activity are crucial elements of cell migration and invasion in primary human GB cell lines . Here we provide evidence for the participation of NKCC1 in the migration and invasion of primary human glioma cells , highlighting its possible role in anchoring the actin cytoskeleton to the plasma membrane . Additionally , through this anchoring , NKCC1 mediates transduction of the cellular contractile forces to focal adhesions that interact with the extracellular matrix . These in vitro findings were confirmed functionally in an in vivo model using primary human BTSCs , where their invasion was decreased significantly after NKCC1 knockdown . Given these findings , NKCC1 inhibition could potentially be used in the clinic to improve glioblastoma treatment , given that Bumetanide ( a commonly used diuretic , FDA-approved ) decreases the invasive potential of glioma cells in vivo [13] . This could potentially improve surgical resection of the tumor mass , as tumor cells lacking NKCC1 activity would form less invasive tumors . Our work shows that NKCC1 protein expression in multiple glioma samples is higher in high-grade gliomas such as GB and anaplastic astrocytomas . Inhibition of NKCC1 transport pharmacologically , as well as genetic inhibition of NKCC1 expression , decreases invasion of multiple primary human GB cell lines . These results are in accordance with previous findings using commercial human glioma cell lines [13] . Our data further indicate that migration speed of GB cell lines in a 2-D nanopatterned substrate is decreased by pharmacological inhibition and by shRNA-based silencing of NKCC1 expression . Interestingly , pharmacological and genetic inhibition of the K+-Cl− cotransporters leads to a more invasive behavior of GB cells in vitro . Moreover , our results suggest that NKCC1 may affect the morphology of focal adhesions , perhaps due to a putative ERM binding motif in the cytoplasmic domain of NKCC1 . We also found that NKCC1 is located at the extending processes of GB cells and that NKCC1 polarization may precede migration towards the direction of this pole . Furthermore , exposure of GB cells to EGF , a factor that promotes migration and invasion of normal and tumor cells [34]–[36] , [55] , [56] , induces phosphorylation ( activation ) of NKCC1 through PI3K-Akt-WNK3 pathway . The concerted action of local anchoring of the actin microfilaments to the plasma membrane and volume regulation may be important for the polarization of cells during migration . Coupling the actin cytoskeleton to the plasma membrane is essential for the regulation of cell morphology and migration [6] . Our immunocytochemistry and live cell imaging experiments using a GFP-NKCC1 fusion protein demonstrate that NKCC1 is localized to the extending processes of migrating GB cells . During the migratory process , cells acquire a polarized morphology where actin , integrin receptors , and ion transporters among other proteins , become asymmetrically distributed in the cell . Some examples of ion transporters that show a polarized localization to the leading edge of the cell include NHE1 and AE2 [9] , [16] , [17] , which also have K+ channels that are polarized to the rear end of the cell [57] . Ezrin-radixin-moesin ( ERM ) proteins bind actin filaments and anchor them to integral plasma membrane proteins . Some of these integral membrane proteins include NHE1 , CD44 , and intercellular cell adhesion molecule 2 , which have ERM binding motifs ( ICAM-2 ) [16] , [18] . The ERM binding motif consists of clusters of positive amino acids , such as lysine and arginine residues in juxtamembranous intracytoplasmic domains of these proteins . By analyzing the peptide sequence of NKCC1 , we found clusters of lysine and arginine residues in the N-terminal cytoplasmic domain that are conserved across mammalian species , which may bind ERM proteins . Indeed , we found that NKCC1 is able to bind to Ezrin and actin with our co-immunoprecipitation assay . The generation of advancing membrane protrusions is necessary for migrating cells to achieve cell translocation . ERM protein binding to NHE1 is necessary during migration to promote extension of advancing processes to anchor the cytoskeleton [16] , [17] . Our results show that NKCC1 is polarized to the extending processes of migrating glioma cells; therefore , it is likely that NKCC1 is necessary during migration to anchor the cytoskeleton , aiding in the extension of lamellipodia , and mediate local volume changes at the same time [14] . NKCC1 expression may be an important determinant of the response of migrating cells to external physical cues . The extracellular matrix surrounding cells presents topographical features ranging from nanometers ( nm ) to microns ( µm ) and affects cell behavior . For example , collagen fibrils form with diameters from 20–200 nm and influence cell polarity and migration through “contact guidance” [58]–[65] . Recent studies have employed more intricate substrates presenting nanoscale features ( e . g . , grooves , ridges , bumps , and pillars ) to more closely model the cellular microenvironment [66]–[68] . In this study , we have employed nanoscale features mimicking the ECM found in the brain . These features include myelinated fiber tracts , upon which brain cancer cells have been shown to migrate [47] . It is known that cell migration is governed by many molecular processes , including attachment to the cell substrate . Our nanopattern provides a quasi-3-D platform that can examine these interactions with ECM by examining speed , direction , and morphologies of migrating cells . While cells move in 2-D , they respond to topographical cues from the substrate presented in 3-D [46] , [69] . By examining cell migration after genetic changes to NKCC1 expression , we sought changes that might inhibit the overall motility of glioma cells . Our observed changes in migratory behavior upon simulated ECM suggests that these cells will be less migratory and invasive , eventually leading to improved medical outcomes . These changes in migratory behavior were further supported by our experiments employing more classical techniques ( e . g . , transwell ) . It has been shown that formation of actin stress fibers precedes the formation of nascent focal adhesions in the lamellipodium of fibroblasts [70] . However , stress fiber formation depends on the anchorage of actin bundles to the plasma membrane , which has been shown in the interaction of NHE1 with ERM proteins [16] . Nascent adhesions are present at the front of the cell and exert traction forces that lead to cell repositioning [21] . As focal adhesions increase in size , they mature and the traction forces that they exert decrease considerably [21] . In migration studies of cells that do not express focal adhesion kinase ( FAK ) , a major regulator of focal adhesion turnover , it was shown that these cells possess larger focal adhesions and display lower cell spreading [71] , [72] . In fact , these observations were also seen in GB cells when NKCC1 is knocked down; NKCC1 knockdown cells display larger focal adhesions and smaller projected cell area than control shRNA cells . These changes in focal adhesion size were accompanied by a decrease in the generation of contractile forces by GB cells . These findings suggest that the localized distribution of NKCC1 to the extending processes plays a role in the modulation of focal adhesion turnover and generation of nascent focal adhesion to maintain cell contractility and traction for efficient migration . Cell volume changes are expected in migrating cells since alterations in shape during extension and retraction occur throughout migration . Multiple ion transport mechanisms are responsible for regulating and maintaining cellular volume in response to changes in extracellular osmolarity and during cell migration [5]; in addition , ion gradients and local volume changes have been described in migrating cells [73] , [74] . These mechanisms include ion transporters such as NKCC1 , NHE1 , KCC transporters , and also ion channels . For instance , neutrophils undergo an increase in intracellular volume in response to the chemotactic factor N-formylmethionyl-leucyl-phenylalanine; this increase in cell volume and increased migration is blunted by NHE1 transport inhibitors and by exposure to hyperosmolar solutions , suggesting that NHE1-mediated volume increase is necessary for neutrophil migration [7] , [75] . Na+-K+-Cl− transport inhibition has also shown to decrease Madin-Darby canine kidney cell migration [74] . Our results are in accordance with the findings discussed above where inhibition of NKCC transport decreases migration . We also show that NKCC1 knockdown decreases GB cell migration , confirming the effects of pharmacologic inhibitors . Pharmacological inhibitors and shRNA-based approaches may have off-target effects , but the fact that the effect of both on GB cell migration is the same confirms that the results seen by manipulating NKCC1 expression/transport are consistent . EGFR activation promotes migration of normal neuroblasts , astrocytes , and glioma cells [34] , [53] , [76]–[78] . EGF signaling affects migration through diverse mechanisms , such as actin polimerization [79] , [80] , focal adhesion kinase regulation [81]–[83] , and matrix metalloproteinase expression [84] . EGF mediates its effects on cell migration and proliferation through activation of its receptor-tyrosine kinase and the various downstream signaling pathways , which include the PI3K-Akt pathway . NKCC1 phosphorylation by WNK3 after activation of the PI3K-Akt pathway supports the hypothesis that NKCC1 activity is necessary for GB cell migration . Furthermore , WNK3 activation after EGF stimulation suggests that phosphorylation and activation of NKCC1 and phosphorylation and inhibition of KCC transporters may result in GB cell migration . Therefore it seems this balance between these opposing transport activities is important in the determination of GB cell invasion . The PI3K-Akt signaling pathway , among many other cell functions , is central in the control of cell motility and polarization . PI3K is activated by receptor tyrosine kinases and Ras [85] . It modulates these functions by bringing diverse proteins that are able to bind phosphatidyl-inositol triphosphate ( PIP3 ) close to the membrane . A notable example of the proteins that are recruited to the membrane is Akt , which is activated after binding to PIP3 and phosphorylated by 3′-phosphoinositide-dependent kinase 1 ( PDK1 ) ; AKT is recruited in a polarized manner to the leading edge of the migrating cell membrane [86] . It is well known that PI3K activation mediates cytoskeletal rearrangements and cell polarization through the action of the guanine nucleotide exchange factors ( GEFs ) [87] , [88] . Moreover , PI3K modulates actin polymerization and membrane insertion at the leading edge of a cell by regulating the activity of Arf [89] . Its activation also promotes cell polarization through Rac regulation [90] . It is still necessary to assess if the cytoskeletal rearrangements and cell polarization mediated by PI3K are important in the generation of the partial distribution of NKCC1 to the extending processes of GB cells . WNK3 regulates ion transport through phosphorylation: it phosphorylates and activates NKCC1 and phosphorylates and inhibits transport of the KCC transporters in a reciprocal manner [27] , [50] . Furthermore , Haas et al . have shown that following a hyperosmotically induced decrease in cell volume , WNK3 may regulate NKCC1 . Also , the reduced expression of WNK3 by shRNA diminished the ability of glioma cells to migrate in vitro [91] . Our results show that inhibition of NKCC and KCC transport result in opposite effects in GB cell migration . Similar to WNK1 , our immunoprecipitation experiments show that EGF induces phosphorylation of WNK3 through Akt [29] , [54] . When phosphorylated , WNK3 then phosphorylates NKCC1 and possibly KCC transporters . This mechanism is similar to the mechanism proposed for increased excitability of neurons , where WNK3 signaling is impaired , resulting in GABA-mediated excitation of neurons and seizure activity [27] . Therefore , it is conceivable that WNK3 activation results in activation of NKCC1 and inhibition of KCC transport , causing increased migration of GB cells . In this study , we show that NKCC1 transport expression and activity are necessary for GB cells to migrate and invade . The mechanism affecting cell contractility that we report in this article may be independent from regulation of volume changes and may be due to regulation of focal adhesion formation and turnover . We also show that EGF regulates NKCC1 phosphorylation through an Akt-WNK3 pathway , linking the PI3K-Akt pathway to the WNK3 kinase . This suggests that WNK3 may have a role in determining GB cell migratory properties . Furthermore , given that NKCC1 is ubiquitously expressed , it is possible that it plays a very similar role in physiological migration such as inflammatory cell diapedesis or neural precursor migration during development , as well as in the process of metastasis of other highly aggressive cancers .
Patient samples of glioma tissues were obtained at the Johns Hopkins Hospital under the approval of the Institutional Review Board ( IRB ) . All human brain tumor cell lines were derived from intraoperative tissue samples from patients treated surgically for newly diagnosed glioblastoma multiforme without prior treatment as listed in Table S2 . Differentiation potential of cell lines for an in vivo experiment was evaluated by immunohistochemistry against GFAP , TuJ1 , and NG2 ( Figure S9 ) . Detailed culture methodology has been previously described [92] , [93] . VSV-G pseudotyped virus was produced by co-transfecting 293T cells with a shRNA transducing vector and two packaging vectors: psPAX2 and pMD2 . G . The shRNA sequence used was 5′-TAG TGC TCT CTA CAT GGC ATG GTT AGA AGC TCT ATC TAA GGA CCT ACC ACC AAT CCT C-3′ . Seventy-two hours after transduction , cells were cultured in the presence of puromycin for selection of cells expressing the shRNA . Knockdown was assessed by quantitative PCR ( Figure S8 ) and immunoblot ( inset in Figure 1F and Figure S1C ) . The sequence of human NKCC1 ( SLC12A2 , accession number NM001046 ) was amplified by PCR using gene-specific primers . The sequence of the primers employed is as follows: sense , 5′- GCG TGC TGC CGG AGA CGT CC-3′; antisense , 5′- AGT CAC CAT TCG CCA TTG TGA TGT T-3′ . The resulting PCR product was cloned into pCR-XL-TOPO ( Invitrogen ) . The cloned sequence was verified in its entirety to confirm the absence of mutations . The EGFP fusion protein was made by cloning the NKCC1 open reading frame into pcDNA3-EGFP using standard cloning procedures . All other procedures are listed in Supplemental Experimental Procedures ( Text S1 ) . Total RNA was extracted from primary glioma cell lines using the RNAeasy kit ( Qiagen ) and reverse transcribed using the SuperScript III First-Strand Synthesis System for RT-PCR ( Invitrogen ) . The target cDNAs were analyzed using SYBR Green PCR master mix ( Applied Biosystems ) in a 7300 Real-Time PCR system ( Applied Biosystems ) . For relative quantification , the results obtained were compared to the levels of target mRNA expression present in the control cell line and normalized for GAPDH expression . Primers are listed in Supplemental Experimental Procedures ( Text S1 ) . NKCC , WNK3 , Akt , Ezrin , and actin were detected using rabbit and mouse primary antibodies . Detection was done with the appropriate horseradish-peroxidase conjugated secondary antibodies and using the enhanced chemiluminescence reagent ( GE Healthcare Life Sciences ) . Antibodies are listed in Supplemental Experimental Procedures ( Text S1 ) . Cell lysates ( 150 µg of protein ) were incubated with anti-NKCC antibody ( T4 antibody , 1 µg; DSHB ) and anti-Ezrin ( cell signaling cat: 3145 , 1∶100 ) overnight at 4 °C on a shaking platform . Indirect immunoprecipitation was done with protein G magnetic beads ( Millipore ) . Proteins were then eluted and denatured in LDS protein loading buffer ( Invitrogen ) . Fifty thousand cells were plated in the top chamber of a matrigel-coated membrane ( 24-well insert; pore size , 8 mm; BD Biosciences ) . Cells were plated in medium containing 0 . 5% of serum , whereas medium with 2% serum was used as a chemo-attractant in the lower chamber . After 48 h cells that invaded were stained and counted for comparison . Migration of glioma cells was quantified using a novel directional migration assay using nano-ridges/grooves constructed of transparent poly ( urethane acrylate ) ( PUA ) , and fabricated using UV-assisted capillary lithography ( see Figure S3A–C ) [94] . Nanopattern surfaces were coated with laminin ( 3 µg/cm2 ) . Cell migration was quantified using timelapse microscopy ( Video S3 ) . Long-term observation was done on a motorized inverted microscope ( Olympus IX81 ) equipped with a Cascade 512B II CCD camera and temperature and gas controlling environmental chamber . Phase-contrast and epi-fluorescent cell images were automatically recorded under 10× objective ( NA = 0 . 30 ) using the Slidebook 4 . 1 ( Intelligent Imaging Innovations , Denver , CO ) for 15 h at 10–20-min intervals . A custom-made MATLAB script was used to identify cell boundaries from phase-contrasted images and to measure cell centroid positions . Average individual cell speed was calculated from individual cell trajectories and durations of the image acquisition . Mean squared displacements at various time intervals were calculated using a previously published method [95] . The spindle shape factor was defined as the ratio of the length of maximum cell width ( maximal axis ) to the minimum value of the cell width in the direction perpendicular to maximum axis , regardless of the orientation with respect to nanogrooves . For each condition , over 60 cells were quantified in total . For quantitative analysis of cell orientation , cells were fixed and stained for F-actin with phalloidin . The orientation angle of polarized cell was determined by measuring the acute angle between the major axis of the cell and the direction of grooves . More than 100 cells for each group were used to construct the polarization angle distributions with range −90° and 90° . A summary of all the migration assays used is presented in Table S3 . The contractile stress arising at the interface between an adherent cell and its substratum was measured with traction microscopy [96] . For each cell analyzed , the traction field was computed using Fourier transform traction cytometry as described previously . The computed traction field was used to obtain the net contractile moment , which is a scalar measure of the cell's contractile strength ( Figure S5 ) [97] . Cells were fixed in 4% paraformaldehyde in phosphate-buffered saline ( pH 7 . 4 ) for 1 h and blocked with 10% normal donkey serum in PBS for 1 h . Subsequently , fixed cells were incubated with primary antibody at 4 °C overnight . The preparation was then incubated with Alexa Fluor-conjugated secondary antibodies ( Invitrogen ) and mounted using Aquamount ( VWR ) . All antibodies and their dilutions are listed in Supplemental Experimental Procedures ( Text S1 ) . All animal protocols were approved by the Johns Hopkins Animal Care and Use Committee . In vivo invasion and tumorigenesis of cells expressing NKCC1 shRNA were assessed in 4- to 6-wk-old male mice ( nude/athymic mice , NCI ) using our brain tumor model as previously described [98] . Mice were sacrificed 8 wk after injection . Brains were fixed using transcardiac perfusion , postfixed overnight at 4 °C in 4% formalin , embedded in OCT compound ( Tissue-Tek ) , and frozen , sectioned , and stained with an antibody against human nestin ( 1∶500 , MAB5326 Millipore ) . Stained cryosections were used to calculate tumor size and invasiveness by computer-based morphometrics using Image J . Please refer to Text S1 for detailed description of the intracranial injection of GB BTSCs . Primary human GB cells expressing the control shRNA and NKCC1 shRNA were treated with 10 µM 5-ethynyl-20-deoxyuridine ( EdU ) . Cells were harvested for detection of EdU incorporation using Click-iT EdU Flow Cytometry Assay Kits ( Invitrogen , Cat . No . C35002 ) following the manufacturer's instructions . The percentage of cells that incorporated EdU was measured using flow cytometric detection of EdU . Data were analyzed using Kaluza software ( Beckman Coulter ) . A tissue microarray was designed and built according to previously established methods [99] . Cores were taken from each tumor mass or control tissue ( see Table S1 ) . The tissue that was included in the cores of the microarray was representative of the tissue blocks from where the cores were obtained . Analysis and correction for cell number was done using the FRIDA software ( free web-based tissue microarray analysis software ) . Unless otherwise noted , data are presented as mean ± standard error of the mean . A t test was used to compare two groups; one-way analysis of variance ( ANOVA ) was used in multiple group comparisons with Bonferroni's post hoc test . Mann-Whitney rank-sum test was used to evaluate the statistical significance in quantification of spindle shape factor where indicated . In order to satisfy the distributional assumptions associated with the ANOVA , cell traction force data were first converted to log scale prior to analyses . For the comparisons between treatments , we used a nested ANOVA . All analyses were performed in Sigma Plot 9 . 0 ( Systat Software Inc . , San Jose , CA ) SAS Version 9 . 2 ( SAS Institute , Cary , NC ) , and a two-sided p value less than 0 . 05 was considered significant .
|
Treatment of many cancers has been hampered by the invasive ability of tumor cells . A notable example is brain cancer , which is incurable due to its invasiveness and resulting high tumor recurrence after surgical resection . Here , we analyze further the function of NKCC1 , an ion transporter that is known to regulate cell volume and intracellular chloride concentration , and to play an important role in brain tumor cell invasion . Our findings suggest that in addition to its conventional function as an ion transporter , NKCC1 may also interact with the cytoskeleton and affect brain tumor cell migration by acting as an anchor that transduces contractile forces from the plasma membrane to the extracellular matrix en route to cell migration . Moreover , we show that regulation of NKCC1 by a family of unconventional enzymes , the WNK kinases , is an important factor that affects the activity of NKCC1 and may determine the invasive ability of brain tumor cells . We postulate that NKCC1 has multiple functions in brain tumor cell migration and that together with its regulatory enzymes may be therapeutic targets in the treatment of brain tumors or other types of cancer , given the wide expression of these proteins throughout the body .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"oncology",
"medicine",
"neurological",
"disorders",
"neurology",
"surgery",
"biology",
"molecular",
"cell",
"biology"
] |
2012
|
Regulation of Brain Tumor Dispersal by NKCC1 Through a Novel Role in Focal Adhesion Regulation
|
Self-organized criticality refers to the spontaneous emergence of self-similar dynamics in complex systems poised between order and randomness . The presence of self-organized critical dynamics in the brain is theoretically appealing and is supported by recent neurophysiological studies . Despite this , the neurobiological determinants of these dynamics have not been previously sought . Here , we systematically examined the influence of such determinants in hierarchically modular networks of leaky integrate-and-fire neurons with spike-timing-dependent synaptic plasticity and axonal conduction delays . We characterized emergent dynamics in our networks by distributions of active neuronal ensemble modules ( neuronal avalanches ) and rigorously assessed these distributions for power-law scaling . We found that spike-timing-dependent synaptic plasticity enabled a rapid phase transition from random subcritical dynamics to ordered supercritical dynamics . Importantly , modular connectivity and low wiring cost broadened this transition , and enabled a regime indicative of self-organized criticality . The regime only occurred when modular connectivity , low wiring cost and synaptic plasticity were simultaneously present , and the regime was most evident when between-module connection density scaled as a power-law . The regime was robust to variations in other neurobiologically relevant parameters and favored systems with low external drive and strong internal interactions . Increases in system size and connectivity facilitated internal interactions , permitting reductions in external drive and facilitating convergence of postsynaptic-response magnitude and synaptic-plasticity learning rate parameter values towards neurobiologically realistic levels . We hence infer a novel association between self-organized critical neuronal dynamics and several neurobiologically realistic features of structural connectivity . The central role of these features in our model may reflect their importance for neuronal information processing .
Self-organized criticality is increasingly postulated to underlie the organization of brain activity [1]–[2] . The notion of self-organized criticality describes an unsupervised emergence of critical dynamics in complex systems dominated by internal interactions [3]–[4] . Critical dynamics emerge at the transition between randomness ( subcritical dynamics ) and order ( supercritical dynamics ) , and are characterized by self-similar ( power-law-distributed ) spatial and temporal properties of system events ( e . g . neural activations ) . The occurrence of these dynamics in the brain is theoretically appealing and is increasingly empirically supported . Theoretically , and increasingly empirically , critical dynamics are associated with optimized information transmission and storage [5]–[8] , maximized dynamic range [9]–[10] and successful learning [11] . Empirically , multielectrode array recordings of spontaneous activity from organotypic cortical slice cultures [5]–[6] and dissociated cortical neuron cultures [12]–[13] show power-law scaling of distributed “avalanche” activity of neuronal ensembles . Multielectrode array recordings of spontaneous cortical activity in the awake rhesus monkey also show power-law scaling of avalanches [14] , suggesting that these dynamics are not confined to in vitro preparations . The temporal and spatial statistics of EEG , ECoG , MEG and fMRI signals likewise show power-law scaling [15]–[18] , although the relationship of these large-scale brain signals to avalanches of neuronal ensembles may not be straightforward . Brain dynamics are thought to be strongly influenced by neuroanatomical connectivity [19]–[22] . Consequently , self-organized critical brain dynamics may be influenced by properties of neuroanatomical organization , such as hierarchical modularity , small-worldness and economical wiring [23]–[26] . Hierarchical modularity is a self-similar organization in which functionally specialized neural clusters ( e . g . cortical lobes ) contain smaller and more specialized neural clusters ( e . g . cortical nuclei , cortical columns ) at multiple spatial scales . Small-worldness is an organization which combines modularity and robust between-module connectivity . Economical wiring is an organization which contains predominantly short connections . The presence of an intuitive association between self-similar brain structure ( i . e . hierarchical modularity ) and self-similar brain dynamics ( i . e . self-organized criticality ) , has not been previously examined . The relationship between brain structure and dynamics is reciprocal: while the structure strongly constrains the dynamics , the dynamics continuously modify the structure through mechanisms such as activity-dependent synaptic depression [27] and spike-timing-dependent plasticity [28]–[29] . We previously showed that this reciprocal relationship is associated with an unsupervised emergence of modular small-world structural connectivity , in a large-scale model of spontaneous brain activity [30] . We now ask whether realistic structural organization is associated with the emergence of self-organized critical dynamics . A number of modeling studies recently reported self-organized critical avalanche dynamics in neuronal networks with nontrivial topology and activity-dependent plasticity [31]–[33] . These studies focused on conceptual features of network organization and plasticity , and hence omitted neurobiologically realistic features such as membrane leakage , axonal delays and spike-timing-dependent plasticity . Other studies are increasingly beginning to examine these relationships in more realistic networks [34]–[35] . Most studies however , remain constrained by assessment of power-law distributions with unreliable linear least-squares-based methods [36] . In contrast , we aim to systematically and rigorously examine the relationship between anatomical connectivity , synaptic plasticity and self-organized criticality , in a realistic network model of neuronal activity . To this end , we extend a recent model of nonperiodic synchronization in networks of leaky integrate-and-fire neurons [37] to incorporate large , sparse , hierarchical modular connectivity , spike-timing-dependent plasticity and other neurobiologically realistic features such as axonal conduction delays and neuronal inhibition . We hypothesize that the neurobiologically realistic features of our model will facilitate the emergence of self-organized critical dynamics .
The studied leaky integrate-and-fire neuron evolves according towhere is the membrane potential , is the membrane capacitance , is the leakage conductance , is the resting potential and and are the external current and synaptic current , respectively . When exceeds a constant threshold , the neuron is said to spike and is reset to the value for an absolute refractory period . The external current maintains a constant low level of background neuronal activity , while synaptic currents couple anatomically connected neurons . In the model , we set and . We set for clarity , but any other value ( e . g . ) results in equivalent dynamics , as long as the above relationship between , and holds . We discuss these and other aspects of the integration scheme in the Supplementary Information ( Text S1 ) . For a postsynaptic neuron , we modeled synaptic currents with decaying exponentials , where the outer sum is over all presynaptic neighbors of , the inner sum is over all previous spike times of each presynaptic neighbor , is the synaptic weight from to , and are the slow and fast decay constants , and is a magnitude parameter . Synaptic coupling incorporated axonal delays , set to uniformly distributed random integers between and . These values are in the range of empirically estimated axonal delays [38] . For computational simplicity we used the same distribution of axonal delays for all hierarchical levels . We note that long-range cortical connections are often more thickly myelinated than short-range connections so there is no simple relationship between inter-level distance and axonal delay . Synaptic weights changed at every spike of a neuron incident to the synapse , according to a spike-timing-dependent plasticity ( STDP ) rule ( Figure 1b ) . The STDP rule potentiates when the postsynaptic neuron spikes shortly after the presynaptic neuron , and depresses when neuron spikes shortly before neuron . More specifically , when or spike , changes as , withwhere and are the latest spike times of and , and are time constants and and are weight dependence functions , The weight dependence functions keep all weights between and , and rescale weight changes by the weight constants and , and by the rate constant . The above functions enable soft weight bounds , or multiplicative weight dependence . Alternative functions , where is the Heaveside step function , enable hard weight bounds , or additive weight dependence . The choice between soft and hard weight bounds has important implications for synaptic weight distributions ( Figure 1c–e ) . The unimodal distribution associated with soft weight bounds has more experimental support [39] , although both hard and soft weight bounds are extensively used in computational studies . We used soft bounds in most simulations , but also explored the robustness of our results to the presence of hard bounds . Parameter values of the model were adapted from the Thivierge and Cisek [37] study and are shown in Table 1 . In the present study , we find that the postsynaptic-response magnitude and STDP learning rate parameters facilitate important internal interactions in the network . We show that high values of these parameters are required to compensate for the relatively small number of neuronal synapses in our networks . We also show that these values may be substantially reduced in larger networks with greater numbers of synapses . Each network comprised neurons , subdivided into modules . Each module comprised neurons , of which neurons were inhibitory and excitatory . Inhibitory neurons only formed synaptic connections with all excitatory within-module neurons . On the other hand , excitatory neurons could potentially form synaptic connections with excitatory or inhibitory neurons in all modules . Initially , excitatory neurons only formed synaptic connections with all other within-module neurons . Subsequently , excitatory synapses were probabilistically rewired within seven hierarchical levels ( Figure 1a ) . The density of intermodular connections , , within each level , was set using power-law ( ) , exponential ( ) or linear ( ) scaling functions , with , and determining density drop-off rates ( Figure 2a ) . Synapses were rewired in a way that preserved the total number of synapses per neuron [40] but not connection reciprocity . For each network , rewiring occurred progressively from the outermost to the innermost hierarchical level . The location of synapses in each network was kept fixed during simulations . The wiring cost associated with each scaling function was computed by estimating the number of synapses in each hierarchical level for that function , equating the cost of each synapse with the number of its hierarchical level ( e . g . synapses in level were assigned a cost of ) , and averaging the cost over all synapses . Higher density drop-off rates were associated with lower wiring cost ( Figure 2b ) . The low wiring cost was in turn associated with higher clustering coefficients and higher characteristic path lengths in the network ( Figure 2b–d ) . Clustering coefficients and characteristic path lengths are simple measures of modular organization and between-module connectivity , respectively [41] . We integrated subthreshold neuronal dynamics exactly , interpolated neuronal spike times between intervals and recorded neuronal activity at bins [42] . We began all simulations by setting all synaptic weights to and setting all membrane potentials to uniformly distributed random values from to . We discarded five minutes of initial activity , ensuring in each case that synaptic weights converged to a stable distribution . We recorded five minutes of subsequent activity and described this activity in terms of module spikes . Module spikes represent simultaneous activations of large numbers of within-module neurons , and hence correspond to network spikes described in empirical data [43]–[44]; we used the term module spike , rather than network spike , to avoid potential confusion with global network synchrony . We explicitly note that module spikes are conceptually distinct from individual neuron spikes . We determined the occurrence of module spikes with a shuffling algorithm that preserved individual spike frequency but destroyed global patterns of network activity . In this algorithm , spike times of all excitatory within-module neurons are randomly shuffled between active time bins . Module spikes are then said to occur when the number of simultaneously active neurons in the original data exceeds a threshold corresponding to the number of simultaneously active neurons in of the shuffled data . For each module , we averaged the spike threshold from shuffled matrices . It is also possible to describe network activity in terms of individual neuron spikes , rather than in terms of module spikes . In our simulations , neurons were likely to spike in module-specific groups , and neuronal spikes were hence strongly correlated with module spikes ( Figure 3 ) . We concentrated on module spike patterns because these describe activations of neuronal ensembles and have clear parallels with population spikes observed through changes of local field potentials in empirical studies of self-organized criticality [5] , [12] . Neuronal spike patterns are studied in more detail elsewhere , e . g . in memory consolidation [45] . We also note that neuronal activity is likely to occur at every time point in large networks; consequently descriptions of avalanches of individual neuron spikes require a global network threshold to remove background activity . In our simulations , this threshold resulted in minimal event sizes of neurons , which , together with maximal event sizes of neurons , made rigorous detection of power-law scaling computationally prohibitive . We defined an avalanche as a sequence of temporally continuous ( in bins ) module spikes , preceded and followed by a period of inactivity [5] . Correspondingly , we defined the avalanche size as the number of module spikes in the avalanche , and the avalanche duration as the total time between onset and conclusion of the avalanche . The minimal avalanche has size module and duration . The maximal avalanche may be arbitrarily large because modules can be potentially active multiple times in the same avalanche . More realistically , the overwhelming majority of avalanches in our simulations , especially in simulations with neurobiologically realistic connectivities ( Figure 7a ) , did not exceed the system size of modules . Probability distributions of avalanche sizes and durations allow a concise quantification of network dynamics . For instance , subcritical dynamics are characterized by small avalanche sizes and rapidly decaying avalanche size distributions , while supercritical dynamics are characterized by large avalanche sizes and slowly decaying avalanche size distributions . Critical dynamics are characterized by avalanche sizes and durations that follow power-law distributions , with a cumulative distribution function , where is avalanche size or duration , is the scaling exponent , and are upper and lower cut-offs and is the generalized Hurwitz zeta function . The functions explicitly incorporate an upper cut-off , as distributions are necessarily bounded by system size [46] . In the following , we set to the maximal event size in each distribution . We rigorously assessed the presence of power-law scaling in avalanche distributions , by adapting the methods described in Clauset et al . [36] . We hence estimated using the method of maximum likelihood . This method is mathematically robust and accurate for large number of samples ( in our simulations ) , unlike linear least-squares-based methods commonly used in previous studies . For a given , we estimated by numerically maximizing the log-likelihood function , where , are the observed values of , such that for all . We imposed the condition and this conservative condition ensured that we considered a wide range of events . We then chose the , pair that minimized the Kolmogorov-Smirnov statistic , where is the cumulative distribution function of the data and is the cumulative distribution function of the fitted model . We formally assessed the power-law goodness-of-fit , by generating synthetic power-law distributions with equivalent , , and . For each generated dataset we individually estimated and , and computed the statistic as above . This procedure gives a -value as the fraction of instances in which the statistic of the generated data exceeds the statistic of the original data . We deemed that [47] did not allow to reject the power-law hypothesis , and hence suggested power-law scaling . Smaller or larger -values ( ) did not qualitatively change our results . We imposed three additional conditions to ensure meaningful power-law scaling . Firstly , we required that maximal avalanche sizes approach system limits ( modules ) , to ensure that power laws did not reflect rapidly decaying subcritical dynamics . Secondly , we required that avalanche distributions extracted from corresponding shuffled module spike matrices had goodness-of-fit . Thirdly , we directly compared power-law and exponential distribution fits , by computing the log-likelihood ratio for the best-fitting power-law and exponential distributions . The corresponding probability distribution , cumulative distribution and log-likelihood functions for the exponential distribution are , respectively , where is the exponential parameter . The log-likelihood ratio compares two distributions and identifies a distribution which fits the data better . A significance test on the log-likelihood ratio gives a -value on the statistical significance of this comparison [48] , [36]] . We deemed that indicated a statistically significant difference in fit between distributions . We did not attempt to compare power-law and log-normal distribution fits because it is very difficult to differentiate these two distributions and hence such comparisons are typically inconclusive [36] . We summarized the presence of power-law scaling in each distribution with a single statistic . For each distribution , equaled the goodness-of-fit -value for the power-law model if the distribution additionally fulfilled the above three conditions; alternatively was set to . We averaged over independent simulations for each type of connectivity , and considered to indicate power-law scaling .
We initially examined dynamics emergent on nonhierarchical modular networks ( Figure 4a ) . We gradually randomized these networks by rewiring excitatory connections in a way that increased the number of connections between modules . At one extreme , ordered nonhierarchical networks had no intermodule synapses . At the other extreme , random nonhierarchical networks had homogeneously distributed intra- and intermodule excitatory synapses . Between these two extremes , nonhierarchical networks had a varying number of homogeneously distributed intermodular excitatory synapses . The location of synapses in each network was fixed during simulations , but synaptic weights continuously fluctuated according to the STDP rule . All nonhierarchical networks had a connectivity-independent neuron spike rate of , and a stable weight distribution ( Figure 1b ) . In addition , these networks had module spike rates of . Ordered networks had no intermodular connections , and correspondingly showed subcritical uncoordinated dynamics . Random networks had large numbers of intermodular connections and correspondingly showed supercritical globally synchronous dynamics . A narrow range of network topologies between these two extremes was associated with critical dynamics , characterized by power-law distributions of avalanche sizes and durations ( Figure 4b , c ) . Distributions of inter-avalanche intervals likewise changed from subcritical to supercritical , but did not follow consistent power laws at this transition ( Figure 4b ) . Despite the stable weight distributions , activity-dependent fluctuations in synaptic weights continuously occurred ( Figure 5a , b ) . In order to investigate the impact of these fluctuations on global network dynamics , we examined the effect of freezing plasticity after five minutes of initial transient simulation . This procedure fixed the values of individual weights , and hence preserved the same neuronal spike rate of . However , this procedure dramatically disrupted within-module neuronal synchrony: module spike rate dropped to less than and dynamics on all networks became highly subcritical ( Figure 5c , d ) . Module spike rate remained negligible despite increases in external current , and consequent increases in neuronal spike rate . Furthermore , module spike rate remained negligible with an even more stringent control condition , which allowed synaptic weight changes at spike times , but made these changes by randomly drawing weights from the distribution in Figure 1c , rather than according to the STDP rule ( results not shown ) . On the other hand , as we show below , a change from soft to hard bounds in the STDP rule preserved equivalent dynamics , despite changing the weight distribution ( Figure 1c–e ) . In addition , halving the STDP learning rate preserved equivalent dynamics when network size was doubled . Together , these findings indicate that the precise patterns of STDP-driven fluctuations enabled the formation of coherent within-module dynamics in our model . Nonhierarchical connectivity is neurobiologically implausible , because of the high wiring cost associated with a large number of long-range connections , and because hierarchical modularity is evident in multiscale neuroanatomical organization [25] . We hence examined a more plausible connectivity by defining a framework in which connections were probabilistically placed within explicit spatial hierarchical levels , according to predefined power-law , exponential and linear scaling functions ( see Methods and Figure 2 ) . Figure 6 compares the critical regimes associated with nonhierarchical connectivity ( Figure 6a ) , and with hierarchical power-law , exponential and linear ( Figure 6b–d ) connectivities . The rows in Figure 6b–d represent different wiring costs for each hierarchical organization . Most strikingly , low-cost power-law and exponential connectivities were associated with a broad critical regime . This regime was especially evident for the power-law connectivity with ( fourth row in Figure 6b ) , as this was the only studied connectivity simultaneously associated with a broad regime of power-law distributed avalanche sizes and power-law distributed avalanche durations . Connectivities with higher wiring cost , such as all linear connectivities , showed narrow critical regimes . Connectivities with very low wiring cost did not show broad critical regimes , presumably because the numbers of long range connections in these connectivities were insufficient to enable the emergence of large events . Figure 7a , b shows statistically significant power-law distributions of avalanche sizes and durations for the optimal power-law , exponential and linear connectivities . The greater number of power-law distributions for the power-law and exponential connectivities , compared with linear connectivity , is clearly visible . Figure 7c illustrates the values of power-law exponents for connectivities in which avalanche sizes and durations simultaneously followed statistically significant power laws . Exponents of avalanche size distributions associated with power-law connectivities were close to and hence accurately resembled empirically estimated exponents of neuronal avalanche size distributions at the same bin size [5] , [14] . Exponents decreased with increasing network randomization . We sought to disambiguate the association between modularity and the broad critical regime by examining dynamics emergent on lattices with optimal power-law connectivity , but no explicit modular structure ( Figure 8a ) . For this purpose , we constructed lattices of the same size and degree as the hierarchical connectivity networks , and we randomized these lattices by distributing off-diagonal connections according to the power-law density scaling function with . In this way , we could focus on the effect of hierarchical modularity by retaining most other features of original network organization , including wiring cost . Dynamics on these lattice networks had substantially reduced module spike rates ( ) and were associated with a rapid phase transition and a loss of the broad critical regime ( Figure 8c , top ) . An increase in external current restored the original module spike rate of and consequently broadened the critical regime , although not to the original level ( Figure 8c , middle ) . On the other hand , when modularity was implicitly reintroduced by rearranging inhibitory synapses into modules ( Figure 8b ) , a broad critical regime reappeared without changes in external current ( Figure 8c , bottom ) . These findings suggest that modularity of inhibitory connections facilitated coherent within-module dynamics . We explored robustness of the broad critical regime ( for the optimal power-law density scaling function ) to other meaningful changes in neurobiologically relevant parameters , such as changes in external current , changes in conduction delays , changes in the postsynaptic response , presence of neuronal inhibition , changes in the STDP rule and changes in network size ( Figure 9 ) . Theoretically , self-organized criticality emerges in systems with low external drive and strong internal interactions , and the responses of our model to variation of parameters were meaningful in this context . It is worth noting that we assessed the strength of external drive by the associated neuronal spike rate . Specifically , we considered the external current of to represent a low external drive even though this value substantially exceeds the minimal value of required to sustain neuronal activity ( see Text S1 for details ) . In our simulations the broad critical regime was robust to moderate variations of external current and delays ( Figure 9a , b ) , but began to disappear when external current exceeded ( as external drive became too strong ) , or when delay lengths were quadrupled to the range of ( as internal interactions lost spike precision ) . The regime was narrowed when the postsynaptic response weakened ( Figure 9c , top ) , but was preserved when the STDP learning rate was reduced ( Figure 9c , bottom ) . In both cases , we controlled for changes in neuronal spike rate by increasing external current . The regime was broadened by a stronger postsynaptic response and by a higher STDP learning rate ( results not shown , as the associated parameter values are unrealistically high ) . We hypothesized that our network models required strong postsynaptic responses and fast STDP learning rates to compensate for the small number of synaptic connections of each neuron . Excitatory neurons in our model connected with only other neurons , while in vivo each neuron is thought to have thousands of synapses . We compensated for the small number of connections in our model by setting the postsynaptic-response magnitude of each neuron to a value which could theoretically exceed the neuron spike threshold and by using an instantaneous STDP learning rate that substantially exceeds empirically observed values ( Table 1 ) . When we doubled our module size to neurons , and consequently doubled our network size to neurons , we were able to simultaneously halve the values of postsynaptic-response magnitudes and STDP learning rates and hence bring these values much closer to empirically observed values [49] . Specifically , the broad critical regime in these larger networks was preserved when the postsynaptic-response magnitude was halved , the STDP learning rate was halved , and the external current was reduced from to ( Figure 9d , top ) . Alternatively , the regime was preserved when the postsynaptic-response magnitude was halved , the STDP learning rate remained unchanged , and the external current was halved ( Figure 9d , bottom ) . These findings show that realistically large numbers of synaptic connections are likely to facilitate strong internal interactions in the presence of biologically realistic parameter values . In addition to these variations , the broad critical regime did not qualitatively change when inhibitory synapses were removed , provided the loss of inhibition was controlled by reductions in external current ( Figure 9e ) . The broad critical regime was likewise preserved when soft weight bounds were changed to hard weight bounds in the STDP rule ( Figure 9f ) .
We found that despite seemingly stable neuronal activity , spike-timing-dependent plasticity enabled coherent within- and between-module neuronal activity . Furthermore , we showed that two variations of the STDP rule produced distinct weight distributions , but enabled a broad critical regime on conducive network topologies . In contrast , fixed or randomly altered synaptic weights were associated with subcritical dynamics and negligible module spike rates . STDP may facilitate coherent within-module activity by intermittently potentiating and depressing synapses between reciprocally connected neurons . In small networks , simulations showed that intermittent synaptic potentiation and depression was associated with pairwise neuronal synchrony , fluctuations of synaptic weights and continuous reversal of phase differences between reciprocally connected pairs of neurons ( results not shown ) . In our networks , within-module weights were potentiated during module spikes , and depressed between module spikes ( Figure 5b ) . These activity-dependent fluctuations hence clearly played an important role in facilitating neuronal ensemble synchronization . Recent studies have shown the importance of short-term synaptic depression in self-organized critical dynamics in networks of spiking neurons , but have not concurrently considered the effects of STDP [33] , [35] . Our study illustrates the importance of STDP in self-organization and hence provides a alternative generative model of critical dynamics in networks of spiking neurons . A principled comparison of the role of these two forms of plasticity in self-organized criticality is hence an important subject of future research . The distinct mechanism of these forms of plasticity may also allow to disambiguate their role empirically with pharmacological manipulations in real neuronal systems . Modular networks with low wiring cost showed a broad critical regime . Modular networks with high wiring cost showed a narrow critical regime , possibly due to high numbers of costly long-range connections , which enabled a rapid onset of globally synchronous , supercritical dynamics . Lattice networks with low wiring cost showed a narrowed critical regime due to uncoordinated inhibition and a consequent loss of coherent ensemble dynamics . Modularity and low wiring cost were hence simultaneously required for self-organized criticality to emerge . This simultaneous requirement is notable , as both properties are thought to be ubiquitously present in neuroanatomical organization . In an early comprehensive exposition , Jensen [4] addressed the potentially confusing meaning of self-organization to criticality: “[s]elf-organization to criticality will definitely occur only under certain conditions; one will always be able to generalize a model sufficiently to lose the critical behavior . Hence the question becomes just what is relevant in a given context . This is where a super-general approach must be supplemented by insight from the specific science to which a given system belongs . ” In this spirit , we examined neurobiologically meaningful variations in parameters such as external current and conduction delays . We found that the broad critical regime was generally preserved despite variations of these parameters and , consequently , finetuning was not required for self-organized critical dynamics to emerge . More specifically , strong synaptic interactions with low external current ( i . e . short delays , strong postsynaptic responses , high STDP learning rate ) favored a broad critical regime , while weak synaptic interactions with high external current ( i . e . long delays , weak postsynaptic response , low STDP learning rate ) favored a narrow critical regime . These findings indicate that critical dynamics primarily emerged through internal interactions , rather than external drive . The findings hence provide further evidence for the self-organizing nature of the observed dynamics . The strong postsynaptic response and STDP learning rate in our model compensated for the relatively low synaptic connectivity , and could be markedly lowered in larger networks without detriment to the broad critical regime . We found that inhibitory neurons in our model did not explicitly enable a broad critical regime . In contrast , recent network simulations of simple stochastic neurons by Benayoun et al . [50] show that inhibitory neurons enable self-organized criticality by balancing the network . However , the differences in neuronal dynamics , and the absence of statistically significant power laws in the Benayoun et al . study , make it difficult to directly compare our findings . We do show however , that the presence of inhibitory neurons in our networks was compatible with self-organized critical dynamics only if these neurons were organized in modules . These modules correspond to realistic local inhibitory connectivity , rather than the less realistic long-range inhibitory connectivity . Inhibitory neurons may also play a more prominent role in other types of network dynamics , such as oscillations . Our findings may be used to generate empirically testable hypotheses of the relationship between anatomical connectivity and emergent network dynamics . For instance , we hypothesize that self-organized critical dynamics in dissociated neuronal cultures emerge on a low-cost modular neuroanatomical connectivity . Recent studies show that dissociated neuronal cultures self-organize towards a critical state , via subcritical and supercritical states [12]–[13] , [51] . Cultured dissociated neurons self-organize by forming axonal and dendritic arborizations , and synaptic connections [44] . In the first week of culture , self-organization is non-activity-dependent , and may show preference towards spatial proximity . After the first week of culture , the network becomes spontaneously active , and self-organization becomes activity-dependent . Our findings may hence be used to explicitly compare structure and dynamics of dissociated neuronal cultures , throughout this period of self-organization . A recent study found that functional activity patterns of dissociated neuronal cultures constitute a small-world network [52] . Novel methods of network reconstruction from avalanche dynamics [53] may allow to study structural network properties of these cultures . For instance , future empirical work may study the relationship between specific anatomical measures ( e . g . wiring cost ) and dynamical measures ( e . g . exponent values of power-law distributions ) in such networks , throughout self-organization . Alternatively , it may be possible to study dynamics in real neuronal networks with externally controlled anatomical connectivity [54] . A clear limitation of our study is the oversimplified symmetric hierarchical organization and the relatively small size of our model . Substantial increases in the number of modules , and in the number of neurons within modules , are required to make realistic inferences about neuronal dynamics at larger scales . The study hence sets the groundwork for simulations of large networks of spiking neurons and for characterization of spatiotemporal activity patterns emergent on these networks . Such simulations may be conducted on increasingly detailed maps of large-scale anatomical connectivity in healthy subjects [55]–[57] and in subjects with connectivity disorders , such as Alzheimer's disease [58] and schizophrenia [59] . These simulations will be the subject of future studies . Studies of neuronal dynamics often employ numerical integration schemes ( such as the Euler method ) , and manually store all previous spike times to compute synaptic currents . An advantage of the integrate-and-fire neuron model is the ability to integrate subthreshold activity exactly and incorporate effects of all previous spikes without the need for explicit summation at each step [60] . In addition , interpolation of spike times between time steps avoids artefactual synchrony and is especially important in simulations with spike-timing-dependent plasticity . Hence , while our results remain subject to numerical error , the particular integration scheme we employ [42] substantially reduces the possibility of numerical artefacts . Despite growing empirical evidence for self-organized criticality , several important studies argue against this evidence , by either noting the potential for spurious reports of power-law scaling , or by attributing such scaling to simpler mechanisms , such as diffusive processes [47] , [61]–[62] . Two observations favor the presence of self-organized criticality in our model . Firstly , we estimate power-law scaling with rigorous statistical tests [36] , rather than the more commonly used unreliable linear least-squares-based methods . We use a method with very high specificity and we can hence be highly certain that the detected power-law distributions are genuine . On the other hand , the method may have potentially low sensitivity , and may hence underestimate the presence of power laws in our data . Secondly , we find that these power-law distributions are associated with a phase transition , suggesting that dynamics evolve at the critical point . In addition , we note that it is not straightforward to compare findings between studies that focus on different scales and types of neuronal activity . Hence , while much evidence for critical brain dynamics comes from studies of low frequency spatiotemporal dynamics ( as in this study ) , these dynamics cannot be trivially related to other phenomena , such as noise-like processes in recordings of high frequency neurophysiological signals [62] . In conclusion , we show an association between modularity , low cost of wiring , synaptic plasticity and self-organized criticality in a neurobiologically realistic model of neuronal activity . Our findings theoretically reinforce the reciprocal relationship between connectivity and dynamics on multiple spatial scales .
|
The intricate relationship between structural brain connectivity and functional brain activity is an important and intriguing research area . Brain structure ( the pattern of neuroanatomical connections ) is thought to strongly influence and constrain brain function ( the pattern of neuronal activations ) . Concurrently , brain function is thought to gradually reshape brain structure , through processes such as activity-dependent plasticity ( the “what fires together , wires together” principle ) . In this study , we examined the relationship between brain structure and function in a biologically realistic mathematical model . More specifically , we considered the relationship between realistic features of brain structure , such as self-similar organization of specialized brain regions at multiple spatial scales ( hierarchical modularity ) and realistic features of brain activity , such as self-similar complex dynamics poised between order and randomness ( self-organized criticality ) . We found a direct association between these structural and functional features in our model . This association only occurred in the presence of activity-dependent plasticity , and may reflect the importance of the corresponding structural and functional features in neuronal information processing .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"neuroanatomy",
"neural",
"networks",
"computational",
"neuroscience",
"biology",
"neuroscience"
] |
2011
|
Neurobiologically Realistic Determinants of Self-Organized Criticality in Networks of Spiking Neurons
|
Toll-like receptor ( TLR ) ligands are being considered as adjuvants for the induction of antigen-specific immune responses , as in the design of vaccines . Polyriboinosinic-polyribocytoidylic acid ( poly I:C ) , a synthetic double-stranded RNA ( dsRNA ) , is recognized by TLR3 and other intracellular receptors . Poly ICLC is a poly I:C analogue , which has been stabilized against the serum nucleases that are present in the plasma of primates . Poly I:C12U , another analogue , is less toxic but also less stable in vivo than poly I:C , and TLR3 is essential for its recognition . To study the effects of these compounds on the induction of protein-specific immune responses in an animal model relevant to humans , rhesus macaques were immunized subcutaneously ( s . c . ) with keyhole limpet hemocyanin ( KLH ) or human papillomavirus ( HPV ) 16 capsomeres with or without dsRNA or a control adjuvant , the TLR9 ligand CpG-C . All dsRNA compounds served as adjuvants for KLH-specific cellular immune responses , with the highest proliferative responses being observed with 2 mg/animal poly ICLC ( p = 0 . 002 ) or 6 mg/animal poly I:C12U ( p = 0 . 001 ) when compared with immunization with KLH alone . Notably , poly ICLC—but not CpG-C given at the same dose—also helped to induce HPV16-specific Th1 immune responses while both adjuvants supported the induction of strong anti-HPV16 L1 antibody responses as determined by ELISA and neutralization assay . In contrast , control animals injected with HPV16 capsomeres alone did not develop substantial HPV16-specific immune responses . Injection of dsRNA led to increased numbers of cells producing the T cell–activating chemokines CXCL9 and CXCL10 as detected by in situ hybridization in draining lymph nodes 18 hours after injections , and to increased serum levels of CXCL10 ( p = 0 . 01 ) . This was paralleled by the reduced production of the homeostatic T cell–attracting chemokine CCL21 . Thus , synthetic dsRNAs induce an innate chemokine response and act as adjuvants for virus-specific Th1 and humoral immune responses in nonhuman primates .
Effective vaccines against infections caused by intracellular pathogens including HIV infection , malaria , or tuberculosis most likely will need to induce strong cellular and humoral immune responses [1] . Current vaccine strategies under development are based on prime-boost immunizations , such as vaccination with plasmid DNA followed by booster injections with replication-incompetent viral vectors ( e . g . , adenoviruses or poxviruses ) , with both DNA and viruses encoding immunogenic proteins of the pathogen [2] . There is concern that these strategies may be insufficiently immunogenic and protective , so alternative vaccine approaches are under development [3] , [4] . While protein based vaccines allow the delivery of large amounts of immunogenic vaccine antigens , particularly when targeted to antigen presenting dendritic cells ( DCs ) [5] , these vaccines require the identification of appropriate adjuvants [6] , which may act by differentiating the DCs to elicit strong immunity [7]–[10] . Monkeys are being used as an animal model to develop AIDS vaccines and are likely to be a valuable preclinical model to identify adjuvants and understand their mode of action . Currently the most widely used adjuvant is aluminum hydroxide . It predominantly induces Th2 immune responses [11] , and as such may be inappropriate for HIV or tuberculosis vaccines or for immune therapy of tumors related to infection by human papillomaviruses ( HPV ) . Ligands for pathogen recognition receptors , e . g . , Toll-like receptor ( TLR ) ligands , can stimulate cells of the innate and adaptive immune systems and have therefore been proposed as promising adjuvant candidates [12] , [13] . We have previously studied the effects of TLR9 ligands , i . e . , CpG-A and CpG-B , on the induction of protein-specific immune responses in nonhuman primates . However , we did not observe strong CD4+ T cell-mediated immune responses as indicated by T cell proliferative assays [14] . This may in part be due to the lack of TLR9 expression in myeloid primate DCs [15] , which can be valuable for the priming of naïve T cells and the induction of cellular immune responses [16] , [17] . In this study , we have focused on synthetic double stranded RNA ( dsRNA ) compounds as adjuvants . They can be recognized by both TLR3 [18] and the melanoma differentiation-associated gene-5 ( MDA-5 ) [19] , pattern recognition receptors that are expressed by many cell types and are involved in anti-viral immune responses [20] . In mice , polyriboinosinic-polyribocytoidylic acid ( poly I:C ) has long been known as a strong IFN-α inducer and provides anti-viral and adjuvant activity [21] , [22] . Poly I:C also works as a mucosal adjuvant for the induction of humoral and cell-mediated immune responses [23]–[25] . MDA-5 is important for the IFN response induced by poly I:C [26] , [27] . In primates , poly I:C is a less effective IFN-α inducer , most likely due to nucleases , which reduce the biostability of poly I:C and are reported to be more prevalent in the serum of primates than rodents [28] . A complex of poly I:C with poly-L-lysine and carboxymethylcellulose ( poly ICLC ) , however , is five to 10 times more resistant to hydrolysis by RNAse in primate serum than the parent poly I:C and induces significant levels of interferon in monkeys under conditions in which poly I:C itself induces no interferon [29] , [30] . Poly ICLC possesses anti-viral activity against a variety of viruses in monkeys [31]–[33] and chimpanzees [34] , and also inhibits malaria infection of macaques [35] . Furthermore , it has shown potent adjuvant activity on the induction of humoral immune responses in the nonhuman primate models of Venezuelan equine encephalomyelitis virus and swine influenza virus [36] , [37] . In humans , dose-dependently , mild to moderate side effects of poly ICLC were observed in a number of phase I and II studies conducted in children and adults [38]–[45] . Another synthetic dsRNA , poly I:C12U ( Ampligen ) , supports the induction of broad antiviral immune responses in mice [46] , [47] , shows low toxicity in humans [48] , and should therefore also be considered as an adjuvant in human vaccine trials . To date , no studies have been reported on the potential of synthetic dsRNA to augment cellular immunity in primates . We therefore have performed studies in rhesus macaques to address the impact of dsRNA on the induction of protein-specific immune responses . As a prelude to studies with protein based vaccines , we selected keyhole limpet hemocyanin ( KLH ) . In contrast to a previous study where TLR7/8 and TLR9 ligands have been used as adjuvants for cellular immunity in rhesus macaques [49] , we injected the dsRNA plus KLH in aqueous solution without additional emulsification in water-in-oil adjuvants , such as Montanide , to minimize the risk of undesired side-effects at the site of injection . To confirm that the adjuvant effect of dsRNA is also manifest in the context of the injection of viral proteins , we injected some animals with the major capsid protein ( L1 ) of HPV16 with or without poly ICLC . HPV16 is the major carcinogenic genotype of HPV in most countries and involved in about 50% of the cases of cervical cancer worldwide [50] . Recently , prophylactic vaccines against HPV16 have been marketed that consist of L1 virus-like particles ( VLPs ) and induce neutralizing antibodies that efficiently protect against persistent HPV infection and premalignant cervical lesions [51] . However , therapeutic vaccines for the use in individuals who are already infected will need to induce cellular immunity , most likely against the E6/E7 antigens of HPV . Subunits of VLPs ( pentameric capsomeres ) have potential advantages over VLPs , i . e . , higher stability and reduced production costs but their immunogenicity has not yet been evaluated in nonhuman primates . To monitor also the innate response to dsRNAs , we concentrated on the rapid innate production of CXCL9 ( MIG ) and CXCL10 ( IP-10 ) chemokines , which are induced by dsRNA [52] as well as CCL21 ( SLC ) , which attracts naïve T lymphocytes and DCs [53] . Here we show that dsRNAs act as adjuvants for the induction of innate and adaptive cellular and humoral immunity in nonhuman primates .
Poly ICLC has adjuvant activity on the induction of humoral immunity at doses as low as 0 . 1 mg/kg [37] . Since we assumed that higher doses might be required for the induction of cellular immune responses , we immunized rhesus macaques subcutaneously ( s . c . ) with KLH and either poly ICLC ( 0 . 5 mg/kg body weight; 6 animals ) , poly I:C ( 0 . 5 mg/kg; 4 animals ) , or without adjuvant ( 4 animals ) . To monitor the development of T cell immunity , we cultured peripheral blood mononuclear cells ( PBMC ) with or without KLH and determined whether immunization resulted in T cell proliferative responses to the administered antigen by 3H thymidine incorporation assays . Peak stimulation indices ( SI; KLH-induced proliferation divided by proliferation in medium alone ) were significantly higher ( p = 0 . 040 ) in animals injected with poly ICLC ( week 0 , 1 . 93±1 . 38; peak , 23 . 00±8 . 02 ) compared with KLH alone ( week 0 , 3 . 28±1 . 55; peak , 8 . 97±7 . 47 ) , individual maximum proliferative responses are shown in Table S1 . The kinetics of the responses are shown in the Figure S1 and reveal significantly stronger proliferative responses in poly I:C co-injected animals than in controls six weeks post injection ( p = 0 . 03 ) . Thus , KLH-dependent proliferation of PBMCs was induced with poly ICLC or poly I:C by the s . c . route . The poly I:C analogue poly I:C12U requires TLR3 to be active in vivo [54] , [55] and shows little toxicity in humans [48] . Like poly I:C and in contrast to poly ICLC , it is not stabilized against primate serum nucleases . We therefore compared the effectiveness of poly I:C12U to poly ICLC in a separate study , using fixed standardized doses per animal rather than adjustment to body weight . To study the effects of dsRNA on cellular immune responses in more detail , we used a carboxyfluorescein diacetate succinimidyl ester ( CFSE ) dilution assay , which allows the separate evaluation of CD4+ and CD8+ T cells ( Figure 1A ) . KLH at 200 µg/animal was administered to five animals each either alone or together with 2 or 6 mg poly I:C12U or 2 mg poly ICLC per animal ( i . e . , 0 . 27 to 0 . 44 , 1 . 00 to 1 . 13 , and 0 . 33 to 0 . 43 mg/kg , respectively ) . The dose of 2 mg per animal has been used in previous studies on TLR agonists as vaccine adjuvants in monkeys [49] and thus facilitates comparison between studies of different adjuvant compounds . KLH-specific CD4+ T-cell proliferation at week 2 after immunization was significantly higher when KLH was given with either 6 mg poly I:C12U ( p = 0 . 001 ) or 2 mg poly ICLC ( p = 0 . 002 ) whereas at that time point no significant difference to KLH alone was observed after the injection of 2 mg poly I:C12U ( p = 0 . 16; Figure 1B ) . The effect of immunization with poly I:C12U or poly ICLC on proliferative responses was sustained over 6 weeks , and there was a significant difference also for the 2 mg poly I:C12U group over KLH alone at this time point ( p = 0 . 013; Figure 1B ) . Thus , all three synthetic dsRNA compounds that we tested could serve as adjuvants for the induction of protein-specific T-cell proliferation in primates . To examine possible dose-dependent adjuvant effects of poly ICLC , we compared in a prime-boost experiment the effects of 0 . 5 mg/kg body weight with those of 0 . 1 mg/kg body weight , which is still sufficient for the induction of humoral immune responses [37] . After the first immunization increased KLH-specific CD3+CD4+ T cell proliferative responses were seen in both animals immunized with 0 . 5 mg poly ICLC/kg , and CD3+CD4− T cells ( as a surrogate for CD8+ T cells ) were expanded to a similar extent ( Fig . S2A ) . CD4+ and CD4− T cell-proliferative responses were less pronounced after the primary immunization together with the lower 0 . 1 mg/kg dose of poly ICLC ( Figure S2B ) . Booster immunization at week 14 enhanced the proliferative CD4+ T cell responses in the animal 13404 immunized with 0 . 5 mg poly ICLC/kg ( Figure S2A ) and in the animal 13406 receiving the lower dose of poly ICLC ( Figure S2B ) . Therefore , 0 . 5 mg/kg of poly ICLC might be more active as an adjuvant for cellular immunity than lower doses . To confirm that dsRNA analogues also serve as adjuvants in the context of a clinically relevant viral antigen , we injected s . c . six animals each with a low dose of HPV16 L1 capsomeres ( 10 µg ) with or without 2 mg of poly ICLC . Another six animals were injected with 2 mg of the TLR9 ligand CpG-C ( ODN 2396 ) , which supports the induction of protein-specific cellular immune responses in monkeys when injected in a water-in-oil emulsion [49] . We selected the L1 pentamers rather than the complete virus-like particles ( VLP; 360 molecules of L1 ) , since capsomeres are promising candidates for 2nd generation vaccines but their immunogenicity in nonhuman primates has not yet been evaluated . The capsomeres were obtained by expression of a modified L1 protein in baculovirus-infected insect cells [56] . In the immune assays , we re-stimulated PBMCs with HPV16 VLPs and used mouse norovirus VLPs ( A447 ) generated in the same expression system as a negative control antigen . All animals were boosted with a second injection of antigen +/− adjuvant eight weeks later . Numbers of IFN-γ secreting cells in the peripheral blood were determined by ELISPOT assay . At week 2 , we detected increased numbers of HPV-specific , IFN-γ secreting cells in PBMCs from 4 out of 6 animals ( Figure 2A ) injected with antigen plus poly ICLC , and the responses waned in all animals by week four . Two weeks after the booster injection ( at week 10 after the first injections ) , however , all six animals injected with antigen together with poly ICLC had developed HPV-specific , IFN-γ secreting cells , which also were maintained two weeks later ( week 12 ) and still present in 3 animals at week 19 . In contrast , IFN-γ secreting cells were detectable at elevated numbers in only one of the CpG-injected monkeys ( animal 13928 ) four weeks after the first injection and this response could not be boosted by the second injection . None of the control animals showed substantial numbers of HPV-specific IFN-γ secreting cells , neither following the first nor the booster injection ( Figure 2A ) . The background responses against A447 might be induced by contaminating protein fractions derived from the expression system , in which both antigens , i . e . , HSV16 L1 capsomeres and A447 , had been generated . When we assessed T cell proliferation in CFSE assays , we found significantly enhanced HPV-specific CD4+ T-cell proliferative responses in the poly ICLC-injected monkeys four weeks after the second application of antigen ( p = 0 . 008 ) ( Figure 2B ) . Figure 2B depicts the proliferation of CD3+CD8− cells , re-stimulated for the last 6 h of the assay with peptide pools 1–4 . Similar results ( p = 0 . 012 , week 12 ) were obtained with cells re-stimulated for the final 6 h with pools 5–8 ( data not shown ) . At week 19 , proliferative responses did not differ significantly in poly ICLC-injected and control animals . To further characterize the Th cell responses , we determined the concentrations of IFN-γ , IL-4 , and IL-17 in supernatants collected from re-stimulated PBMCs 2 d after setting up the assays . We used ELISAs for the detection of monkey cytokines or , in the case of IL-17 an ELISA for the detection of the human protein but known to cross-react with monkey IL-17 [57] . Following the booster injection at week 8 , we detected significantly more IFN-γ in the supernatants of cells collected from poly ICLC-injected animals than in those of cells from control animals and these responses were sustained until week 19 ( Figure 2C ) . In contrast , we were unable to detect IL-4 or IL-17 in the supernatants from assays set up with PBMCs from either group of animals . Thus , poly ICLC supports the induction of HPV-specific Th1 immune responses , i . e . , CD4+ T cell proliferative responses and IFN-γ secretion . We also determined the humoral immune responses induced by the injection of HPV16 L1 capsomeres with or without adjuvants . Injection of poly ICLC or CpG-C resulted in up to 1000fold increased titers of binding antibodies ( measured by ELISA ) compared with control animals ( Figure 3A; p<0 . 01 for both adjuvants for weeks 4 , 8 , 10 , and 12 ) , and at weeks 4 , 8 , and 10 , poly ICLC also induced higher titers than an equal dose of CpG-C ( p<0 . 05 for week 4 , p<0 . 01 for weeks 8 and 10 ) . The individual antibody titers of all animals are shown for all points in time in the Table S2 . In addition , we performed neutralization assays using the serum samples collected 12 weeks after first immunization and HPV16 pseudovirions as targets . Sera of the animals from both adjuvant groups showed considerable neutralizing activity while samples from the control animals were not able to neutralize the activity of the pseudovirions in our assay ( Figure 3B ) . Poly ICLC injected animals showed stronger responses than monkeys that had received CpG-C ( p = 0 . 03 for serum dilutions of 1∶1000 ) . There was a good correlation between ELISA and neutralization titers in the sera of the individual animals ( Figure S3 ) . Therefore , while CpG-C mainly affects the induction of antibodies , poly ICLC acts as adjuvant for both humoral and cellular immunity . Since we have previously observed that poly I:C activates monkey DCs [58] , immunohistochemistry was performed to determine the number and activation status of DCs present in lymph nodes taken prior to immunization and at 18 h after injection of poly ICLC . The numbers of phenotypically immature ( CD1a+ ) and mature ( CD83+ or CD208+ ) DCs varied between animals but did not show a clear decrease or increase after immunization ( Figure S4 ) . Draining inguinal lymph nodes were also analyzed for CXCL10 , CXCL9 , and CCL21 by in-situ hybridization and immunohistochemistry . In comparison to control lymph nodes removed before immunizations , elevated expression of CXCL10 ( Figure 4A and 4B ) and CXCL9 ( Figure 4C and 4D ) was detected in the T cell areas of draining lymph nodes at 18 hours after immunization . Chemokine mRNA expression correlated with protein expression detected by immunohistochemistry ( insets in Figure 4 ) . Expression of CCL21 mRNA ( Figure 5A ) and protein ( Figure 5B and 5C ) 18 hours after immunization was markedly decreased in draining lymph nodes compared with control lymph nodes obtained before immunization . Thus , the innate response to dsRNA is detectable in lymph node cells . The administration of poly ICLC or poly I:C together with KLH led to a significant increase of serum levels of CXCL10 ( Figure 6A; p = 0 . 001 for both compounds at 18 or 24 h ) . Furthermore , 48 h after immunization , serum levels of CXCL10 were significantly higher in poly ICLC- than in poly I:C-injected monkeys ( p = 0 . 027 ) . Like poly I:C at 0 . 5 mg/kg , poly I:C12U or lower doses of poly ICLC ( 0 . 1 mg/kg ) induced increased CXCL10 levels , which were less sustained ( Figure 6B and 6C ) . We detected increased CXCL9 serum levels in animals injected with poly I:C or poly ICLC ( 0 . 5 mg/kg ) , and there were minor changes in CXCL9 concentrations in monkeys receiving KLH alone ( Figure 6D ) . No changes of CXCL9 serum concentrations were observed when 0 . 1 mg/kg poly ICLC were administered ( data not shown ) . At 6 , 24 , or 48 h after infection , no significant differences in serum levels of IFN-α , IFN-γ , TNF , IL-12p40 , and CCL3 ( MIP-1α ) were observed between groups receiving KLH alone or together with dsRNA ( data not shown ) . We were not able to detect considerable serum concentrations of IFN-α at any point in time including 1 h post injection . Since immunohistochemistry and in-situ hybridization revealed that CXCL10 was mainly produced in the T cell-areas of the draining lymph nodes ( Figure 4 ) , we considered DCs as a potential source for this chemokine in vivo . Unfortunately , double-labeling with DC identifying mAbs was not possible on formalin-fixed specimens . We therefore tested whether dsRNA may directly induce CXCL10 secretion by highly purified rhesus macaque DCs in vitro . When monocyte-derived monkey DCs were incubated with poly ICLC at two different concentrations ( 50 and 200 µg/ml ) , significantly elevated CXCL10 concentrations were detectable 48 h later in the cell culture supernatants ( p = 0 . 002 compared to un-stimulated controls ) , and both doses of of poly ICLC induced comparable levels of CXCL10 ( Figure 7 ) . Thus , primate DCs produce CXCL10 upon stimulation with synthetic dsRNA , making DCs one of the candidate sources of CXCL10 observed in the draining lymph nodes .
This study shows that s . c . injection of synthetic dsRNA , i . e . , poly I:C , poly ICLC , or poly I:C12U supports the induction of cellular immune responses to protein antigens in nonhuman primates . These responses could also be boosted by a second injection of antigen together with dsRNA . We observed antigen-specific T cell proliferation of CD3+CD4+ and CD3+CD4− T cells . High but nontoxic doses ( toxicity starts in M . mulatta at i . v . doses >2 mg/kg , i . m . or s . c . injections are better tolerated than i . v . injections; unpublished observations ) of poly ICLC ( 0 . 5 mg/kg or 2 mg/animal ) might be more potent than lower doses ( ≤0 . 1 mg/kg ) . Using HPV16 capsomeres at low doses ( 10 µg/animal ) as a relevant viral antigen with low immunogenicity , we also showed that poly ICLC , but not CpG-C ( which supported the induction of humoral responses , however ) , supports the induction of HPV16-specific Th1 responses . The lack of effect of CpG-C in our system compared to other studies where the same compound helped to elicit cellular immunity in nonhuman primates is most likely due to the fact that we injected the antigens in PBS , while others injected CpG-C and antigens in the synthetic water-in-oil emulsion , Montanide [49] . Amongst the three different formulations of synthetic dsRNA , poly ICLC appears to possess the most potent adjuvant activity on the induction of cellular immune responses . Subsequent studies will show whether it will help to induce protective immune responses against other pathogens , e . g . , SIV . Both adjuvants supported the induction of humoral immune responses , including neutralizing antibodies . Therefore , subsequent in vivo studies should compare poly ICLC with the adjuvants currently used in vaccine formulations , e . g . , alum , and investigate whether its co-application might allow fewer injections than required today for the currently licensed vaccine formulations . In order to understand the activity of dsRNA , we examined the innate response since this includes events that can improve the function of antigen presenting DCs and T cells . Surprisingly , we did not detect the expected increase of serum IFN-α shortly after injection of poly I:C or poly ICLC . This might be due to the s . c . route of injection . While i . v . injections of poly ICLC give rise to high serum interferon levels [30] , the s . c . application of dsRNA may lead to a more protracted release from the site of injection and a delayed bioavailability . In mice , type I interferon induced by poly I:C has been shown to be essential for its adjuvant effect on humoral immunity and isotype switching [59] , and it also seems essential for TLR3-mediated cross-priming of CD8+ T cells [60]–[62] . Likewise , type I interferon is critical for the CD8+ T cell expansion induced by TLR agonists in combination with CD40 [63] . Poly I:C and poly ICLC induce proliferation of CD8+ T cells , both have been shown to be effective as an adjuvant for the induction of specific CD8+ T cell responses in mice [64]–[66] , and this effect partially depends on NK cells [67] . Thus , poly I:C , and most likely also poly ICLC , support the induction of CD8+ T cell responses , and the KLH-specific responses expressed by CD3+CD4− T cells observed by us might reflect true CD8 responses . In contrast to our inability to detect IFN-α in the serum in response to dsRNA , we did detect enhanced levels of CXCL10 . These were sustained over 48 hours in animals injected with 0 . 5 mg/kg poly ICLC but decreased more rapidly in monkeys following injection of lower concentrations of poly ICLC , 0 . 5 mg/kg poly I:C , or a comparable dose ( 2 mg/animal ) of poly I:C12U . This may reflect the reduced biostability of the nonstabilized poly I:C and poly I:C12U compared with that of poly ICLC as described before [29] , [30] . CXCL10 is known for its activity to attract effector Th1 cells through interaction with its receptor CXCR3 at sites for the expression of Th1 immune responses [68] , e . g . , rejection of allografts or the inflammatory response upon mycobacterial infection [69] , [70] . CXCL10 is also required for resistance to protozoan or viral pathogens [71] , [72] . Studies in mice revealed additionally that CXCL10 is secreted early ( e . g . , earlier than CXCL9 , which we did not detect at the same levels in the serum as CXCL10 ) [73] , and stimulates T cell proliferation [74] . In fact , CXCL10-deficient mice have impaired T cell responses following primary immunization with exogeneous protein antigen indicating a role for CXCL10 in effector T cell generation [75] . Since CXCR3 also is induced early in CD4 T lymphocyte differentiation [76] , the literature suggests an enhancing role for CXCL10 in both the expression and induction of Th1 immune responses . Notably , while we have previously detected an increase in serum CXCL10 after injection of CpG-A or CpG-B [14] , these concentrations were around ten-fold lower than in animals injected with 0 . 5 mg/kg poly ICLC . Since both the two forms of CpGs and low doses of poly ICLC had only marginal adjuvant effects on the induction of cellular immunity , high and sustained serum levels of CXCL10 after injection of an adjuvant seem to be indicative of its ability to support the induction of cellular immune responses . Interestingly , the Th2-adjuvant alum considerably inhibits TLR-induced production of CXCL10 [77] . Expression of CXCL9 and CXCL10 was primarily in the T cell areas of the draining lymph node . Thus , DCs should be considered as a potential source of these chemokines , since they are abundant in this area of the lymph node . We show that monocyte-derived DCs produce CXCL10 upon activation with dsRNA , which suggests a direct role of these cells in the production of the pro-inflammatory chemokines and induction of cellular immune responses in our system . Monkey DCs express TLR3 ( manuscript in preparation ) and can be activated by poly I:C [58] , so pattern recognition receptors on DCs likely contribute to the observed adjuvant effects of dsRNA for CD4+ T-cell proliferation . Synthetic dsRNA , however , may also target and activate other TLR3+ or TLR3− leukocyte subsets . In vitro , it activates human NK cells [78] , [79] , γ/δ TCR+ T cells [80] , CD8+ α/β TCR+ T cells [81] , and also monocytes/macrophages , which are TLR3− [82] , [83] . These cells ( or the corresponding cells in lymphoid tissues ) could contribute to its adjuvant activity , e . g . , through the secretion of pro-inflammatory cytokines and notably type I and II interferons . While it remains to be determined whether dsRNA can promote survival of primate CD4+ T cells as recently shown for murine cells [84] , analyses of human blood leukocytes shortly after poly ICLC injection revealed increased percentages of CD4+ T cells , but also effects on the activity of NK cells and the frequency of HLA-DR+ cells [85] . Nevertheless , cells other than leukocytes including keratinocytes and neurons also can produce type I interferons and other pro-inflammatory cytokines upon stimulation with poly I:C [86] , [87] . After injection of dsRNA , we observed a down-regulation of the homeostatic chemokine CCL21 , which attracts CCR7+ cells , such as DCs and naïve T cells , to lymph nodes . In agreement with the advuvant effect of poly ICLC on the induction of HPV-specific Th1 immune responses shown in the present study , this process has recently been described for the early phase of the induction of Th1 but not Th2 immune responses in mice and is controlled by the production of IFN-γ [88] . This is mirrored by our findings using HPV16 capsomeres as viral protein antigen with relevance to the human system . Animals injected with HPV together with poly ICLC developed Th1 immune responses characterized by antigen-specific T cell proliferation and IFN-γ secretion in the absence of detectable IL-4 or IL-17 production . In conclusion , dsRNA compounds induce the innate production of CXCL10 in the draining lymph nodes and high CXCL10 concentrations in the serum early after injection , and these compounds are effective adjuvants for the induction of adaptive pathogen-specific T cell and humoral immune responses .
Healthy young adult male and female rhesus macaques ( Macaca mulatta ) housed at the German Primate Center ( Deutsches Primatenzentrum , Göttingen , Germany ) were used . The animals were antibody negative for simian T-lymphotropic virus type 1 , simian D-type retrovirus , and simian immunodeficiency virus . All animal care operations were in compliance with the guidelines of the German Primate Center and approved by the local authorities . For immunizations and collection of blood samples animals were sedated with ketamine . For the removal of lymph nodes , a deeper anesthesia consisting of a mixture of xylazine , atropine , and ketamine was used . 200 µg endotoxin-free KLH ( Calbiochem , San Diego , CA , USA ) or 10 µg HPV 16 capsomeres alone or in combination with either poly I:C ( Invivogen/Cayla , Toulouse , France ) , poly ICLC ( Hiltonol , Oncovir , Washington , D . C . ) , poly I:C12U ( Ampligen , Celldex Therapeutics , Bloomsbury , NJ , USA ) , or CpG-C ( ODN 2396 , generously provided by Coley Pharmaceutical Group , Wellesley , MA , USA ) were administered bilaterally s . c . at doses indicated in the text at volumes between 1 . 0 and 2 . 0 ml , partially diluted in PBS , by injecting close to the inguinal lymph nodes . All animals remained well following the application of dsRNA plus antigen and no local signs of inflammation apart from transient lymph node swellings and mils hyperemia were observed at sites of injection . Blood samples were drawn at 0 , 1 or 6 , 18 or 24 , and 48 h after injections for measurements of serum cytokine and chemokine concentrations . To determine humoral and cellular immune responses additional blood samples were drawn at points in time indicated . Axillary lymph nodes were removed before the immunizations , and 18 h after the injections one draining lymph node from each immunized animal was removed . Lymph nodes were divided in two parts . One part was fixed in 4% neutral-buffered formalin overnight and embedded in paraffin . The other halves were embedded in tissue-freezing medium ( Leica , Nussloch , Germany ) , snap-frozen in liquid nitrogen , and stored at −70°C until use . HPV16 L1 capsomeres were produced using recombinant baculoviruses containing the mutated L1 ( L1_2xCysM: C175A+C428A ) as described previously [56] . In short , High Five insect cells ( Invitrogen , Germany ) were infected with recombinant baculoviruses and harvested by centrifugation . Proteins were extracted by sonification from cell pellets resupended in 20 ml of extraction buffer ( 5 mM MgCl2 , 5 mM CaCl2 , 1 M NaCl , 0 . 01% Triton ×100 , 20 mM Hepes pH 7 . 4 and 1 mM PMSF ) . The cleared lysate was loaded on a two-step gradient consisting of ( 30% w/v ) sucrose and CsCl ( 58% w/v ) , followed by a centrifugation at 96 , 500 g at 10°C for 3 h in a SW32 rotor ( Beckman Ultracentrifuge ) . The interphase between the sucrose and CsCl and the complete CsCl layer were centrifuged again in Quickseal tubes ( Beckman , USA ) for 16–18 h at 20°C at 184 , 000 g in a Sorval TFT 65 . 13 rotor . Fractions of 1 ml fractions were collected and the L1 containing determined by antigen-capture ELISA and western blot analysis and the structure of the particles was characterized by electron microscopy [89] . The control antigen ( mouse norovirus VP1 VLPs ) were generated by the identical protocol . The VP1 clone was kindly provided by W . Nicklas , DKFZ Heidelberg . Standard proliferation assays were set up with 1×105 PBMCs/well in 96-well round-bottom trays ( Nunc ) with KLH ( 100 µg/ml ) in cell culture medium consisting of RPMI 1640 , supplemented with 2 mM L-glutamine , penicillin ( 100 U/ml ) -streptomycin ( 100 µg/ml ) , 10 mM HEPES ( all GIBCO , Invitrogen , Karlsruhe , Germany ) , 50 µM 2-mercaptoethanol ( Sigma ) , and 10% heat-inactivated FCS ( Biochrom , Berlin , Germany ) . Controls included PBMCs in medium alone and PBMCs stimulated with 5 ng/ml staphylococcal enterotoxin B ( SEB; Alexis Corp . , Lausen , Switzerland ) . All conditions were set up in triplicates and cultures were incubated at 37°C and 5% CO2 . Supernatants were harvested on day 2 and frozen at −80°C for analyses of cytokine concentrations . 3H-thymidine ( 1 µCi/well , NEN , Perkin Elmer , Boston , MA , USA ) was added to the wells on day 3 ( for SEB and medium alone ) or day 5 ( KLH and medium alone ) . Cells were harvested 24 h later onto glass fibre filter mats ( ICN Biomedicals , Aurora , OH , USA ) , and incorporated 3H-thymidine was measured in a liquid scintillation counter . To facilitate the comparison of proliferative responses , SIs were calculated by dividing the mean counts per minute ( cpm ) of triplicates of antigen-containing wells by the mean cpm of triplicate wells with unstimulated PBMCs . Additionally , CFSE ( Invitrogen/Molecular Probes , Karlsruhe , Germany ) assays were used to determine proliferation . PBMCs at 1×107 cells/ml were stained with 0 . 25 µM CFSE in pre-warmed PBS for 15 min at 37°C , washed in medium , incubated in pre-warmed medium for another 30 min , and washed again . The cells were then adjusted to 1×106 cells/ml and cultured in medium with or without SEB or KLH as described above or and incubated for 6 to 7 days . Alternatively , cells were incubated at 1 . 25 µg/ml with HPV16 VLPs or an unrelated control antigen , i . e . , mouse norovirus VLPs similarly produced as the HPV antigen ( A447 ) , at the same dose . After 7 days cells were harvested and washed in PBS/5% FCS/0 . 05% sodium azide , stained with anti-CD3 PerCP- and anti-CD4 APC-conjugated mAbs , washed , and fixed . T cell proliferation was assessed as the percentage of CFSElow cells , gating on live CD3+CD4+ or CD3+CD4− cells ( Figure 1A ) . Alternatively , cells were re-stimulated with eight pools of HPV16-specific , 15mer peptides ( 124 peptides , pool 1–4 with 16 peptides each , pool 5–8 with 15 peptides each ) , 2 µg/ml SEB , or medium alone in the presence of 1 µg/ml co-stimulatory mAbs CD28 and CD49d ( BD Pharmingen ) for 6 h , and Brefeldin A ( Sigma ) was added at a final concentration of 10 µg/ml for the last 4 . 5 h . Cells were then washed in PBS/5% FCS/0 . 05% sodium azide , stained with anti-CD3 PerCP- and anti-CD8 APC-conjugated mAbs , washed , fixed with 4% paraformaldehyde , and stained with PE-conjugated mAbs against IFN-γ after cell permeabilization with 0 . 5% saponin in PBS/5% FCS/0 . 05% sodium azide . T cell proliferation was assessed as the percentage of CFSElow cells , gating on live CD3+CD8+ or CD3+CD8− cells , and IFN-γ secretion was measured as the percentage of PE-stained , CFSElow cells in the gated cell populations . ELISPOT assays were preformed using commercially available reagents ( Mabtech AB , Hamburg , Germany ) as previously described [90] . Briefly , PBMCs were resuspended in culture medium and seeded at 1×105 cells/well in 96-well plates ( MAIP S4510 , Millipore , Schwalbach , Germany ) , which had been coated with 1 µg/well of anti-human IFN-γ monoclonal antibody overnight at 4°C . For antigen stimulation , HPV16 L1 protein or control antigen ( A447 ) was added at 1 . 25 µg/ml to the wells in triplicates . Positive and negative controls consisted of cells stimulated by SEB ( 1 µg/ml , Sigma ) and cells kept in medium alone . After 20 h of incubation at 37°C in 5% CO2 , cells were removed and biotinylated anti-human IFN-γ detector antibody was added ( 0 . 1 µg/well ) , followed by the addition of streptavidin-alkaline phosphatase conjugate at 1∶1000 in PBS/0 . 1% FBS . Spots were developed with NBT/BCIP solution ( 25 µg NBT and 15 µg BCIP in 0 . 1 M Tris–HCl pH 9 . 5 per well ) for 30 min , the wells were washed with distilled water and air-dried , and spots were counted using a BIOSYS2000 ELISPOT reader . The counts were extrapolated to 106 PBMCs . Average spot numbers of background responses ( to A447 ) plus twice the standard deviation were considered positive responses . The presence of L1-specific IgG antibodies in plasma samples from immunized monkeys was determined by VLP-ELISA as described earlier [56] . Briefly , 96-well plastic plates were coated overnight at 4°C with VLP produced in baculovirus infected High Five insect cells and purified according to a previously published method [89] . After washing with PBS-T , plates were blocked with MPBS-T ( 5% skim milk in PBS- 0–05% Tween ) for 1 hr at 37°C . Prediluted sera ( in two-fold dilutions starting from 1∶50 to 1∶819 , 200 ) were added , and plates were incubated for 1 hr at 37°C . After washing , plates were incubated for 1 hr at 37°C with 1∶2000 diluted HRP-coupled antihuman IgG F ( ab' ) 2 secondary antibody ( Dianova , Germany ) in MPBS-T , TMB ( 3 , 3′ , 5 , 5 -tetramethylbenzidine ) substrate solution ( Sigma , Germany ) was used as substrate . OD was measured in an ELISA reader at 450 nm after 10 min and 30 min incubation at room temperature . Nonspecific binding was determined by using the same dilutions on plates coated with extracts of High Five cells infected with wt baculovirus . IgG titers were expressed as the reciprocal of the highest dilution giving an absorbance above the cut off value ( the average of the negative controls plus three times standard deviation ) . Pseudovirions were prepared by transfecting 293TT cells ( cultivated in DMEM containing 50 µg of hygromycin/ml ) with a plasmid coding for the humanized HPV16 L1 and L2 genes , together with a plasmid containing the gene for secretable alkaline phosphatase ( SEAP ) under the control of the CMV promoter . For pseudovirion extraction , cells were harvested 3–4 days later by trypsination , washed once with PBS and resuspended in 1 ml PBS containing 1 mM CaCl2 and 5 . 6 mM MgCl2 per 5×107 cells and lysed by 50 µl Brij58 ( Sigma ) in the presence of Benzonase ( 250 U/ml ) for 5 min on ice . The cellular lysate was centrifuged after the addition of NaCl to a final concentration of 710 mM , and the cleared supernatant containing the pseudovirions was used for infection of 293TT cells . For this purpose , pseudovirions were diluted 1∶5000 in DMEM and preincubated with the sera ( dilution from 1∶50 to 1∶100 , 000 ) for 15 min at room temperature . Pseudovirions were then added to the cells , followed by incubation at 37°C for 5 days . Detection of SEAP activity in cell culture supernatant was measured by using a commercial assay ( Roche , Mannheim , Germany ) according to the manufacturer's recommendations . Chemokine and cytokine concentrations in serum or plasma samples and cell culture supernatants were measured using ELISA kits for human CXCL10 , CXCL9 ( both R&D Systems , Wiesbaden , Germany ) , CCL3 ( Antigenix America , Huntington , NY , USA ) , IFN-α ( PBL , Brunswick , USA ) [14] , IL-17 ( eBioscience , NatuTec , Frankfurt/Main , Germany ) [57] , and human TNF as well as monkey IFN-γ , IL-4 , and IL-12p40 ( all U-Cytech , Utrecht , The Netherlands ) . Cryostat sections were cut , fixed in acetone for 30 min and incubated with monoclonal antibodies against human CD1a ( dilution: 1∶100; Medac , Hamburg , Germany ) , CD83 ( dilution: 1∶100 ) or CD208 ( 1∶70; both Immunotech , Hamburg Germany ) . Antibody binding was visualized by the alkaline phosphatase anti-alkaline phosphatase method using New Fuchsin as chromogen . The sections were counterstained with hemalaun and mounted . The numbers of DCs were quantified with a Zeiss AxioImager M1 microscope ( Carl Zeiss , Jena , Germany ) . Using a 40× objective , a standard area was set ( unit area ) . Ten non-overlapping unit areas were selected . The positive cells were counted using AxioVision ( Release 4 . 6 ) software ( Zeiss ) . The values were averaged to represent the numbers of positive cells per unit area . Due to inadequate immunohistochemical staining the draining lymph node from animal number 13408 was omitted from the examination . Immunohistochemistry on paraffin sections was performed as previously described [92] . The following antibodies diluted in antibody diluent ( S3022 , DAKO , Glostrup , Denmark ) were used: mouse anti-CXCL10 ( MAB266 , R&D Systems , 1 µg/ml ) , goat anti-CXCL9 ( AF392 , R&D Systems , 1 µg/ml ) , and goat anti-CCL21 ( AF366 , R&D Systems , 1 µg/ml ) . After over night incubation , sections were washed and incubated with rabbit anti-mouse ( E0413 , DAKO ) or rabbit anti-goat ( E0466 , DAKO , ) biotinylated antibodies followed by streptavidin-alkalyne phosphatase complex ( K0391 , DAKO ) , following the manufacturer's instructions . Positive cells were detected using New Fuchsin ( K0698 , DAKO ) as substrate , and tissue sections counterstained with Meyer's Haematoxylin ( 1 . 09249 , Merck , Zug , Switzerland ) . 35S-labeled sense and antisense CXCL9 , CXCL10 , and CCL21 mRNA probes , 411 bp in length corresponding to position 26 to 437 of the CXCL9 sequence ( NM_002416 ) , 372 bp corresponding to position 28 to 400 of the CXCL10 sequence ( NM_001565 ) , and 367 bp corresponding to position 27 to 394 of the CCL21 sequence ( NM_002989 ) , respectively , were generated by in vitro transcription ( Roche Molecular Biochemicals , Indianapolis , IN ) . Tissue sections were dewaxed , rehydrated in graded ethanol solutions , and subjected to in situ hybridization , according to a previously described method [93] . Finally , the sections were dipped in photo emulsion NTB-2 ( Kodak , Rochester , NY ) and exposed in complete darkness for 2 to 4 weeks at 4°C . Development and fixation were performed according to the instructions provided by Kodak , and counterstaining was done with haematoxylin . Rhesus macaque monocyte-derived DCs were generated from heparinized peripheral blood as previously described [57] . CD14+ monocytes were magnetically separated ( Miltenyi Biotec , Bergisch-Gladbach , Germany ) and cultured at 1 . 5–2×106 cells/3 ml in RPMI 1640 , supplemented with 5% human AB serum ( PAN Biotech , Aidenbach , Germany ) , human rGM-CSF ( 1000 U/ml , sargramostim , Leukine , Berlex , Richmond , CA , USA ) , human rIL-4 ( 100 U/ml , R&D Systems , Wiesbaden-Nordenstadt , Germany , and L-glutamine , 2-mercaptoethanol , HEPES , and penicillin-streptomycin as described under T cell assays . At day 6 , DCs at 1×105/well were stimulated for 48 h with 50 or 200 µg/ml poly ICLC in 96-well round bottom plates . Supernatants were harvested for analysis of cytokine and chemokine secretion . Data are expressed as means±standard error of the mean ( SEM ) , standard deviation ( SD ) , or median , where appropriate . Statistical significance of differences was determined by Student's t-test or Mann Whitney U-test . Differences were considered statistically significant for p<0 . 05 .
|
Novel adjuvants that facilitate the induction of strong cellular immunity could be of help in the design of vaccine strategies to combat infections such as HIV or tuberculosis . Our immune cells possess archaic receptors recognizing structures of infectious pathogens , and the interaction of these receptors with their ligands results in an activation of the immune system . Here we exploited synthetic forms of one of these ligands , i . e . , dsRNA , to define an adjuvant for the induction of cellular immune responses in primates . We injected model and viral proteins together with three different forms of dsRNA subcutaneously ( s . c . ) in rhesus macaques , and all compounds served as adjuvants for the induction of cellular immunity without the incidence of major side effects . These adjuvant effects depended on the adjuvant dose and coincided with profound alterations in the chemokine production in the draining lymph nodes . dsRNA also helped to induce cellular and humoral immune responses against capsomeres of low immunogenicity derived from the human papillomavirus 16 , the causative agent in about 50% of all cases of cervical cancer worldwide . Therefore , formulations involving synthetic dsRNA are promising candidates for development of novel vaccines .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"immunology/immunity",
"to",
"infections",
"virology/vaccines"
] |
2009
|
Synthetic Double-Stranded RNAs Are Adjuvants for the Induction of T Helper 1 and Humoral Immune Responses to Human Papillomavirus in Rhesus Macaques
|
Canine rabies is one of the most important and feared zoonotic diseases in the world . In some regions rabies elimination is being successfully coordinated , whereas in others rabies is endemic and continues to spread to uninfected areas . As epidemics emerge , both accepted and contentious control methods are used , as questions remain over the most effective strategy to eliminate rabies . The Indonesian island of Bali was rabies-free until 2008 when an epidemic in domestic dogs began , resulting in the deaths of over 100 people . Here we analyze data from the epidemic and compare the effectiveness of control methods at eliminating rabies . Using data from Bali , we estimated the basic reproductive number , R0 , of rabies in dogs , to be ∼1·2 , almost identical to that obtained in ten–fold less dense dog populations and suggesting rabies will not be effectively controlled by reducing dog density . We then developed a model to compare options for mass dog vaccination . Comprehensive high coverage was the single most important factor for achieving elimination , with omission of even small areas ( <0 . 5% of the dog population ) jeopardizing success . Parameterizing the model with data from the 2010 and 2011 vaccination campaigns , we show that a comprehensive high coverage campaign in 2012 would likely result in elimination , saving ∼550 human lives and ∼$15 million in prophylaxis costs over the next ten years . The elimination of rabies from Bali will not be achieved through achievable reductions in dog density . To ensure elimination , concerted high coverage , repeated , mass dog vaccination campaigns are necessary and the cooperation of all regions of the island is critical . Momentum is building towards development of a strategy for the global elimination of canine rabies , and this study offers valuable new insights about the dynamics and control of this disease , with immediate practical relevance .
Rabies transmitted by domestic dogs is a re–emerging public health problem in Asia . In recent years incidence has increased dramatically in China [1] , [2]; multiple incursions have been reported from Bhutan [3] , [4]; and the disease has spread across several previously rabies–free islands in Indonesia ( Flores 1997 [5] , Maluku 2003 , North Maluku 2005 , West Kalimantan 2005 , Nias 2009 [6] ) , including the popular tourist destination of Bali [7] . The island province of Bali was historically rabies–free until late 2008 , when several local people died in the southernmost peninsula showing signs of the disease . An incursion is thought to have occurred approximately seven months earlier , when a fisherman landed on the peninsula with a dog that was incubating the virus [8] . Initial control efforts by the Balinese government attempted to contain the outbreak to the two administrative districts ( Regencies ) within the peninsula . However , in August 2009 a human case was diagnosed beyond the outbreak locality , and by July 2010 cases had been confirmed in all nine Regencies of Bali and 62 people had died ( Fig . 1 ) . As is common with an unexpected incursion: the island lacked surveillance , medical staff trained in rabies diagnosis , and contingency planning . The ensuing epidemic generated local and international pressure to eradicate rabies and led to plans for island–wide mass vaccination of the dog population ( ProMED-mail archive number 20100806 . 2673 ) . The government's main concern for the effectiveness of any proposed rabies control programme on Bali was the high density of domestic dogs . Dogs are an important part of Balinese culture; the majority of households own at least one dog [9] , though most are unconfined and not easy to restrain for parenteral vaccination . However , a pilot vaccination campaign that used trained dog–catchers equipped with nets showed that more than 80% of dogs could be vaccinated [10] , with a team of six vaccinating around 100 dogs per day ( Fig . 1A ) . From initial estimates of the human∶dog ratio ( 8∶1 ) the Bali dog population was extrapolated to be 400 , 000 , with densities exceeding 250 km−2 in urban areas [11] . The basic reproductive number , R0 , measures the average number of secondary cases arising from a primary infected individual in an otherwise fully susceptible population , and determines the critical level of vaccination coverage needed to protect the population ( ‘herd immunity’ ) and bring a disease under control [12] . For directly transmitted diseases such as rabies , R0 is often assumed to depend on population density [12] , implying that such high–density dog populations could limit the success of mass vaccination . Estimating R0 for rabies on Bali was therefore a priority for determining whether vaccination would be a feasible control strategy and for setting coverage targets . The relationship between dog rabies incidence and human rabies deaths was a further important consideration for estimating public health impacts of proposed strategies . Considerable successes have been achieved in the control of rabies in many parts of the world through the mass vaccination of domestic dogs [13] , [14] , [15] , [16] and mounting evidence demonstrates that regional elimination of canine rabies is possible through sustained annual campaigns that attain 70% coverage [17] , [18] . However , there are no operational guidelines on how to roll out dog vaccination campaigns strategically in the face of an emerging epidemic . We developed a model to capture the inherent variation in epidemic trajectories , particularly as eradication is approached , to guide strategic choices in planning Bali's first island–wide mass vaccination campaign . We fitted dog rabies incidence to human deaths in order to link the model output to potential human deaths averted . The model addressed concerns over the extremely dense population of dogs and presumed high levels of dog population turnover . We used the model to investigate whether vaccination campaigns that reach 70% of dogs on Bali could provide herd immunity , and how many campaigns would be needed to achieve eradication . We investigated how campaign effectiveness might be affected by use of locally–produced ( potentially more affordable and sustainable ) vaccines versus longer–acting , imported vaccines , by the speed of delivery and strategic rolling out of the programme across Bali and by the interval between campaigns . Then we examined how robust campaign performance would be to human–mediated movement of dogs around Bali and heterogeneities in coverage arising from political , logistical and operational constraints . Finally we explored the impacts of the vaccination campaigns that have since been implemented on Bali and their prospects for achieving eradication , and provide advice for how these prospects may be enhanced .
Figure 2 provides a visual summary of the model of dog-dog transmission and spread across Bali , as well as the functional form used to predict human rabies cases . We assumed that each infectious dog case causes κ secondary dog cases ( ‘offspring’ ) , drawn from a negative binomial distribution ( κ∼negative binomial ( R0 , k ) , Table 1 , Fig . 2Ai ) , with R0 as its mean [20] , [21] . Each secondary case was assigned a generation interval selected from a gamma distribution [18] ( Table 1 , Fig . 2Aii ) representing an incubation period plus a period of infection prior to transmission , to determine when new infections were generated . Using an explicit spatial representation of Bali based on 1 km2 grid cells ( Fig . 2A ) , we probabilistically allocated the location of each secondary case . To capture human–mediated transport of dogs across the island , exposed offspring were assigned to a randomly chosen grid cell with probability p so that infected dogs could potentially travel much further distances than a rabid dog is capable of running . To capture the local movement of rabid dogs , secondary cases were displaced from their direct epidemiological predecessors according to a gamma–distributed dispersal kernel [18] ( Table 1 , Fig . 2Aiii ) , with probability 1–p . We estimated the initial epidemic growth rate λ from the monthly time series of confirmed dog rabies cases ( Fig . 1A ) using a generalized linear model with negative binomial errors [18] . We converted the inferred initial epidemic growth rate to an estimate of R0 using the probability distribution function of the generation interval ( Gt ) for rabies based on data from natural infections [18] , according to Wallinga & Lipstich [22]: The R0 estimate for Bali was used as the mean of the offspring distribution in the model ( Fig . 2Ai ) . To estimate the relationship between confirmed dog rabies cases and human deaths , we fitted several functions using maximum likelihood and used AIC to select the best fitting model ( Fig . 2Av ) . The probability of human-mediated transport of dogs across the island , p , was inferred by incrementally increasing the proportion of rabid and incubating dogs that were moved randomly across the island until the modelled speed of spread matched the observed spread of the epidemic ( Fig . 2Aiv , assuming a case detection probability of 0 . 07 [6] ) . Other parameters used in the model ( Table 1 , Fig . 2 ) were derived from epidemiological data on naturally infected rabid dogs in Tanzania [18] . We modeled vaccination coverage ( the proportion of dogs vaccinated , V ) in each cell as waning exponentially from the coverage achieved at the time of vaccination , at a rate ( Δt = one day ) determined by dog population turnover ( b = birth rate and death rate , assuming constant population size ) and the duration of the vaccine–induced immunity ( τ , where v = 1/τ ) :Parameter estimates are provided in Table 1 . We made the conservative assumption that coverage did not accumulate over multiple vaccination campaigns ( see Supporting Information for more details ) . Dog vaccination is represented in the model by reducing the number of secondary cases per primary infection in direct proportion to vaccination coverage at the time of transmission . In effect , each potential secondary case becomes infectious with probability 1−Vt , so in a vaccinated population the number of secondary cases attributed to each case is κv∼binomial ( κ , Vt ) . The branching process formulation does not account for any effects of depletion of the susceptible population as disease incidence increases . However , since detected incidence on Bali did not exceed 0 . 2% per annum , depletion of the susceptible population is assumed to play a negligible part . Likewise , we did not include the effects of rabies incidence on the proportion of dogs vaccinated . The island–wide mass vaccinations on Bali began in October 2010 by which time 477 cases of rabies had been confirmed in dogs . We suspect that samples were retrieved from less than 10% of rabies cases ( based on [6] , [23] and previous experience during intensive contact tracing studies in northern Tanzania that suggest samples are recovered from around 5–10% of identified cases ) , therefore we commence vaccination in the model after 7 , 000 cases had occurred in model realizations . We assume the vaccinations failed to eradicate rabies if 40 , 000 cases were reached . The expected behavior of the epidemic under alternative scenarios was estimated using two measures: 1 ) the probability of eradication of rabies from Bali , and 2 ) the time to eradication from the onset of vaccination . For each scenario we ran 1 , 000 realizations of the model . Statistical analyses were carried out in R ( version 2 . 14 . 2 , R Core Team 2012 ) and the model was built in MATLAB ( version 7 release 14 , The MathWorks Inc . ) . Codes are available upon request to the corresponding author . We explored the sensitivity of performance measures to variation in R0 ( between 1 and 2 based on estimates of R0 from rabies outbreaks around the world [18] ) , vaccination coverage; domestic dog population turnover ( assuming constant population size and birth/death rates varying from 0 . 1 to 2 . 3 year−1 spanning a range of population replacement from 10% to 90% per year ) ; duration of vaccine–induced immunity; and variation in long-distance dog movement , to investigate the potential impact of restrictions on human–mediated transport of exposed or infected dogs . The island grid was aggregated into 24 rectangular blocks of similar size ( mean 277 km2 , range 49–500 km2 ) to evaluate strategies . We analyzed repeat campaigns ( 1 , 2 and 3 campaigns ) under a range of coverage levels ( 40% , 60% or 80% ) and inter–campaign intervals ( 0 , 6 or 12 months ) . We considered one synchronous campaign vaccinating all 24 blocks in the same month ( A in Table 2 ) , four proactive strategies each of six–month duration vaccinating four blocks each month in different sequences ( random , rotate , source and furthest , B–E in Table 2 ) and two reactive strategies of six–month duration ( F–G in Table 2 ) . To examine the impact of heterogeneity in vaccination coverage we compared the effect of leaving unvaccinated areas distributed across the island in two ways: either randomly distributed unvaccinated 1 km2 grid cells , or equivalently–sized contiguous blocks of unvaccinated grid cells . Videos of model simulations of a sample of the scenarios we considered are available as Supporting Information . To estimate vaccination coverages achieved in Bali , data on vaccination dates , numbers of dogs vaccinated and post-vaccination surveys ( counts of dogs with or without collars signifying vaccination ) were compiled at the banjar ( sub-village ) level , where possible . Where data were only available at courser resolution , numbers of dogs vaccinated were split between corresponding villages and banjars . Dog population size was calculated from post-vaccination surveys in banjars as: dogs vaccinated/ ( collared dog count/total dogs counted ) . If surveys were not available , dog populations were estimated from the human∶dog ratio for the village , district or regency as available . To obtain vaccination coverages by 1 km2 grid cell , banjar centroids were assigned randomly within their village polygon , and coverage averaged from banjar centroids within the grid cell or , if empty , assigned from the nearest banjar centroid . We assumed lakes , reservoirs , forested areas and mountain peaks were not inhabited by dogs . Coverage was assumed to wane as described above , and epidemic trajectories were simulated across the resulting dynamic coverage landscape .
We observed an exponential relationship between modeled R0 and the median time to rabies eradication ( Fig . 3A ) . Above a threshold value ( R0 between 1 . 3 and 1 . 4 ) , the probability of eradication fell to below one even for annual campaigns that achieved 70% coverage ( Fig . 3A ) . When R0 was equal to 1 . 2 , vaccination programmes with annual campaigns eventually eradicated rabies if coverage targets of at least 40% were met ( Fig . 3B , Fig . 4 ) . If campaigns achieved the WHO–recommended target of 70% coverage , the probability of eradication was largely insensitive to population turnover and duration of vaccine–induced immunity . Only at the highest turnover rates ( >70% ) and shortest vaccination immunity durations ( <1 year ) was the time to eradication substantially prolonged ( Fig . 3C&D ) . The number of consecutive island–wide annual campaigns and coverage achieved strongly influenced the probability of eradicating rabies ( Fig . 5A ) . A single high coverage ( 80% ) campaign did not guarantee eradication , but had a reasonable probability of success ( ∼0 . 6 ) , whereas a single 40% or 60% coverage campaign had no prospect of achieving eradication ( Fig . 5Ai ) . Subsequent campaigns greatly increased eradication prospects: two campaigns of 80% coverage or three campaigns of 60% coverage eradicated rabies in more than 90% of model runs , but three consecutive low coverage ( 40% ) campaigns still had a very low prospect of achieving eradication ( Fig . 5Aii & iii ) . Six consecutive low coverage campaigns increased the likelihood of eradication to ∼90% ( Fig . 3B ) . Thus , a roughly equivalent reasonable chance of eradication ( ∼90% ) can be achieved with a two high coverage ( 80% ) , three annual moderate coverage ( 60% ) or six annual low coverage ( 40% ) campaigns . Increasing campaign frequency did not greatly affect the probability of eradication , but annual campaigns of six–month duration with six–month inter–campaign intervals could be slightly more effective than back–to–back campaigns ( Fig . 5Aii & iii ) . Based on the pilot vaccinations ( Fig . 1A ) , it was estimated that the methods used could be feasibly scaled–up to cover the entire island within a six–month period , but more intensive vaccinations ( 1–month synchronized ) might compromise coverage because of insufficient availability of trained teams . Completing campaigns in six months rather than one month ( ‘sync’ ) delayed eradication by a few months , but these delays could be compensated for by a small increase in coverage ( Fig . 5C ) . Therefore on the basis of six-month long campaigns , we compared strategies for how to vaccinate the island , based upon different patterns of rollout under consideration at the time of planning the first campaign ( Table 2 B–E ) . Time to eradication under different strategies varied depending on the spatial evenness of cases and thus was sensitive to potentially long distance , human–mediated transport of dogs ( Fig . 5B ) . When human-mediated dog movement was restricted or at low frequency ( p = 0 and 0 . 02 , Fig . 5B ) cases were less evenly distributed and the strategy that most rapidly eradicated rabies started vaccinations in the southernmost Regency where the index case occurred ( ‘source’ ) . In contrast , the strategy that ended in the South ( ‘rotate’ ) took longest and the random strategy and the wave–like strategy from West to East ( ‘furthest’ ) were intermediate in performance . When human-mediated dog movement was frequent ( p = 0 . 05 , Fig . 5B ) all four strategies performed similarly . We also compared two six-month reactive strategies ( Table 2 F–G ) : the strategy that vaccinated blocks solely based on incidence ( ‘reactive’ ) , produced the most variation in eradication times ( Fig . 5B ) . This strategy eradicated rabies more rapidly than all others , including the synchronized campaign , when there was no human-mediated dog movement , but took longest when human–mediated movement was frequent ( p = 0 . 05 ) . The performance of the reactive strategy that did not return to previously vaccinated blocks within the same campaign ( ‘react w/o repeat’ ) was more robust to long distance movement ( Fig . 5B ) . We looked at the probability of eradicating rabies when there were gaps in coverage and under the scenarios of low and high frequency human-mediated dog movement where dogs could potentially be transported to any point on the island . When human-mediated dog movement was relatively low ( p = 0 . 02 ) , and gaps were modelled by excluding randomly distributed 1 km2 grid cells during vaccinations , the effect on the probability of eradication was negligible if the total area omitted was less than ∼10% of the island ( Fig . 5D ) and declined in a roughly linear fashion , reaching 0 . 9 when ∼20% of the island was not vaccinated ( Fig . 5D ) . In contrast , when the same proportion of unvaccinated cells were left in contiguous blocks , the probability of eradication dropped rapidly , reaching 0 . 9 when just 0 . 4% of the island's area was omitted , which equates to just three neighboring villages of Bali's ∼700 villages ( Fig . 5D ) . In both situations , the probability of eradication reaches zero when ∼50% of the island's area is left unvaccinated , but the decline is exponential when unvaccinated grid cells are aggregated ( Fig . 5D ) . More frequent human-mediated dog movement ( p = 0 . 05 ) amplifies the effects of gaps in coverage on the probability of eradication , with a greater chance of rabies reaching and persisting in unvaccinated areas ( Fig . 5D ) . Incorporation of all recorded vaccination efforts on Bali was necessary to generate simulated epidemics that matched the observed epidemic trajectory ( Fig . 6 ) . This included initial localized low coverage vaccinations using locally produced vaccines that required 3-month boosters which nevertheless played an important role in building up coverage and slowing the momentum of the epidemic ( Fig . 6 ) . Control was subsequently achieved through improving the scale , coverage and orchestration of vaccination , including switching to a longer lasting vaccine ( Fig . 1A ) : in late 2010 and early 2011 the first island-wide campaign achieved target coverages of 70% , although because the campaign took several months to implement , the average island-wide coverage was around 40% ( with ongoing turnover and waning immunity continually eroding coverage , Fig . 4 ) . A second campaign was completed later in 2011 building up island-wide coverage to around 60% ( Fig . 6 ) . The overall trajectory towards eradication appears very promising especially if gaps are addressed in a third campaign currently underway ( Fig . 5Aiv & 6 ) . However , if control measures lapse , there is a more than 30% chance that within three years rabies will re-emerge to an endemic situation ( Fig . 5Aiv & 6 ) with around 55 human deaths per year occurring on the basis of the relationship between confirmed cases and human deaths ( Fig . 2Av ) . Over a ten-year time horizon , under the best-case scenario of rapid eradication from Bali as a result of a 3rd comprehensive coverage vaccination campaign , approximately 550 human rabies deaths would therefore be averted in contrast to the endemic situation . Whereas if control measures are maintained , but not to the level required for eradication , low levels of rabies persistence would avert around 440 human rabies deaths but would require indefinite administration of expensive post-exposure prophylaxes ( ∼$1 . 5 million/year ) . These calculations assume awareness of rabies and the availability of PEP remain the same as over the course of the epidemic to date .
There are strong incentives for carrying out a mass dog vaccination programme to eradicate rabies from Bali . More than 100 human deaths have occurred since the start of the outbreak in 2008 [24] . Costs for the provision of post–exposure vaccine to bite victims in 2010 alone exceeded USD$2 million and would remain high in an endemic situation . If rabies was eradicated by mass dog vaccination , and assuming bite incidence returns to pre-outbreak levels ( one tenth those in 2010 ) , then precautionary use of post–exposure vaccine would also be ten–fold lower ( ∼100 , 000 USD per year ) . Following official declaration of freedom from rabies ( 2 years with no detected cases under effective surveillance [25] ) these costs should reduce to zero . Our results suggest that eradicating rabies from Bali through mass dog vaccination is feasible; it would prevent hundreds of human rabies deaths , save millions of dollars , alleviate the trauma and panic that is currently widespread in local communities , and mitigate potential impacts on Bali's tourist industry . We investigated operational aspects of vaccination strategies to determine which are most critical to achieving eradication rapidly . Our R0 estimate of 1 . 2 for rabies in Bali is remarkably similar to estimates for canine rabies elsewhere , which range from 1 to 2 [18] , despite population densities varying by an order of magnitude . Even under a range of assumptions about the timing and extent of reactive control measures following confirmation of rabies on Bali , R0 remains between 1 and 2 . Indeed , improvements in surveillance on Bali during the first year of the epidemic would likely lead to R0 being overestimated rather than underestimated . The low R0 observed on Bali challenges assumptions that canine rabies transmission depends on population density [12] , [17] . The relationship between R0 and density is in many ways parallel to the functional responses in predator prey interactions in population ecology . Borrowing existing concepts from population ecology helps to embed epidemiological phenomena in a different context , and may be helpful in understanding possible mechanisms underlying this relationship . The ( much studied ) mechanisms underlying Type 2 functional responses in predator prey interactions would be an obvious starting point suggested by the analogy . While further investigation is required to understand this phenomenon , our results suggest that moderate reductions in dog density are unlikely to have any beneficial effects on rabies control . Dog population management is often a common component of rabies control programmes , either exclusively or in combination with dog vaccination . Such programmes should be aware that the mass culling or sterilisation of dogs may not be an effective means of controlling rabies , and that as long as a high proportion of the dog population can be reached with vaccination , rabies should be brought under control . The sensitivity of vaccination success to R0 ( Fig . 3A ) highlights the importance of estimating R0 locally and accurately and the need to prioritize surveillance including collection of incidence data . Overall , the low R0 suggests that only 17% of the population would need to be vaccinated to control rabies ( Pcrit = 1−1/R0 ) [12] , [17] . However , when realistic operational features are taken into account , particularly the pulsed nature of vaccination campaigns , and the birth of susceptible dogs , we find that coverage of less than 30% may never achieve eradication ( Fig . 3B ) . At least 40% of dogs must be vaccinated to maintain island–wide coverage above 17% at all times ( Fig . 4 ) and consecutive annual campaigns are needed to ensure eradication given the stochastic nature of rabies spread ( Figs . 3 & 4 ) . With annual comprehensive vaccinations achieving uniformly high coverage of at least 70% as recommended by WHO [12] , [17] we would expect rabies to be eradicated from Bali within 1–3 years of initiating comprehensive vaccinations ( Fig . 3B ) . While we find that achieving high vaccination coverage is a decisive factor for disease elimination , follow–up campaigns are essential for achieving eradication , especially when achieving high coverage is problematic . At lower coverage , rapid population turnover and use of vaccines that confer only short–lived immunity could cause population–level protection to fall below Pcrit and reduce or preclude the chance of eradication ( Fig . 4 ) . Therefore use of long–acting vaccines particularly in populations with high turnover is recommended ( Fig . 3C&D ) . We found a positive effect of six–month intervals between campaigns ( Fig . 5Aii ) probably because coverage levels were maintained above Pcrit for longer than with equivalent effort in back–to–back campaigns [26] . Our results highlight that a successful vaccination programme requires comprehensive and even coverage . Missing randomly distributed small pockets ( totaling <10% of the total area ) may not be overly detrimental , but omitting an equivalent contiguous area such as an administrative unit , could jeopardize an entire programme . Hence , mass vaccination programmes which are not perfectly implemented everywhere are of less concern than lack of participation from all communities . High intensity mass vaccinations conducted over short periods that eradicated rabies from other regions [27] raised concerns about the need to complete campaigns on Bali as rapidly as possible . Our findings suggest taking longer to vaccinate a population ( six–months versus one–month ) has little impact on the success of otherwise equivalent campaigns , thus easing considerably the otherwise daunting logistical and financial challenges of synchronized mass vaccination campaigns [28] . In practice , increasing the speed with which a campaign is delivered might result in trade–offs if , for example , constraints include availability of personnel . Such logistical considerations are important: for instance , a slower six–month , but higher coverage ( 70% ) campaign takes the same time to eradicate rabies as a one–month synchronized lower coverage ( 60% ) campaign ( Fig . 5C ) . Taking longer to reach more dogs will have a greater impact than achieving low coverage quickly , offering further optimism that eradication is still feasible where resources are limited or hard to synchronize ( e . g . community–based ) . In terms of spatial roll out , there may be advantages to starting vaccinations where an outbreak began , because this is probably where there are most cases and is the most intuitive starting point for policy makers . However , this may only improve the chances of success if long distance human-mediated dog movement is restricted . The reactive strategy emphasizes this point: with no long distance transport , eradication times were fastest using this strategy because the most infected areas were vaccinated repeatedly . Yet with frequent long distance transport ( 5% , and as was estimated on Bali ) the reactive strategy performed worse than all others . Thus , while in some situations the reactive strategy could pay dividends , it is risky for at least two reasons: first , movement restrictions to slow rabies spread may be difficult to implement; and second , the potential to control an outbreak depends not only on the speed of transmission ( R0 and dog movement ) , but also the quality of surveillance [6] and responsiveness of control measures [29] . In Bali surveillance was not in place before the incursion , which led to delays in initiating a response , and the culling of dogs caused some people to move their dogs to safer areas . Establishing national surveillance and emergency response procedures should be prioritized given the continuing spread of rabies in the region . Further investigation into the potential of reactive strategies is warranted , including contact tracing in focal areas of transmission , and modeling to predict undetected infections [30] and to identify locations posing the greatest risks [26] . Future data collection on the human transport of dogs would be valuable for modeling realistic patterns of spread that may help direct targeted vaccination . Overall , our analyses strongly support the feasibility of rabies eradication from Bali and our modeling conclusions are borne out by the vaccinations campaigns carried out to date ( Fig . 6 ) . Whilst logistical difficulties of mobilization and implementation proved challenging , and heterogeneities in coverage compromised overall effectiveness , the extensive vaccination campaigns conducted have brought the epidemic under control . Further campaigns will be needed to eradicate rabies from Bali , and improving the comprehensiveness of these campaigns should be a high priority to achieve this goal . Once rabies does reach very low levels , then control measures may lapse and the risk of new incursions becomes an obvious danger , which we have not considered here . These risks are being evaluated in on-going field and modelling studies but , in the long term , genetic data could provide valuable information about the frequency and source of incursions . Eradication of rabies from Bali would not only save hundreds of lives , and millions of dollars by mitigating the indefinite need for expensive post-exposure prophylaxis , but would provide a valuable precedent for the feasibility of rabies eradication in very large and dense dog populations through effectively conducted mass vaccinations . More generally we make the following practical recommendations: 1 ) There is no evidence that rabies transmission in domestic dogs is density dependent over commonly encountered ranges of dog densities , so controlling rabies in higher density dog populations should not require higher vaccination coverage; 2 ) Vaccines that provide at least one year of protection should be effective , but the use of vaccines of shorter duration that require a booster could compromise the effectiveness of vaccination campaigns; 3 ) The advantages to spatially strategic roll-out or intensified synchronous effort for implementing vaccination campaigns are not justified if the increased logistical challenges compromise coverage; 4 ) Human–mediated transport of dogs expedites the spread of rabies and vaccination performance could be improved by restricting dog movement . However , there is currently no infrastructure to achieve this on Bali and indeed some dog owners in Bali reportedly moved animals to avoid culling or to replace dogs that had been culled , which could jeopardize spatial targeting of vaccination; 5 ) While achieving high coverage ensures the best possible chance of rabies eradication , repeat campaigns are vital to guarantee this . 6 ) The greatest concern for eradication programmes would be the lack of participation from any administrative areas , for example in Bali , omission of even the smallest of the nine Regencies that consists of 59 villages or 6% of the island could dramatically reduce the odds of achieving eradication to one third or less ( Fig . 5D ) . Our findings about the impact of omitting contiguous subpopulations may help explain why eradicating disease is so difficult without comprehensive coverage , particularly in landlocked areas with recurrent introductions from neighboring populations [3] , [4] , [26] . Determining the impact of neighbouring endemic areas on the effort required to eradicate rabies is an important question to address in future studies . Nonetheless , our results further emphasize the need for regional coordination in large–scale control programmes , as evidenced by successful control of rabies in the Americas [31] in contrast to Africa [16] .
|
Canine rabies continues to cause tens of thousands of horrific deaths worldwide , primarily in Asia and Africa . Momentum is building towards development of a global elimination strategy for canine rabies , but questions remain over how best to eliminate rabies epidemics . This paper uses data generated from the recent high-profile rabies outbreak in Bali , Indonesia to evaluate different control options . We find that , despite high dog densities , the spread of rabies on the island was remarkably similar to canine rabies spread elsewhere , suggesting that the practice of dog culling is an ineffective control strategy . We then simulate rabies transmission and spread across the island and compare the effectiveness of mass dog vaccination strategies in terms of how many lives are saved and how long it will take for elimination to be achieved . We find that the effectiveness of campaigns is not improved by being more reactive or synchronized but depends almost entirely upon reaching sufficient coverage ( 70% ) across the population in successive campaigns . Even small ‘gaps’ in vaccination coverage can significantly impede the prospects of elimination . The outputs of this study provide the kind of evidence needed by rabies program coordinators to help design effective national control programmes , and to build the evidence-base to drive forward the development and implementation of effective global rabies policy .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"disease",
"mapping",
"infectious",
"diseases",
"veterinary",
"diseases",
"zoonoses",
"rabies",
"veterinary",
"epidemiology",
"infectious",
"disease",
"epidemiology",
"epidemiology",
"zoonotic",
"diseases",
"spatial",
"epidemiology",
"population",
"biology",
"biology",
"animal",
"management",
"veterinary",
"science"
] |
2013
|
Designing Programs for Eliminating Canine Rabies from Islands: Bali, Indonesia as a Case Study
|
The bacteriophage φ29 generates large forces to compact its double-stranded DNA genome into a protein capsid by means of a portal motor complex . Several mechanical models for the generation of these high forces by the motor complex predict coupling of DNA translocation to rotation of the head-tail connector dodecamer . Putative connector rotation is investigated here by combining the methods of single-molecule force spectroscopy with polarization-sensitive single-molecule fluorescence . In our experiment , we observe motor function in several packaging complexes in parallel using video microscopy of bead position in a magnetic trap . At the same time , we follow the orientation of single fluorophores attached to the portal motor connector . From our data , we can exclude connector rotation with greater than 99% probability and therefore answer a long-standing mechanistic question .
As part of its viral infection cycle , the Bacillus subtilis bacteriophage φ29 packages its double-stranded DNA genome into a preformed capsid shell , or prohead , by means of a powerful molecular motor [1 , 2] . The DNA-packaging motor is situated at a unique 5-fold vertex of the prohead and is a complex assembly of multiple components . At the core of the motor is the dodecameric head-tail connector , gene product 10 ( gp10 ) . Associated with the connector is a ring of RNA molecules ( prohead RNA or pRNA ) , which is required for packaging . A ring of ATPases ( gp16 ) interacts with the pRNA to complete the packaging machinery . gp16 belongs to the Her A , FtsK superfamily of ATPases [3] . Hydrolysis of ATP powers the motor and drives viral DNA into the prohead . While numerous biochemical , structural , biophysical , and theoretical studies have elucidated many details of the packaging process [1 , 2 , 4–15] , a complete mechanistic understanding of how the components of the portal motor force the DNA into the capsid has not been presented . In particular , many theoretical models for the function of the connector have been proposed [5 , 7 , 16–18] . Most of these models include a rotation , either passive or active , of the connector with respect to the prohead shell—an idea first introduced by Hendrix in 1978 [16] . Recently , a study of DNA packaging in T4 , using cross-linking of bulky domains to the connector that could interfere with connector rotation , provided indirect evidence that the connector of T4 does not rotate during packaging [19] . However , no direct structural , biochemical , or biophysical experiments have been published that address the rotation hypothesis . Here , we directly test this hypothesis using single-molecule fluorescence polarization ( SMFP ) spectroscopy in combination with single-molecule force spectroscopy . Single-molecule force spectroscopy has proven to be a powerful method for studying the movement of motor proteins . In recent years , a wealth of different systems has been studied , such as actin- and microtubule-based molecular motors [20 , 21]; motors that move along DNA , like DNA polymerase [22] , RNA polymerase [23–26] , exonuclease [27 , 28] , and DNA pumps [2 , 29]; polymerization motors [30]; as well as motors that move pili [31] or whole bacteria [32] . Here , we use a magnetic tweezers set-up to observe the packaging of many DNA-packaging complexes simultaneously . Single-molecule fluorescence spectroscopy has been used recently to study conformational changes of single-motor complexes [33–39] . In particular , detection of changes in the polarization of emission by a single-dye molecule is well suited to investigate conformational changes involving rotation events [33 , 34 , 37] . We utilize this method to investigate the putative rotation of the connector protein with respect to the prohead shell . To test the hypothesis of connector rotation , we track orthogonal polarization components of fluorescent light emitted by single-dye molecules attached to the connector . Several issues must be overcome to ensure experimental fidelity: ( i ) DNA packaging by the labeled motor complex must be observed simultaneously with the fluorescence detection; ( ii ) dye labeling must be specific to the connector; ( iii ) the dye molecule must stay at a fixed angle relative to the connector; and ( iv ) the prohead itself must be immobilized without possibility of rotation . In order to fulfill all these requirements , we designed and implemented a combined SMFP and magnetic tweezers packaging assay as described below . The results of our experiment allow us to rule out connector rotation during packaging with more than 99% probability .
Figure 1 shows a schematic of the experimental geometry . Packaging is initiated in bulk as described previously [2] , and complexes stalled with non-hydrolyzable ATP analog ( ATP-γS ) and enriched for active particles are bound to streptavidin-coated superparamagnetic beads via a biotin modification to the distal end of the DNA ( see Materials and Methods for details ) . Antibodies against the capsid protein gp8 are used to anchor the stalled complexes via the prohead shell to a quartz slide . The experiment is observed through a glass cover slip , which together with the quartz slide forms a fluid chamber , using a 1 . 2-numerical aperture ( NA ) water-immersion objective . Magnets placed on either side of the objective pull the beads away from the quartz surface , stretching the unpackaged DNA under a force of about 0 . 1 pN . We monitor the beads via video microscopy and can calculate the length of the DNA tether through either Brownian motion [40] or change of focus of the bead image [41] . Upon exchange of buffers to remove ATP-γS and reintroduce ATP , the complexes resume packaging as confirmed by gradual reduction of tether length . The packaging is monotonic and ATP-dependent . In a typical experiment , about 80% of stalled , tethered complexes package to completion at rates consistent with previous bulk and single-molecule measurements [8] . Conformational changes that occur during the enzymatic cycle of a motor can be efficiently probed by attaching dye molecules as local reporters to specific proteins of a large multi-subunit complex . Single-dye molecules are sufficiently small so as not to interfere sterically with biological activity in most cases . The simplest way to attach a dye molecule as a reporter to a specific site of a protein is to make use of the high specificity of a cysteine-maleimide reaction . To this end , one needs to make point mutations in order to remove native cysteines , which could be inadvertently labeled , and introduce new cysteines ( one at a time ) to desired exposed locations . A dye molecule with a reactive maleimide group can then be covalently attached to that site . For our experiments , we replaced the two native cysteines of the connector with serines by site-directed mutagenesis and introduced new cysteines at various positions at the inside and outside of the connector . Figure 2 shows the X-ray crystallographic structure of the connector highlighting the position of the amino acids that were mutated to cysteines ( see Materials and Methods ) . Since current techniques are unable to assemble the free connector protein gp10 into a prohead in vitro , it was necessary to label the connector in the presence of the complete capsid , including capsid ( gp8 ) , head fiber ( gp8 . 5 ) , and scaffold proteins ( gp7 ) ( none of which contain cysteines ) . Figure 3A shows , as an example , the fluorescence image of a denaturing gel ( SDS-PAGE ) of the proheads with a cysteine at amino acid 170 ( 170C ) of the connector , labeled with a Cy3-maleimide . About 80% of the total fluorescence signal is in the single band of the connector , while capsid protein ( which makes up more than 95% of the total mass ) , head fibers , scaffolding protein , and residual contaminants of the Escherichia coli particle expression system show only very weak Cy3 fluorescence . Furthermore , the labeled proheads package DNA in vitro with the same efficiency as unlabeled particles , as can be seen by bulk packaging experiments ( see Figure 3B and Materials and Methods ) . Having established a specific labeling scheme in a complex that is competent for in vitro packaging , we also need to ensure that the dye molecule can act as a reporter for a rotational movement of the connector . Several criteria must be satisfied: the dye must be bright enough to allow integration times less than one rotational period; the time before photobleaching of the dye molecule must be long enough to observe several rotations; the dye must be attached at angles relative to the connector axis and relative to the objective optical axis such that changes in the dye dipole moment can be seen with our instrument; and the dye molecule must be attached rigidly . To address the latter concern , we measured the fluorescence anisotropy of an ensemble of labeled proheads . The resulting anisotropies of the labeled mutant particles were typically r ∼0 . 3 . The common interpretation is that the dye can freely rotate within a cone that is defined by steric limitations; in this case , the measured anisotropy corresponds to a cone half-angle of around 25° . However , the bulk anisotropy measurements only test the freedom of the dye molecules to reorient on the time scale of the fluorescence lifetime ( a few nanoseconds ) . In order to measure the time evolution of the orientation of a connector monomer ( a few seconds ) and to probe the time before photobleaching of the dye , as well as its brightness and orientation , we had to perform single-molecule fluorescence experiments . The long-term rotational rigidity of the dye , the rigidity of immobilization of the packaging prohead complexes , and the suitability of the dye orientation relative to the rotation axis can all be studied in single-molecule experiments [42] . To this end , we measured the polarization of the emitted fluorescence from proheads labeled with single-dye molecules . The labeled , stalled prohead complexes are attached to the surface of the micro-fluidic chamber ( see Materials and Methods ) and illuminated in prism-type total internal reflection geometry , with a green laser ( λ = 532 nm ) . A schematic of the experimental setup is shown in Figure 4 ( see also Materials and Methods ) . The fluorescence signal is collected by the objective and spatially separated into two perpendicular polarization components that are simultaneously imaged on a charge-coupled device camera . An overlay of the two polarization images yields a time-dependent fluorescence signal for each fluorophore in two orthogonal polarizations , thus allowing us to track the changes in relative orientation of a single dye in real time . φ29 packaging complexes stalled by incubation with ATP-γS and therefore incapable of enzymatically driven rotation were used to assess the dye suitability . Figure 5 shows the fluorescence signal of a single fluorophore attached to a stalled complex , with the excitation polarization being rotated between the two orthogonal directions , henceforth called horizontal and vertical , with a frequency of 0 . 7 Hz ( see Figure 5 for details ) . Fluorescence signals were integrated for 75 ms , and a typical time before photobleaching of the dye molecule was 10 s . The measured intensities in the two emission channels ( perpendicular polarizations ) are shown in black and red . The change in excitation polarization can be seen through an oscillation of the intensity in each channel . We observed a correlated signal ( which is a very clear indication for a stable orientation of the dye ) in more than 50% of the test experiments . Furthermore , it is important to note that the average signal intensity in the two channels remains almost constant over the lifetime of the dye , which shows that the dye keeps a stable orientation on timescales larger than the integration time . In contrast , a free diffusive rotation of the dye on the timescale of the integration time would lead to anti-correlation of the vertical and horizontal fluorescence signals . Anti-correlation ( and therefore free rotation of the dye molecule ) is likely caused by imperfections in the surface attachment of the stalled complexes . Finally , the fluorescence disappears in a single step at around time t = 22 s , indicating the presence of a single dye that was photobleached . These experiments ( and an additional control discussed in Text S1 ) , demonstrate that the polarization of the emitted fluorescence is an accurate reporter of the position of the connector protein , and that our instrument is indeed capable of detecting the changes in the fluorescence polarization , and hence the connector orientation , due to rotation on the single-fluorophore level . Constrained by the properties of single dyes and the camera , we can measure connector rotation rates from about 0 . 1–2 . 5 Hz , which corresponds to actual signal frequencies of 0 . 2–5 Hz due to the fact that the dipole emission has a 2-fold rotational symmetry . Current models for rotation predict a frequency within our detection bandwidth . Simpson et al . [5] have proposed a model for connector rotation based on the symmetry of the capsid and the connector . The connector is a homododecamer that sits at the 5-fold portal vertex of the icosahedral capsid . This 12/5 symmetry mismatch dictates that the relative alignment between connector and capsid is recapitulated with every 12°-rotation between the two structures . Furthermore , in vitro packaging experiments have measured the DNA-packaging rate at saturating ATP to be ∼100 bp/s [2] , and biochemical studies have shown that the step-size of the motor is 2 bp per ATP [8 , 43] . Given these figures , this rotation hypothesis predicts a rotation frequency of 1 . 65 Hz . Such rotation would result in a frequency of 3 . 3 Hz for the measured fluorescence signal in our experiment . Alternatively , if the connector were to track the helicity of the DNA [16] , the rotation would be 36° per basepair and the resulting signal frequency would be ∼20 Hz . In our experiments , the packaging velocity was reduced 2- to 4-fold by simply reducing the ATP concentration [8] in order to reach a signal frequency below 5 Hz , which is within our experimental time resolution . For actual packaging experiments , we illuminated the dye with homogeneously polarized light by time-sharing two different excitation polarizations . The emitted fluorescence was then detected in two orthogonal detection channels ( see Materials and Methods for details ) . If the dye were to rotate in the plane parallel to the chamber surface around the DNA being packaged , the intensity of horizontally polarized light would oscillate according to a sine-squared function , while the vertically polarized intensity would oscillate in the same manner with a phase shift of 90° . The two channels , therefore , would show an anti-correlated modulation of the fluorescence intensity . ( If the rotation axis is not perpendicular to the chamber surface , other phase shifts might be observed . We performed extensive simulations that suggested various limitations to our detection ability , and these are discussed below . ) Figure 6 shows typical examples of the single-molecule fluorescence signal during DNA packaging by the φ29 motor complex . Simultaneously , packaging activity is observed using a magnetic bead attached to the free end of the viral DNA . The red and black time traces show the fluorescence intensity detected in the two perpendicular polarization directions—horizontal and vertical , respectively . Figure 6A shows the signal for the mutant 170C . At t = 0 . 5 s , the signal of two single fluorophores can be seen . The first fluorophore bleaches after about 4 s , the second after about 19 s . This multi-step bleaching demonstrates our ability to quantify the number of dyes observed , and in rotation experiments only single-fluorophore traces were analyzed . The ratio of the intensities in the black ( vertical ) and red ( horizontal ) channel indicates that the first fluorophore is aligned almost horizontally , while the second fluorophore is at an angle of about 45° in between the two polarization directions . After about 100 s , scattered light from the magnetic bead , which is attached to the free end of the DNA , becomes visible in the trace . As the prohead continues packaging the DNA , the bead is pulled further into the evanescent field of the excitation by the green laser until it touches the surface at about t = 165 s . The packaging , assuming an approximately 10-kb tether , was therefore about 60 bp/s , consistent with bulk and optical tweezers measurements at this ATP concentration . It should be noted that we did not observe beads that were slowly pulled toward the packaging complexes in control experiments without ATP . Therefore , this behavior can be , without doubt , identified with the ATP-dependent DNA packaging of the motor complex . The center of the bead and the center of the initial single-dye fluorescence signal are within one pixel from each other , demonstrating the colocalization between the fluorophore and the packaging complex . By considering the density of fluorescent spots on the surface when using highly overlabeled proheads , we estimate that over 98% of such colocalized events are due to a tethered bead and dye molecule attached to the same packaging prohead ( see Text S1 for details ) . We have recorded more than 50 of these single dye/bead colocalized traces from six different mutants ( see Text S1 for a complete list ) . The data were analyzed in various ways . Details are given in Text S1; in short: First we used a normalized sliding correlation function that measures the correlation of the two perpendicular signals over a window of variable size . If the motor rotated during packaging , our simulations ( including noise ) suggest that one out of four traces should give an average correlation coefficient less than −0 . 3 for many seconds , assuming a random dye orientation . We never observed such correlation coefficients for an extended period of time ( several seconds ) in any of the data collected ( Text S1 ) . Second , we checked for periodicities in the channels by looking at Fourier transforms and cross-correlations . Our simulations predict that motor rotation generates a periodicity in more than 90% of the traces in the accessible frequency range , assuming random orientation of the dye molecule relative to the axis of rotation and random orientation of this rotation axis with respect to the optical axis . We did not observe this periodicity in a single trace .
We have studied a possible rotation of the bacteriophage φ29 portal motor protein with respect to the capsid on the single-molecule level during DNA packaging . With six different connector mutants , we did not observe a single trace resembling a signal expected for connector rotation . These results permit us to rule out with very high probability ( see below ) the compressive-ratchet model for connector rotation proposed by Simpson et al . [5] that involves a rotation by 6° per basepair . At the experimental ATP concentration of 25–50 μM , this rotation would lead to a rotation frequency in the fluorescence signal of about 1–2 Hz , which is easily detectable with our instrument . Similarly , a rotation model in which the DNA is wrapped around the external surface of the connector , with rotation providing an indirect translocating force [1 , 44] , would generate a rotation rate of 3° per basepair , also within the detection range of this experiment . A rotation rate below our detection limit is very unlikely , as a frequency of 0 . 1 Hz ( our lower detection limit ) would already correspond to as few as 0 . 6° per basepair—which would not fit any current model for rotation . On the other hand , we would be capable of observing a rotation of the connector if it were to follow the DNA double helix pitch in a nut and bolt fashion ( 10 . 5 bp per 360° ) , which would yield a rotation frequency in the fluorescence signal of 5 Hz at 25 μM ATP . While we can rule out complete rotation of the connector relative to the shell , we cannot rule out partial rotation over a small angle followed by return to the original position . In order to detect such transient rotation that would be consistent with models of connector flexure , polarization sensitive fluorescence correlation spectroscopy experiments would have to be performed . There remains a small uncertainty about packaging motor rotation due to the unknown orientation of the dye molecule with respect to the putative rotation axis of the connector and due to the lack of absolute labeling specificity . Simulations of dye emission ( unpublished data ) show that there are orientations of the dye where neither a correlated nor an anti-correlated signal can be observed , even if the connector is rotating . Given the random orientation of the rotation axis with respect to the substrate and of the dye axis with respect to the rotation axis , this situation should happen in about one out of ten traces for the signal levels shown . Here , we reduced the likelihood of an unfavorable orientation of the dye molecule with respect to the putative rotation axis by investigating six different mutants . We cannot rule out that all mutants result in unfavorable dye molecule orientations , although we consider this possibility highly unlikely . Considering the labeling specificity , there is currently no method to separate the connector protein ( gp10 ) from the capsid protein ( gp8 ) and re-assemble them again . Therefore , we have to label the connector in intact prohead particles , which might allow some dyes to attach to the nonrotating capsid . However , the fact that the capsid does not contain cysteines allowed us to achieve a specificity of more than 80% of dye on the connector of the 170C as observed on denaturing ( SDS-PAGE ) gels ( Figure 3 ) . We can also rule out that our selection of actively packaging complexes leads to a selection of nonlabeled complexes , since we have shown that labeling does not affect packaging efficiency or speed ( Figure 3B and unpublished data ) . As a result , about four out of five of the observed dyes can be expected to be attached to the connector . While no rotation of the connector was observed experimentally , does that mean that the connector does not rotate ? In order to answer this question , we have performed a mathematical analysis of the remaining uncertainties . This analysis is described in great detail in Text S1 . In brief , if one includes the uncertainty of unfavorable orientation , colocalization , labeling specificity , and rigidity of attachment , one can rule out connector rotation with more than 99% probability . In our experimental design , we were able to eliminate several limitations of previous efforts to combine single-molecule fluorescence and magnetic tweezers [45] . Magnetic beads scatter a great deal of light and are therefore not compatible with most single-molecule fluorescence experiments . However , as demonstrated by our results , if one uses highly localized excitation fields , like an exponentially decaying fluorescence excitation used here , they can be combined with single-molecule fluorescence . For a processive motor like the bacteriophage φ29 packaging complex , the disadvantage of using spatially separated trapping and fluorescence detection can be overcome by the colocalization of single-molecule fluorescence and bead fluorescence after the bead has been pulled close to the surface . Our experiments have shown that beads kept at a distance of larger than ∼1 μm do not introduce a significant scattered signal at the Cy3 fluorescence wavelength . They do block part of the emitted fluorescence light , but for tethers longer than 1 μm and bead sizes of 1 μm , the detected dye fluorescence is calculated to decrease by less than 10% . Another advantage of this setup is the possibility of parallel observation . In some preparations , we could observe the packaging of more than five complexes simultaneously . In previous single-molecule fluorescence studies on biological systems , the fluorescence signal itself was the only evidence of biologically relevant activity . Therefore , there had to be a characteristic and expected feature in the single-molecule fluorescence signal and sufficient statistics to confirm that the biological system is the cause for the observed signal . Obtaining these statistics can necessitate the observation of tens of thousands of fluorophores [33] . Here , we have presented a method that overcomes this problem . We can select for fluorophores that are attached to independently monitored active biological systems and observe their single-molecule fluorescence . In the present application , we can colocalize a fluorophore with a translocated bead . This leads to a 98% confidence that the observed fluorescence signal originates from a packaging prohead . The setup , therefore , opens exciting opportunities for the study of a number of different systems , such as RNA polymerase transcription initiation or elongation complexes , ribosomes , spliceosomes , or as shown here , nucleic acid translocation or packaging complexes in real time and with high resolution . In conclusion , we were able to test the connector rotation hypothesis , a long-standing prediction of several DNA packaging models . Our single-molecule experiments exclude with very high probability ( more than 99% ) the predominant model that the connector rotates with respect to the capsid . Having established that the connector does not rotate during packaging , it is important to ask how DNA is driven into the capsid . A model consistent with all experimental data was proposed recently by Chemla et al . [8] . In this model , ATP binding , hydrolysis , and release of products induce conformational changes in the ATPases that are directly involved in the translocation of the DNA [8] . Specifically , the translocation step of the DNA is triggered by , or performed by , the ring of ATPases via a conformational change that follows release of phosphate after ATP hydrolysis [8] . Here , we add to this model the idea that the connector could function as a valve to prevent DNA from leaking out . The spring-like shape of the connector suggests , indeed , that through compression and expansion , the connector may act as a “Chinese finger trap” ( George Oster , personal communication ) allowing the passage of the DNA in one direction during packaging but preventing its exit in the reverse . A complete understanding of the coupling that occurs between the ATPase , the pRNA , and the connector substructures is needed to refine our picture of the molecular mechanism of this powerful motor . Finally , the possibility that translocation could be driven by a ( nonrotating ) compression/extension ratchet mechanism is an intriguing idea , but one for which there is no direct experimental evidence to date and that is distinct from the mode of action established for certain AAA+ -related ATPases .
For all experiments , we used quartz slides ( Finkenbeiner Incorporated , http://www . finkenbeiner . com ) cleaned in piranha at 60 °C overnight , rinsed with purified water ( Barnstead , E-pure ) , sonicated in 2% Hellmanex ( Hellma , http://www . hellma . com ) , rinsed again , sonicated , and stored in pure water . Slides were blown dry with nitrogen immediately before being placed in 1 ml vectabond ( Vector Laboratories , http://www . vectorlabs . com ) and 100 ml acetone for 5–10 min . The slides were then washed in 100 ml water , slowly pulled out of the water bath such that no water remained on the hydrophobic surface , and placed in a wet box . 100 μl of a mixture of 3 mg Biotin-PEG-NHS ( Mw3400 , Nektar Therapeutics , http://nektar . com ) , 80 mg mPEG-SPA ( Mw5000 , Nektar Therapeutics ) , and 550 μl 0 . 1 M bicarbonate buffer was then placed on each slide and kept in the dark for 3 h . Afterward , the slides were rinsed thoroughly with water , dried with nitrogen , and assembled into a flow chamber by placing a Nescofilm gasket in between the quartz slide and a cover slip and heating for 3 min at 100 °C . The assembled chamber was then rinsed with 1 ml phosphate-buffered saline and incubated with 0 . 2 mg/ml streptavidin for about 20 min . After being rinsed again with phosphate-buffered saline and incubated with biotinylated antibodies against the capsid protein , gp8 , ( 0 . 1 mg/ml ) for 25 min , the chamber was rinsed with 1 ml 0 . 5× TMS ( 25 mM Tris-HCl , 50 mM NaCl , 5 mM MgCl2 ) and then with 250 μl of buffer XS ( 800 μl of 0 . 5× TMS , 10 μM ATP , 10 μM ATP-γS , 0 . 2 mg/ml BSA , 2 mg/ml glucose , 1% w/v beta-mercaptoethanol , 0 . 02 mg/ml catalase , 0 . 1 mg/ml glucose oxidase , and 0 . 8 μl of RNase inhibitor ( SuperaseIn , Ambion , http://www . ambion . com ) . With the exception of prohead particles , components for the φ29 in vitro packaging system ( DNA-gp3 , gp16 , 120-base pRNA ) were produced as previously described [46 , 47] . Prohead particles were produced in E . coli by overexpression of prohead structural proteins in HMS ( DE3 ) pAR 7-8-8 . 5-9-10 [48] . Two wild-type cysteines ( C76 and C265 ) in the φ29 connector , gp10 , were replaced by serines using standard site-directed mutagenesis to produce a cysteine-free clone , and individual amino acids in gp10 were replaced with cysteines to produce a library of single-cysteine mutants . Particles were produced by induction of mid-log cultures grown in LB media with 0 . 5 mM IPTG for 2 h . Cells were pelleted and re-suspended in a lysis buffer containing 50 mM Tris HCl ( pH 8 . 0 ) , 20 mM NaCl , 2 mM EDTA , 2 mM DTT , and 10 mg/ml lysozyme . After a 20-min incubation to produce sphaeroplasts , MgCl2 was added to 4 mM final concentration and DNase was added to a final concentration of 10 μg/ml to digest cellular DNA . Complete lysis was achieved with the addition of deoxycholate to 0 . 25% w/v . Extracts were clarified by centrifugation , and prohead particles were purified on 10%–40% w/v sucrose zonal gradients ( 45 kilo-rotations per min , 45 min , 20 °C ) in a SW55 rotor ( Beckman ) buffered with 1× TMS buffer ( 50 mM Tris-HCl , 100 mM NaCl , 10 mM MgCl2 [pH7 . 8] ) . Particles were collected , concentrated by pelleting , and re-suspended in 1× TMS . Labeling was conducted by adding an equal volume of Cy3-maleimide in H2O to prohead samples to provide the appropriate molar amount of dye with respect to available connector cysteine . After labeling for 1 h at room temperature , particles were purified away from free dye by sucrose gradient sedimentation ( 5%–20% w/v sucrose in 1× TMS , 45 K , 30 min , 20 °C ) . Particles were pelleted and re-suspended as above . Particles were quantified and checked for purity and labeling efficiency by SDS-PAGE and for DNA-packaging activity ( Figure 3 ) . Labeled proheads were reconstituted with 120-base pRNA at a ratio of 10 pRNAs/prohead by mixing 2 μl of pRNA ( 0 . 07 mg/ml ) with 2 μl of labeled proheads ( 1 mg/ml ) for 15 min in 0 . 5× TMS . Reconstituted proheads were then added to a packaging reaction containing 2 μl of biotinylated DNA ( 0 . 44 mg/ml ) and 2 μl of ATPase gp16 ( 0 . 025 mg/ml ) in a final volume of 18 μl of buffered in 0 . 5× TMS ( for a final ratio of two proheads:one DNA-gp3:15 gp16 ) . After 5 min , packaging was initiated with 4 μl of ATP ( 250 μM ) . After 60 s , 2 μl of ATP-γS ( 1 mM ) was added to stall the reaction . Magnetic beads ( MyOne Dynabeads , Invitrogen , http://www . invitrogen . com ) were prepared by washing three times in 0 . 5× TMS , then blocked by adding 2 μl of beads to 25 μl of 2 mg/ml BSA in 0 . 5× TMS . Blocked beads were then treated with 0 . 5 μl of RNAse inhibitor . Freshly prepared stalled complexes from above ( 4 μl ) were bound to beads by mixing 0 . 1 μl of Superase Inhibitor , 3 μl of BSA ( 10 mg/ml ) , 0 . 5 μl of ApaLI ( 10 U/μl ) in 80 μl of buffer X ( 0 . 5× TMS , 10 μM ATP , 10 μM γ-S ATP ) . After gentle mixing , the sample was incubated for 1 h for the ApaLI restriction digestion which cleaves at both the extreme right and left ends of the DNA , to reduce the presence of biotinylated DNA-gp3 on the bead surface not associated with stalled-packaging complexes , which can form nonspecific tethers due to the stickiness of gp3 . ( Left ends of packaged DNA-gp3 are in the prohead and protected from digestion . ) Using a magnet , we washed the magnetic beads three times with buffer X . These washes removed free and cut DNA-gp3 ends , all free dye , and proheads that did not initiate packaging . Finally , the beads were flowed into the chamber and incubated for 10 min . To restart packaging , 0 . 5× TMS buffer containing 0 . 2 mg/ml BSA , 2 mg/ml glucose , 1% w/v beta-mercaptoethanol , 0 . 02 mg/ml catalase , 0 . 1 mg/ml glucose oxidase , and 50 μM ATP was flushed into the chamber . During SMFP , the sample is illuminated in prism type total internal reflection geometry with a green laser ( 532 nm , CrystaLaser , http://www . crystalaser . com ) . There are two ways for a dye molecule to report on the rotation of a macromolecule: First , one could excite the dye molecule with linearly polarized light . The emitted fluorescence intensity would then be proportional to the square of the scalar product of dipole orientation and polarization direction . For a rotating dye molecule in the plane of the evanescent field , the emitted intensity would oscillate between a maximum and minimum within a given polarization as the molecule changes its orientation . A second option for using a dye molecule as a direction sensor is to illuminate with both horizontal and vertical polarization ( with equal intensities , i . e . , homogeneous polarization ) but detect the polarization state of the emitted fluorescence . Here , the angle between emission dipole and polarizer in the detection path becomes important . For a molecule rotating in the polarization plane , the result would be an anti-correlated signal between the vertical and horizontal polarization . We chose to use the second option , since an anti-correlated signal cannot be confused with other events such as blinking or changes in molecular brightness . In order to achieve an illumination with homogeneous polarization , the light of the laser is coupled into an electro optical modulator ( Linos Photonics , Incorporated , http://www . linos-photonics . com ) that switches between two perpendicular polarization directions with a frequency of 10 kHz . This switch is orders of magnitude faster than the integration time during the experiment , and therefore , only homogenous polarization is observed . The two perpendicular polarizations are then split by a polarizing beam splitter ( PBS ) . The resulting beams are both brought to the chamber in s-polarization , but from perpendicular directions—one from the side and one from the top . For this reason , we used a custom-made prism with two input ports . We checked the light intensity from both directions by comparing the signal scattered by beads attached to the flow chamber surface and adjusted the intensities to be equal in the center of the field of view; the intensities varied by less than 50% across the field of view . The fluorescence light is collected by a high NA objective ( Nikon , 1 . 2 NA , http://www . nikon . com ) and separated in two perpendicular polarization components by a PBS . The two beams are spatially offset and recombined by another PBS . The two beams are then focused on a CCD camera ( Cascade 512B , Photometrics , Roper Scientific Incorporated , http://www . roperscientific . com ) , such that two images , one for each polarization , can be read out simultaneously . One pixel on the camera ( physical size 14 μm ) corresponds to about 400 nm . Because of the huge dilution of phages , and therefore dyes on the substrate ( less than one dye per 10 μm2 ) , we used hardware binning by 3 × 3 pixels . Furthermore , since the fluorescent spot was not always centered on one point , we added up to four adjacent points for signal optimization . Two band-pass filters ( 580BP50 , Omega Optical , http://www . omegafilters . com ) are used to separate excitation and LED illumination from the single-molecule fluorescence signal . At the same time , the sample is illuminated with a red LED to observe the magnetic beads that are pulled away from the surface by two magnets . The red light scattered from the magnetic beads is separated from the fluorescence light with a dichroic mirror ( 630 DCLP , Omega Optical ) and detected on a separate CCD camera ( Watec , 902C; Watec Company , http://www . watec . com ) . In addition , an epi illumination can be used to focus onto the chamber surface without illuminating ( and therefore bleaching ) the surface .
|
The life cycles of many viruses include a self-assembly stage in which a powerful molecular motor packs the DNA genome into the virus's preformed shell ( the capsid ) . Biochemical and biophysical studies have identified essential components of the packaging machinery and measured various characteristics of the packaging process , while crystallography and electron microscopy have provided snapshots of viral structure before and after packaging . In bacteriophage φ29 assembly , the DNA passes into the shell through a channel formed by a structure called the connector . Structurally motivated models over the past 30 years have coupled DNA movement to rotation of the connector relative to the capsid . We describe a direct test of the connector rotation hypothesis , combining magnetic single-molecule manipulation techniques and single-molecule fluorescence spectroscopy . In our experiments , we use a single-dye molecule attached specifically to the connector as a reporter for its orientation and simultaneously observe the translocation of a magnetic bead attached to the DNA that is being packaged . From our data , we can exclude connector rotation with greater than 99% probability and therefore answer a long-standing mechanistic question .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"viruses",
"biochemistry",
"virology",
"biophysics",
"molecular",
"biology"
] |
2007
|
Experimental Test of Connector Rotation during DNA Packaging into Bacteriophage φ29 Capsids
|
Expression quantitative trait loci ( eQTL ) studies have generated large amounts of data in different organisms . The analyses of these data have led to many novel findings and biological insights on expression regulations . However , the role of epistasis in the joint regulation of multiple genes has not been explored . This is largely due to the computational complexity involved when multiple traits are simultaneously considered against multiple markers if an exhaustive search strategy is adopted . In this article , we propose a computationally feasible approach to identify pairs of chromosomal regions that interact to regulate co-expression patterns of pairs of genes . Our approach is built on a bivariate model whose covariance matrix depends on the joint genotypes at the candidate loci . We also propose a filtering process to reduce the computational burden . When we applied our method to a yeast eQTL dataset profiled under both the glucose and ethanol conditions , we identified a total of 225 and 224 modules , with each module consisting of two genes and two eQTLs where the two eQTLs epistatically regulate the co-expression patterns of the two genes . We found that many of these modules have biological interpretations . Under the glucose condition , ribosome biogenesis was co-regulated with the signaling and carbohydrate catabolic processes , whereas silencing and aging related genes were co-regulated under the ethanol condition with the eQTLs containing genes involved in oxidative stress response process .
eQTL studies aim to uncover the genetic architecture underlying expression regulation . In the past decade , they have been conducted in many organisms , including yeast , drosophila , mouse , human and many others [1]–[5] . A common approach in eQTL data analysis is to consider association between each expression trait and each genetic marker through regression analysis , and attention is usually focused on those trait-marker pairs whose associations are significant after multiple comparison adjustments . Despite great success with this approach , some regulatory signals may not be detected due to the complex nature of regulatory networks . For example , genetic buffering relationships often exist in the phosphorylation regulatory network in yeast [6] , where pairs of regulators have overlap in function . Similar phenomenon has also been observed in the transcriptional regulatory network in yeast [7] . Single marker analysis may not capture such regulatory patterns , where the genetic effects act through interactions between markers , necessitating the need to incorporate interactions in the analysis . However , extending beyond single marker analysis presents many challenges including the computational demand and the lack of statistical power , because a much larger number of models need to be considered and the need to control the overall false positive results . Storey et al . [8] developed a step-wise regression method to detect epistasis on the genome-wide scale . This method is computationally feasible but may miss epistatic effects involving markers having weak marginal effects . Wei et al . proposed a Bayesian partition model which may detect more loci having epistatic effects but weak marginal signals [9] . However , this Bayesian approach did not compare favorably with an exhaustive search scheme to detect features with weak marginal signals but strong epistatic effects in practice [10] . To reduce the model search space and increase statistical power , Lee et al . adopted genetic interaction networks identified by large-scale synthetic genetic array ( SGA ) analysis as prior for detecting epistasis in yeast [11] . Since they only consider interacting SNPs that have already been identified , its application is limited to those organisms where comprehensive prior knowledge is available , which is rare in practice . Although most eQTL studies considered the expression levels of individual genes as response , a conceptually different approach was proposed by Li et al . [12] to consider “liquid association” ( LA ) between a pair of genes . LA aims to identify differential co-expressions , versus differential expressions , across different samples/conditions and the identified LA may offer insights that may not be captured by analysis based on single genes . Li and colleagues later introduced this 2D-trait concept into eQTL study [13] . The goal of such 2D-trait based eQTL analysis is to identify genetic markers that can affect the co-expression patterns between two genes . Since co-expression patterns reflect co-regulation status , such 2D-trait analysis can assess whether the co-regulatory relationship between two genes is associated with certain genetic markers , which is complementary to analyzing the expression patterns of individual genes . For example , in signal transduction pathways , transcriptional factors ( TFs ) are often regulated by post-transcriptional regulation such as phosphorylation and dephosphorylation . Such regulations are difficult to detect because there may be little change at the expression levels for these genes . However , post-transcriptional regulation does affect TFs' activities , which further affect the expression levels of their target genes . In this case , if a genetic marker affects post-transcriptional regulation , its effect may be captured by the change of co-expression patterns of the targets of TFs , so a LA analysis may lead to the identification of such markers , where it may be difficult to detect these signals using single gene expressions as the response . Recently , Ho et al . [14] proposed a conditional bi-variate normal model to analyze LA that simultaneously captures means , variances , and correlation between a pair of genes . Under a similar framework , Chen et al . [15] proposed a penalized likelihood approach to effectively detecting causal genetic loci using iterative reweighted least squares , and Daye et al . [16] further considered the heteroscedastic problem . Although these methods have broadened the scope of eQTL analysis , none have considered the possibility that markers may have no or weak marginal effects but strongly interact to affect the correlations patterns among gene expressions , which may happen if there is genetic buffering between the markers and this is the focus of our current manuscript . One major challenge to consider interactions effects on 2D-traits is the large number of models to be examined . For example , with 6000 genes in yeast , a total of 18 million 2D-traits can be formed . If we collect 4000 markers from each yeast strain , considering each pair of markers for their interaction effects will involve 8 million pairs of markers . Therefore , an exhaustive search of all 2D traits versus all marker pairs will evaluate models , a prohibitive number with the current computing power based on the existing methods mentioned above . In this manuscript , we propose a computationally efficient algorithm to identify these Epistasis-2D associations based on conditional bivariate models and likelihood ratio test . In our procedure , we proposed to use a statistic called PA ( Potential of Association ) to filter out trait and marker sets that are unlikely to be significant before performing the more rigorous likelihood ratio tests . When we applied our method to a yeast eQTL dataset , we were able to identify many “Epistasis-2D” associations that could not inferred from single marker based analysis , where 2D refers to our focus on gene co-expression patterns and epistasis refers to our focus on detecting how loci interact to affect 2D-traits .
We applied our method to a dataset containing gene expressions and genotypes for 109 segregants from a cross between laboratory ( BY , noted as 1 ) and wild ( RM , noted as 0 ) strains of Saccharomyces cerevisiae [3] . The expression levels were measured under two different conditions: glucose and ethanol . We applied our method to the expression data collected under these conditions , and identified 225 and 224 pairs of genes ( 2D-traits ) , respectively , whose correlation patterns were under the epistatic control of pairs of markers at an estimated false discovery rate ( FDR ) ( Materials and Methods , Text S1 , Table S3 ) . As far as we are aware , none of these detected marker interactions have been reported to affect expression traits , and our results revealed a new group of regulation patterns that have been overlooked in the literature . Among the 225 and 224 gene pairs , there is an enrichment of pairs having the same functional annotations ( 31 out of 225 with a p-value of 0 . 05 and 58 out of 224 with a p-value of ) according to GO slim . Despite this statistically significant enrichment , most pairs have different functional annotations suggesting either unknown functions for these genes or interactions between different biology processes . We observed that the functional distributions of the Epistasis-2D associations are dependent on the environment condition under which the eQTL data were collected . This is consistent with the literature on the importance of the environment on gene expression regulations [17] . Also consistent with previous finding that the trans-acting linkages differ under different environmental conditions [18] , our results suggest that trans-acting loci are related to the environment related stress response pathways . The modules identified by our method may be followed up with experimental studies for validation and learning to gain further insights on their biological relevance .
We have developed a novel statistical approach to identifying gene pairs whose co-expression patterns are jointly regulated by interacting loci through the analysis of eQTL data . Our approach is based on modeling the joint expression levels with a bivariate normal distribution whose covariance matrix is dependent on the joint genotypes at two candidate loci . Although different model search strategies have been proposed to jointly analyze multiple markers and their interactions based on genome wide data , e . g . marginal search , forward search and exhaustive search , the ability to conduct an exhaustive search allows us to identify interacting loci with weak marginal effects [63] . To facilitate an exhaustive search of all gene pairs versus locus pairs , we also proposed a filtering process to only focus on those modules that are likely to be statistically significant . This filtering process is one important component of our strategy to reduce the computational burden without reducing statistical power for discoveries . The application of our method to a yeast data set has identified many interacting loci with weak marginal signals which would not have been found without the exhaustive search strategy . Compared to the existing methods to detect epistasis , we considered the 2D-trait , especially their co-expression patterns , as the phenotype . As discussed in the introduction section , using such 2D-traits may help to detect post-transcriptional regulation from the change of expression correlations between downstream genes . As shown with the examples in the results section , we detected many regulatory loci containing candidate genes encoding kinases or phosphatases that regulate the co-expression correlation between the targets of their TF substrates . None of these modules could have been detected using 1D-traits . Since we only focus on modules which can not be detect using 1D-mapping in this paper , we may miss potential Epistasis-2D modules with genotype-dependent mean values ( ) through 1D-Map filtering and assuming in our model . About 13% modules were filtered using 1D-Map filtering , therefore around 13% Epistasis-2D modules could be missed by our method . Although we could introduce more parameters in our model to allow for genotype-dependent mean values , this may introduce noises that lead to reduced statistical power with limited sample size . More detailed discussion on the trade-off between statistical power and model adequacy is provided in the supplementary materials ( Text S1 , Figures S6 and S7 ) . We applied our strategy to a well studied yeast eQTL dataset and detected many epistasis modules , most of which have not been discovered to date and many may be interpreted with existing biological literature . We found that the co-regulated genes in the modules inferred under different environments were enriched for different biological processes . For example , under the glucose condition , ribosome biogenesis tends to be co-regulated with glucose response and glucose metabolic processes . The loci jointly regulating their expression patterns are enriched with genes in the glucose response pathway . Under the ethanol condition , silencing and aging related genes were found to be co-regulated . The loci jointly regulating these genes are enriched with genes in the oxidative response pathway , consistent with the hypothesis that the metabolism of ethanol would induce aging through increased damage from ROS produced in oxidative stress response . Through detailed discussion of several identified modules , we proposed potential regulatory mechanisms between oxidative stress signal and aging process . Interpretation is difficult in eQTL linkage studies because the detected eQTLs often have low resolution , e . g . large intervals , with many candidate genes . Traditional linkage analysis with single genes and one locus often offers limited information to identify a candidate gene around the locus to understand the linkage signal . Since Epistasis-2D modules detected in our study involve two genes and two loci , the biological association of the two genes offers additional information to prioritize candidate genes in the inferred loci as shown in the examples in the results section . However , significant challenges remain to identify candidate genes in the inferred loci and interpret the results . First , there is genetic buffering in a robust regulatory network , and we may not be able to infer all the direct linkages from eQTL studies . The mediators not observed between indirectly linked loci and genes make it more difficult to interpret the regulatory linkage . Second , in the Epistasis-2D modules , the genetic loci may affect one of the two target genes or both of them , and either situation will cause the variation of the co-expression pattern . This also increases difficulty for interpreting the linkage results . Therefore it is often necessary to incorporate information from other resources to interpret the detected modules . For example , in the oxidative phosphorylation pathway modules we illustrated in the results section , the co-expression patterns between Dbp8 and other oxidative phosphorylation genes are co-regulated . Since the candidates we predicted are all involved in the oxidative phosphorylation pathway , it is quite possible that only the expression of oxidative phosphorylation genes in Dbp8-related modules , but not the expression of Dbp8 , is actually affected . Similarly , we also investigated different types of databases to collect evidences and interactions for interpreting other discussed modules . It is important to integrate other data sources including protein interactions , transcription and proteomics data under a consistent framework to better interpret the results . This generic idea has been formalized in different ways for interpreting one-to-one linkages [64]–[67] , and more work is needed to adapt these methods to interpret the modules identified by our method . Utilizing our results through integration of multiple data sources is an interesting future direction . Our strategy could also be applied to other eQTL data in mouse or human . For example , in the mouse eQTL data , there are around 2000 markers which is comparable to the yeast data and the number of differentially expressed transcripts was around 8000 [4] . In this case , the search space is on the order of , which can be readily handle by paralleling our algorithm . In the human eQTL data , up to over 5 , 000 , 000 SNPs may be genotyped and up to 50 , 000 transcripts may be profiled . This will dramatically increase the computation time . We may reduce the computational burden by focusing only on those transcripts of interest ( e . g . those known to be relate to diseases ) or setting more stringent cutoffs in the filtering process to accelerate the processing . However , more computationally efficient methods need to be developed to identify Epistasis-2D modules for these data if we want to consider all the traits and markers . One possible direction is to jointly consider multiple markers within a region as those done for GWAS data [68] , [69] .
We define a module in this manuscript as the collection of a pair of loci and a pair of genes , denoted as , where and represent two loci and and represent two genes . Our objective is to identify Epistasis-2D modules where and interact to affect the co-expression patterns of and . To formally describe our model , we use to denote the genotypes of and and the expressions of and . We assume that , ( 5 ) whereis the covariance matrix , and ( 6 ) where I is the indicator function , i . e . I ( A = i ) = 1 if A = i and 0 otherwise , and T is the set of genotypes . For example , in the yeast dataset . In this study , we focus on associations that can not be detected using 1D-trait ( expression level ) , i . e . we assume that and are independent of and . This simple model may have overall good statistical power to detect Epistasis-2D modules as discussed in detail in the supplementary materials ( Text S1 , Figures S6 and S7 ) . Without loss of generality , we let which approximately hold after applying the following transformation to the expression data: The normal quantile transformation based on individual genes is a means to “normalize” the sample observations so that our procedure is robust to the effects of extreme observations and/or highly skewed distributions [12] , [15] ( Text S1 , Figures S1 and S2 ) . In our model , we assume that and are independent of the genotype because we found this specification achieved a good balance between model adequacy and simplicity . We illustrate this through the analysis of simulated data and a subset of the real data in the supplementary materials ( Text S1; Figures S3 , S4 , S5 ) . We found that although it was feasible to fully consider genotype dependent variances , there may be overall power loss due to additional model parameters , especially when the sample size is limited . Considering a sample with n individuals , let represent the genotypes and expressions in the kth sample . The model parameters in ( 8 ) can be estimated using the maximum likelihood estimates ( MLE ) , where the log-likelihood function is , ( 7 ) Our goal is to identify gene pairs whose correlations depends on the joint genotypes of the two loci . We consider 12 epistatic models ( Table S1 ) versus the null hypothesis that the correlation is the same for different joint genotypes . To focus on epistatic interactions , we then compare the most significant model with two single association models . The comparisons are based on the likelihood ratio ( LR ) test . Since MLE needs to solve a numerical optimization problem , applying the tests above to all possible modules is computationally expensive . Therefore , we introduce a statistic “PA-score” ( Potential of Association ) to estimate the lower bound of the p-value for each module . The PA-score is defined as , ( 8 ) where is the number of individuals with genotypes and , is the Pearson correlation coefficient of the expression levels among the individuals and is the correlation coefficient among all the individuals . We prove in the Text S1 that the expectation of PA corresponds to the lower bound of p-value for each module . In this case , we could control the sensitivity by choosing the cutoff for PA to filter out modules before performing the LR tests . In this paper , we chose a cutoff value of 45 for PA which has an estimated sensitivity 0 . 995 . The sensitivity here is one minus the percentage of the significant LR test modules which will be filtered out by PA score filtering . The details of the sensitivity estimation are provided in the supplementary materials ( Text S1 , Figure S8 ) . After the PA-score filtering , more than modules remained for each condition . Using the LR tests , we identified 225 and 224 2D-traits whose correlation patterns were under the epistatic control of pairs of markers . Therefore we estimated that epistatic controlled 2D-traits was filtered out by PA score in each condition . Since PA can be directly calculated from the data , the filtering process can reduce the total computing time by a factor of 16 from our experiments with the data ( Text S1 ) . We analyzed the yeast dataset collected by Kruglyak and colleagues [3] . The expression data were downloaded from http://www . plosbiology . org/article/info%3Adoi%2F10 . 1371%2Fjournal . pbio . 0060083 , with 4 , 482 genes measured in 109 segregants derived from a cross between BY and RM . The experiments were performed under two conditions , glucose and ethanol . We removed 63 genes with more than 10 missing values in either condition for a total of 4 , 419 genes analyzed . The authors provided genotypes at 2 , 956 loci . We combined neighboring loci having fewer than 5 discordant calls among the 109 samples , leading to 820 merged markers . In this manuscript , we still call these merged markers as markers to simplify the discussion . For each marker pair , an individual can have four joint genotypes , ( 0 , 0 ) , ( 0 , 1 ) , ( 1 , 0 ) , and ( 1 , 1 ) . We only considered marker pairs where there were at least 15 individuals for each joint genotype . There were a total of 305 , 301 such pairs . Therefore , we tested 305301 modules . The algorithm was implemented in R . Applying our procedure to one condition took one week of one CPU on a Linux cluster with 2 . 40 GHz CPU . We estimate the false discovery rate ( FDR ) through a permutation technique similar to previous study [8] . Specifically , we randomly permutated the expression data across all the genes and applied our procedure to the permuted data set using exactly the same setting as the real dataset . That is we used the same cutoff 45 for PA , and the same cutoff for p-values of the LR tests ( also select the best model ) . For a given threshold for LR tests , we counted the number of unique 2D-traits involved in modules with their p-values lower than the threshold . Note that we did not use the number of modules to calculate FDR because a 2D-trait may be mapped to multiple neighboring marker pairs since neighboring markers tended to have similar genotypes . Hence , we use 2D-trait to label the modules for FDR estimation . We performed ten permutations for each condition to yield ten sets of simulated null modules , and the results were consistent across these ten permuted data sets ( Table S2 ) . For example , at the threshold value of , the average number of unique 2D-traits involved in modules with a statistical significance level less than in the permuted dataset was 36 . 4 ( SD = 4 . 9 ) and 38 . 5 ( SD = 5 . 1 ) , respectively . Therefore , with a total of 225 and 224 significant 2D-traits identified for the observed data under the two conditions , the estimated FDR was for both conditions ( Text S1 , Table S3 ) . Among the inferred Epistasis-2D modules , neighboring markers tended to show similar patterns of interactions as discussed previously [70] . We merged neighboring markers with fewer than 15 individuals showing discordant genotypes among all samples , leading to 266 merged markers for clustering analysis . Table S4 listed all detected Epistasis-2D modules after the merging . We define an epistasis map E under a specific condition as We performed hierarchical clustering on this map using Cluster 3 . 0 . For each gene , we used GO slim to annotate its function . The gene pair in each Epistasis-2D module were annotated with a pair of functions . To investigate whether a particular pair of functions were over-represented among the Epistasis-2D modules , we performed the following hypergeometric test , ( 9 ) where N is the total number of gene pairs , M is the number of gene pairs with two specific functions , n is the number of gene pairs from Epistasis-2D modules , and k is the number of Epistasis-2D gene pairs having the specific two functions . The p-values were Bonferroni corrected for multiple testing . Before analyzing the results , we note that many genes involved in these function categories are overlapped . Under the glucose condition , 18 genes annotated as “precursor metabolites/energy” actually consist of carbohydrate metabolic genes ( 7/18 ) and cellular respiration genes ( 11/18 ) . Genes annotated as “cofactor metabolic” are also highly overlapped with these two processes ( 7/13 ) . In addition , genes annotated as “RNA metabolic process” are mainly involved in ribosome biogenesis ( 33/80 ) . Under the ethanol condition , most genes annotated as “transcription” and “chromosome organization” are involved in the RNA metabolic process ( 32/34 , 9/19 ) . According to these overlaps , the main differences between the two conditions can be summarized as shown in Figure 4 . Since a chromosomal interval encompassing the markers may contain multiple candidate genes , we need to perform enrichment analysis to investigate whether there is statistically significant evidence for the enrichment of certain function . We performed hypergeometric test to investigate whether a particular function was over-represented among the genes located at the chromosomal intervals within one or several modules . The p-value was calculated as , ( 10 ) where n is the total number of considered chromosome intervals , N is the total number of annotated genes , M is the number of genes with specific function , is the number of genes in ith chromosome interval , and is the number of genes having the specific function in ith chromosome interval . The gene function is defined by GO annotation at level 5 ( downloaded from DAVID Knowledgebase [71] , [72] ) . The calculation detail of the examples that discussed in the results section was listed in Table S6 .
|
eQTL studies collect both gene expression and genotype data , and they are highly informative as to how genes regulate expressions . Although much progress has been made in the analysis of such data , most studies have considered one marker at a time . As a result , those markers with weak marginal yet strong interactive effects may not be inferred from these single-marker-based analyses . In this article , using joint expression patterns between two genes ( versus one gene ) as the primary phenotype , we propose a novel statistical method to conduct an exhaustive search for joint marker analysis . When our method is applied to a well-studied dataset , we were able to identify many novel features that were overlooked by existing methods . Our general strategy has general applicability to other scientific problems .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genome-wide",
"association",
"studies",
"genetics",
"biology",
"computational",
"biology",
"genetics",
"and",
"genomics"
] |
2013
|
Statistical Analysis Reveals Co-Expression Patterns of Many Pairs of Genes in Yeast Are Jointly Regulated by Interacting Loci
|
Lichtheimia species are the second most important cause of mucormycosis in Europe . To provide broader insights into the molecular basis of the pathogenicity-associated traits of the basal Mucorales , we report the full genome sequence of L . corymbifera and compared it to the genome of Rhizopus oryzae , the most common cause of mucormycosis worldwide . The genome assembly encompasses 33 . 6 MB and 12 , 379 protein-coding genes . This study reveals four major differences of the L . corymbifera genome to R . oryzae: ( i ) the presence of an highly elevated number of gene duplications which are unlike R . oryzae not due to whole genome duplication ( WGD ) , ( ii ) despite the relatively high incidence of introns , alternative splicing ( AS ) is not frequently observed for the generation of paralogs and in response to stress , ( iii ) the content of repetitive elements is strikingly low ( <5% ) , ( iv ) L . corymbifera is typically haploid . Novel virulence factors were identified which may be involved in the regulation of the adaptation to iron-limitation , e . g . LCor01340 . 1 encoding a putative siderophore transporter and LCor00410 . 1 involved in the siderophore metabolism . Genes encoding the transcription factors LCor08192 . 1 and LCor01236 . 1 , which are similar to GATA type regulators and to calcineurin regulated CRZ1 , respectively , indicating an involvement of the calcineurin pathway in the adaption to iron limitation . Genes encoding MADS-box transcription factors are elevated up to 11 copies compared to the 1–4 copies usually found in other fungi . More findings are: ( i ) lower content of tRNAs , but unique codons in L . corymbifera , ( ii ) Over 25% of the proteins are apparently specific for L . corymbifera . ( iii ) L . corymbifera contains only 2/3 of the proteases ( known to be essential virulence factors ) in comparision to R . oryzae . On the other hand , the number of secreted proteases , however , is roughly twice as high as in R . oryzae .
The basal lineages of terrestrial fungi , formerly Zygomycota , were recently shown to be polyphyletic and were therefore separated into four separate subphyla [1] . Especially the order Mucorales of the Mucoromycotina encompasses several human pathogenic species . Although infections with mucoralean fungi ( mucormycosis ) are less common as compared to aspergilloses or candidioses , these fungi are increasingly recognized as the source of infection in immunocompromised patients [2] . Mucormycoses are associated with rapid blood vessel invasion and massive destruction of tissue ( necrosis ) [3] , [4] . Mortality rates are high ( ∼50% ) and treatment mainly includes a combination of antifungals and extensive surgery [2] , [5]–[7] . In addition , mucoralean pathogens are resistant to a variety of antifungals including voriconazole which makes treatment even more complicated [8] . The order Mucorales comprises 240 described species , of which at least 20 have been found to be involved in mucormycosis . Genome sequences have been published for only two important pathogenic species within the Mucorales , namely Rhizopus oryzae ( = R . arrhizus ) and Mucor circinelloides . These species are closely related and represent derived lineages within the group . However , a large proportion of pathogenic Mucorales ( 10 species ) belong to more basal groups including the genera Lichtheimia , Rhizomucor , Apophysomyces , Saksenaea and Syncephalastrum . Recently , the first report of the involvement of Thamnostylum lucknowense , an ancient mucoralean fungus , in human infections has been published [9] . To date , almost nothing is known about the genomic structure and pathogenicity mechanisms of these basal groups . Lichtheimia species are ubiquitous saprophytic molds and represent the second and third most common cause of mucormycosis in Europe and worldwide , respectively [2] , [7] , [10] , [11] . The genus Lichtheimia was formerly included in the genus Absidia based on morphological similarities [12] . However , based on the higher growth optimum as well as morphological and molecular data Lichtheimia species were separated from the mesophilic Absidia species [13] . Today the genus encompasses five thermotolerant species , of which three are known to be clinically relevant , namely L . corymbifera , L . ramosa and L . ornata [14] . In addition to the distinct phylogenetic position at the base of mucoralean fungi , Lichtheimia species exhibit differences in physiology compared to the sequenced pathogens M . circinelloides and R . oryzae , including a higher maximum growth temperature ( 48–52°C vs 37°C and <45°C ) and differences in susceptibility to certain antifungals [8] , [15] . Moreover , filamentously growing Mucor and Rhizopus species have been shown to be able to form yeast cells which were also found in patient material and thus might be of relevance during infection [16]–[18] . In contrast , no yeast-like growth forms of Lichtheimia species have been observed to date . In addition , pulmonary Lichtheimia infections following solid organ transplantation seem to be associated with a higher risk to develop disseminated disease [19] . Besides its role in human infections , L . corymbifera is also believed to be associated with Farmer's lung disease ( FLD ) , a hypersensitivity disorder resulting from frequent contact of mouldy material in agriculture [20] . Nothing comparable has been described for other mucoralean species . In addition to their pathogenicity towards humans , several Lichtheimia species are known as contaminants of several food products ( e . g . cocoa , peanuts , olive products ) [21]–[23] . However , despite the known role of Lichtheimia species in infection and diseases , several Lichtheimia species play an important role in the fermentation of soy products in Asian cuisine [24] . The large evolutionary distance and notable differences in infection strategies between Lichtheimia and the two sequenced mucoralean pathogens indicate that they independently evolved their ability to infect humans by developing specific pathogenesis mechanisms . To gain insight into the genomic differences between these groups of pathogens , here we report the genome sequence of the type-strain of L . corymbifera ( FSU 9682 , CBS 429 . 75 , ATCC 46771 ) which has been shown to be a typical strain in terms of virulence and physiology for this species [25] and compare it to published genomes of mucoralean fungi and other fungal phyla .
The genome of the type-strain of L . corymbifera ( FSU 9682 , CBS 429 . 75 , ATCC 46771 ) was sequenced by a combination of 454 sequencing of a shotgun and 8 kb paired-end library in combination with Illumina sequencing of a paired-end read library ( Materials and methods , Table S1 ) . The final assembly comprises 209 scaffolds with a N50 scaffold size of 367 , 562 nt and a total length of 33 . 6 Mb ( Table 1 ) , which is comparable to the genome size of other zygomycetous fungi [26] . Mucoralean genomes are generally believed to contain large amounts of repetitive elements representing around 35% of the genome [27] . However , analysis of the L . corymbifera genome shows a much smaller content of repetitive elements , with only 4 . 7% of the assembly representing repetitive elements including DNA transposons , LTR and non-LTR retrotransposons ( Table S2 ) . This finding is consistent with the results of the k-mer analyses on the Illumina reads where only low amounts of potential repetitive regions were found . Of note , all previous estimates of repetitive elements in mucoraleans correspond to species with large genomes such as R . oryzae ( 46 Mb; 20% repetitive elements ) , Absidia glauca ( 52 Mb; 35% ) and P . blakesleeanus ( 54 Mb; 35% repetitive elements ) [27] , [28] . Interestingly , a Lichtheimia-specific gene expansion in the heterokaryon incompatibility genes was discovered ( see section gene expansion ) which are involved in the recognition of non-self DNA and may contribute to the low amount of repetitive elements . Another mechanism of protection against transposons and viruses is RNA interference resulting in sequence specific RNA degradation [29] . Several predicted proteins with functional domains associated with this mechanism were found including a dicer-like protein , one argonaute-2 protein and a translation initiation factor 2C homolog . However , the exact effects of these mechanisms on the amount of repetitive elements remain to be determined . Heterozygosity was shown for several fungi including the basal lineage fungus Batrachochytrium dendrobatidis [30] . In order to test for potential heterozygous regions and estimate the genome size of L . corymbifera , k-mer analyses based on the Illumina reads were performed using an algorithm described previously [31] ( Material and Methods ) . Analysis resulted in a relatively clear single peak with a slightly trailing left flank for all k-values ( Figure S1 ) . The distribution could be dissected into three components , each showing a normal distribution with similar variance , but different means and different proportions . The main component represents the potential homozygous part of the genome ( 94% ) , whereas two small components represent the potential heterozygous part of the genome ( 4% ) , and most likely some repeat regions that occur at relatively low frequency ( 2% ) . It has to be noted that the potential heterozygous part is rather small and could as well be explained e . g . by regions that are difficult to sequence and therefore have lower k-mer coverage . The lack of heterozygocity is in accordance with the general assumption that mucoralean fungi are haploid during vegetative growth . Based on the k-mer analysis for different k-mer lengths ( 41 , 59 , 69 , and 79 nt ) a total genome size of around 35 Mb was predicted which is close to our total scaffold length of 33 . 6 Mb ( 96% of k-mer predicted size ) . We annotated 174 tRNAs in L . corymbifera . Although R . oryzae ( 239 ) comprises many more tRNAs , we found unique anticodons among the basal fungi in Lichtheimia: CCC ( Gly ) , AAA ( Phe ) and GAT ( Ile ) . In contrast , only L . corymbifera misses the anticodons CAC ( Val ) , CCT ( Arg ) and TAT ( Ile ) . Three GTA ( Tyr ) tRNAs were predicted with introns in L . corymbifera , while the number was higher in other mucoralean fungi ( up to 10 ) . No selenocysteine and possible suppressor tRNAs were predicted . We found the downstream half of 28S rRNA only , but no 18S rRNA in the current assembly . We expect at least two operons ( 18S – 5 . 8S – 28S rRNA ) as found in R . oryzae . In addition to 5S rRNAs located close to the operons , we were able to identify several independent 5S rRNA copies ( Table S3 ) . Another housekeeping ncRNA , present in all kingdoms of life , is the ribozyme RNase P , which processes tRNAs by cleaving off nucleotides on the 3′ end of tRNAs [32] . We detected this gene as expected in a single copy per genome , but two identical copies are apparently present in the genome of R . oryzae , which may result from whole genome duplication in R . oryzae [28] . The pseudoknot in the centre of the molecule is accredited with the catalytic function and highly conserved in evolution [33] . However , the L . corymbifera candidate varies exceptionally in sequence , while the secondary structure is maintained . Whether the function of the molecule is affected has to be analyzed . The evolutionary related RNase MRP was invented at the origin of eukaryots with dual function: ( a ) initiation of mitochondrial replication and ( b ) separation of 18S rRNA from 5 . 8S rRNA [34] . One copy per basal fungal genome was detected . The signal recognition particle containing a ncRNA part ( SRP RNA ) guides proteins to the endoplasmatic reticulum [35] . One copy was detected in Lichtheimia , whereas two copies were identified in the genome of R . oryzae . Surprisingly , the covariance model of mucoralean fungi , in agreement with Rhizopus , Batrachochytrium and Monosiga ( RF00017 ) is much closer related to metazoans than to other known fungi SRP RNAs . We detected the RNA components of the major spliceosome and collected indications for a functional minor spliceosome . Except for U4 snRNA all five RNAs involved in U2-splicing were detected in Lichtheimia . U4 snRNA was not part of the assembly; however an U4-candidate was identified in the originally sequenced read data . Additionally , four of five RNAs involved in AT-AC-splicing were found . However , several special secondary structures were discovered , which may alter the functionality of the minor spliceosome: ( i ) The third stem of U12 snRNA is atrophied and the last stem is shorter than expected for all basal fungi . ( ii ) U4atac is not detected in Lichtheimia . The other basal fungi show one inconspicuous copy , which is assumed to to be an assembly mistake . However , no similar homologous gene was detected in reads either . ( iii ) The second half of U6atac is highly divergent ( Figure S2 and supplemental material: http://www . rna . uni-jena . de/supplements/lichtheimia/index . html ) . Eleven CD-box snoRNAs and 3 H/ACA snoRNAs were identified , which are mainly conserved in sequence and structure among basal fungi . For further details we refer to the supplemental material ( www . rna . uni-jena . de/supplements/lichtheimia/index . html ) . Additionally , several ncRNA candidates could be proposed , which have to be functionally characterized in future experiments . A riboswitch , binding to thiamine pyrophosphate ( TPP ) was found in all basal fungi . For Lichtheimia a potential telomerase RNA is suggested , which is surprisingly closely related to the shortest known telomerase RNAs in ciliates ( 150 nt Tetrahymena paravorax ) . This is unexpected , since the longest telomerase is known from the fungus Saccharomyces cerevisiae ( 1 , 220 nt ) . Although the alignment of the usually extremely divergent telomerase RNA is very convincing in sequence and secondary structure ( see supplemental material ) , no homologs in another basal fungus and no interacting ciliate protein homolog were found in our current assembly . U7 snRNA is known to interact with the downstream region of histone mRNA for inhibition of degradation . Four similar candidates for this short ncRNA were identified . In eukaryotes , polymerase III transcripts ( e . g . U6 snRNA , RNase P , RNase MRP , SRP RNA , U6atac snRNA ) usually display a typical promoter region: −10 nt TATA box , PSE element , Oct region [36] . Therefore , a search for conserved motifs was conducted in Lichtheimia promoter regions . However , we were not able to identify even one of these motifs . This highlights a possible modified polymerase III activity for basal fungi and has to be investigated in detail in further work . A phylogeny among basal fungi and Schizosaccharomyces pombe as outgroup based on ncRNAs ( except 18S and 28S rRNA ) was reconstructed , see Figure S2 ( B and C ) . In accordance with protein and traditional rRNA phylogeny Lichtheimia groups basal to P . blakesleeanus and the other two investigated fungi . To aid prediction of protein-coding genes , RNA-seq analyses were performed for three different growth conditions in three biological replicates ( see Material and Methods ) . The use of different conditions should ensure a higher number of expressed genes , thereby allowing evidence-based gene predictions for many gene models . On average , each replicate has a 70-fold genome coverage , which sums up to a 630-fold genome coverage ( Table S4 ) . Prediction of protein-coding genes was performed using AUGUSTUS [37] , resulting in 12 , 379 predicted genes . Genes were functionally annotated by comparing to GenBank sequences using BLASTp ( E-value≤10−25 ) , and by scanning for the presence of conserved domains using the InterProScan function of BLAST2GO [38] . BLAST hits were obtained for 7 , 917 genes , InterProScan results were found in 10 , 066 genes and at least one Gene Ontology ( GO ) term was assigned to 7 , 435 genes based on the union of BLAST and InterProScan results . The raw reads of the DNA- and RNA-seq experiments , the final genome assembly , the structural and functional gene prediction are available at http://www . ebi . ac . uk/ena/data/view/PRJEB3978 ) . The genome data are also accessible via HKI Genome Resource ( http://www . genome-resource . de/ ) . An exhaustive comparison of L . corymbifera genome with other 24 completely sequenced genomes including the major fungal groups ( Chytridiomycota , Mucoromycotina , Asco- and Basidiomycota ) was performed . This comparison included the reconstruction of L . corymbifera phylome , which encompasses the complete set of evolutionary histories of L . corymbifera genes ( Material and Methods ) . It was carried out using the previously described PhylomeDB pipeline [39] . In brief , for each L . corymbifera protein-coding gene we searched for homologs , and multiple sequence alignments were built , and Maximum Likelihood analyses were performed to reconstruct a phylogenetic tree . The phylome is available through phylomeDB ( http://phylomedb . org ) , with the phylome ID 245 . The phylome was used to establish phylogeny-based orthology and paralogy relationships among genes in the species considered [40] , and to detect gene expansions ( see below ) . In addition , we used two complementary approaches , gene concatenation and super-tree [41] , to reconstruct the species tree that represents the evolution of the 25 species considered . In the first approach , 58 genes were selected that were present in 21 out of 25 in single copy . Their corresponding alignments were then concatenated and a maximum-likelihood species tree was reconstructed ( Material and Methods ) . In the second approach , 9 , 478 trees present in the phylome were used to build a super-tree using a gene tree parsimony approach , a method which finds the topology that minimizes the total number of duplications in the phylome [42] . Both resulting trees presented a similar topology , which placed L . corymbifera at the base of the other Mucorales species ( Figure 1 ) . The only difference found between the trees by the complementary approaches was the position of Schizosaccharomyces pombe , which appeared at the base of Ascomycota in the super-tree tree while in the concatenated tree it grouped with S . cerevisiae . To assess the level of overlap in genetic content between the different species an all-against-all comparison of the 25 genomes was performed . The results indicate between 50% and 75% of the proteins encoded in the other three Mucorales species had homologs in L . corymbifera ( Figure 1 ) . Surprisingly , this percentage of shared gene content with L . corymbifera was similar to that of Schizosaccharomyces pombe ( 60 . 6% ) , which is higher than that found in the more closely related chytrid B dendrobatidis ( an average of 41 . 1% ) . Figure 1 also shows how these homologs are distributed in differently defined groups . Most interestingly , the fraction of species-specific proteins ( grey bars in the figure ) is particularly high in large genomes ( e . g . , over half of the largest genomes Laccaria bicolor and Puccinia graminis ) . Over 25% of the proteins apparently are specific for L . corymbifera . Since the Lichtheimia lineage separated early in mucoralean evolution we can expect that severe genomic re-arrangements have taken place during evolution , causing substantial differences between the genome structures of L . corymbifera and other Mucorales . Only 57 . 7% of the gene families present in Lichtheimia are also present in at least one of the other mucoralean genomes while only 36 . 7% were found in all four genomes representing 70 . 4% and 53 . 7% of the total L . corymbifera genes , respectively ( Figure 2 A ) . Conserved regions , in terms of gene order , between mucoralean genomes were examined and evaluated with respect to the amount of conserved genes of these regions . A total of 230 regions with a minimum of 3 conserved genes of L . corymbifera were found that were present in at least one of the other genomes . These regions were interspersed over 41 . 1% of the scaffolds but covered only 7 . 6% of the L . corymbifera genome reflecting the high dissimilarity between the mucoralean genomes ( Figure 2 B ) . Only 6 regions were shared with all species . The total number of shared clusters was found to be consistent with the phylogenetic distance between the species ( Figure 2 B+C ) . Genes in the conserved regions are members of different gene families and contain a variety of functional domains . In accordance with former results [28] a higher number of gene families with two members were detected for R . oryzae but also for L . corymbifera as compared to other fungi ( 6 . 99%±0 . 58% ) ( Figure 3 A ) . Whole genome duplication has been previously described for R . oryzae based on the presence of gene duplications and duplication of large genomic regions ( segmental duplications ) [28] . To investigate whether segmental duplications and thus a potential WGD also occur in L . corymbifera , the genomes were scanned for the presence of duplicated regions using GECKO2 [43] , [44] . Consistent with the former findings , our analysis showed a high number of segmental duplications in R . oryzae covering more than 10% of the genome [28] while fewer duplicated regions were found in L . corymbifera covering less than 4% of its genome ( Figure 3 B ) . Thus , the gene duplications seem not to result from recent WGD as in R . oryzae but may result from an ancient genome duplication in mucoralean fungi as suggested by Marcet-Houben et al . [45] which is no longer detectable in the duplicated gene clusters . The possibility of ancient WGD in mucoralean genomes is currently investigated in more detail ( Corrochano et al . , pers . comm . ) . The genome of L . corymbifera also shows increased numbers of gene families with a higher number of genes indicating that gene duplication and the preservation of the gene copies seem to be a common process in mucoralean genomes and may be independent from WGD . This will be further addressed in the next section . In addition to gene duplications shared by all mucoralean fungi a high amount of species-specific duplications was detected . Therefore , the phylome was scanned in search of expansions of protein families that occurred specifically in L . corymbifera . For each tree , ETE [46] was used to find nodes that contained at least five L . corymbifera sequences and no other fungal sequence ( Figure 4 A and B ) . Overlapping expansions were fused when they shared more than half of their members . We found 75 expansions that fulfilled those requirements . Five expansions contained more than 30 members , with the largest containing 331 paralogous genes . In contrast , the large genome of R . oryzae contains approximately twice the number of expansions , with the largest encompassing 1 , 888 members . As some of those expansions are likely the result of the presence of transposons , we scanned them for the presence of transposon-linked domains using the Pfam database and HMMER3 [47] , [48] . If the expansions that contain transposons were excluded , 66 groups of paralogous proteins will be left in L . corymbifera comprising a total of 820 genes ( Figure 4 B ) . The most abundant expanded groups ( with 331 and 242 members ) are rather heterogeneous in terms of functional domains , thus there is no particular function that could be assigned to them . The largest group with a dominating domain contains in total 56 members , of which 52 possess a heterokaryon incompatibility protein ( HET ) domain ( PF06985 ) ( Figure 4 C ) . Interestingly , this domain was so far attributed nearly exclusively to ascomycetes ( with only one exception for the basidiomycete Moniliophthora perniciosa ) , where the HET proteins control somatic allorecognition ( non-self-recognition ) during the formation of heterokaryons [49] . However , Mucorales , opposite to ascomycetes , do not form heterokaryons by fusion of somatic cells but only during sexual reproduction and zygospore-formation . Since HET domain proteins are absent in all other sequenced zygomycetous genomes it is unlikely that they play a general role in the sexual reproduction but seem to be specific for Lichtheimia . Interestingly , these HET genes were differentially regulated under stress conditions . Several copies of the HET domain proteins were down-regulated under iron-depletion and hypoxia . Since these genes are absent in all other mucoralean fungi it is unclear which functions they serve in L . corymbifera . In addition it is unclear where these genes originate since they do not occur in other basal fungi and show only very weak similarity with the HET proteins of dikaryan fungi . Several expansions contain transporters: major facilitator superfamily ( MFS , PF07690 , PF12832 , PF05977 , PF13347 ) , ABC transporters ( PF00005 , PF00501 , PF01061 , PF00664 , PF06422 ) , sugar ( and other ) transporters ( PF00083 ) . In addition , some interesting expansions are connected to the transcription regulation function , which is discussed in more detail in a separate section and signal transduction pathways ( see supplemental Material , Table S5 and S6 ) . Four expanded groups are characterized by the cytochrome P450 ( PF00067 ) domain ( Figure 4 C ) . Interestingly , mucoralean pathogens like L . corymbifera have been shown to be resistant to several antifungals including voriconazole [8] , [50] which could be explained by high copy numbers and isoforms of the target genes . Thus , gene duplication and expansion might be important for the success of L . corymbifera in human infections . Strikingly , these domains ( MFS transporters , HET and cytochrome P450 ) were also the dominant domains in genes which were localized in tandem duplications ( Figure S3 ) . Tandem duplicated genes were found to be present in 42 of the 66 gene expansion groups covering 38% of all genes in the expanded groups ( Figure 4 D ) . However , additional smaller tandems were found which did not fit the criteria of gene expansions . A total of 701 genes are organized in such tandem repeats ( Figure S3 ) . In addition , duplicated genes were frequently found to be located on the same scaffold which may result from older tandem duplications . Thus , tandem duplications and a high amount of gene retentions may give an additional explanation for the high amounts of duplicated genes in L . corymbifera comparable to the observations in plant genomes where segmental duplications ( resulting from WGD ) and tandem duplications play different roles in the enrichment of genes of several gene families [51] , [52] . Tandem duplication and the retention of duplicated genes would be an explanation for the severe differences in the size of gene families between mucoralean fungi with only 53% of gene families with the same size in L . corymbifera and R . oryzae ( see Figure S3 C ) . To investigate if the different gene copies may have different functions and thus may contribute to rapid adaptations to different environmental conditions we analysed the expression of tandem duplicated genes under infection-associated stress conditions ( iron depletion and hypoxia; see Table S7 ) . Differential expression of at least one gene of the tandem clusters under at least one of the conditions was found for 71 tandems . Strikingly , only 7 tandems were co-regulated while in 64 cases expression of the copies was different including six cases were copies were antithetically regulated ( Figure S4 ) . These results are consistent with the hypothesis that the high prevalence and maintenance of duplicated genes leads to diversification of gene functions . Duplicated genes can lead to the diversification of gene functions of the two copies which has been discussed in the section above . In addition , alternative splicing ( AS ) can increase the functional diversity . Gene prediction resulted in 841 alternative splicing events in a total of 683 genes ( 5 . 5% of total genes ) comparable to the situation in S . cerevisiae [53] . Based on the analysis of the RNAseq data alternative splicing could be verified for 273 genes ( 2 . 2% of total genes ) ( Figure S5 A and Table S8 ) . Alternative donor and acceptor are the dominant groups of alternative splicing events ( >75% of the total events ) which is similar to the situation in several higher eukaryotes [54] ( Figure S5 A ) . Comparison of alternatively spliced genes with genes in tandem duplications and gene expansions showed that only 12 ( 4 . 4% of genes with AS ) in these groups are also alternatively spliced . If AS occurs in tandem duplicated genes , it occurs in only one of the copies except in one case . This is in accordance with recent results in S . cerevisiae which show that duplicated genes can replace one alternatively spliced gene and that alternative splicing is often lost after gene duplication [55] . To test if AS plays a role in the stress adaptation of L . corymbifera we analysed the potential alteration in alternative splicing pattern during stress adaptation . Significant changes were only detected for 16 and 23 genes under iron depletion and hypoxia ( <0 . 2% of the total genes ) , respectively ( Figure S5 B ) . Based on the high incidence of gene duplication and the differential expression of the copies as well as the comparably low number of alternatively spliced genes , maintenance of duplicated genes seems to play a more important role for the generation of functionally distinct paralogs than alternative splicing . The genome of L . corymbifera represents the first insight into the genome structure of basal mucoralean pathogens . Despite the growing recognition of Mucorales as life-threatening clinically important human pathogens , little is known about the virulence traits of these fungi . The high dissimilarity between L . corymbifera and the other sequenced mucoralean pathogens R . oryzae and M . circinelloides in both evolutionary and functional sense underlines the importance of additional genome projects . This study revealed a high proportion of duplicated and expanded genes in the L . corymbifera genome comparable to the situation in R . oryzae . However , clear evidence for a WGD can be detected only for R . oryzae , but not for L . corymbifera indicating that additional mechanisms contribute to the higher incidence of duplicated genes in mucoralean fungi . Tandem repeats seem to be an important source for gene duplication in L . corymbifera and may explain the rapid development of lineage-specific gene duplication and expansion in mucoralean fungi . Several species-specific gene duplications point at potential virulence traits including iron uptake genes , hydrolytic enzymes and genes which may contribute to resistance against antifungal agents like azoles ( cytochrome P450 gene expansion ) . In contrast , alternative splicing does not seem to play an important role in the generation of orthologs and the adaption to stress conditions . Based on these results , we postulate a relationship between genome fluidity by the generation and retention of additional gene copies and dynamics of adaptation to new environments . Higher genome flexibility results in a higher likelihood for a saprobic zygomycete to become a pathogen . In addition we were able to shed light on the genes involved in iron uptake , which is a crucial step for virulence and thus for the development of an infection . We could identify additional genes which might be involved in iron-uptake besides the known virulence factor FTR1 L . corymbifera including transcription factors , siderophore transporters and a potential regulator involved in siderophore biosynthesis that has not been described in mucoralean fungi . Our data represent a valuable resource for future research and the understanding of infection-associated mechanisms of mucoralean pathogens .
A combination of Illumina and 454 sequencing was used for the L . corymbifera genome . A shotgun library and an 8 kb paired-end library were created and sequenced on a half plate on a Roche GS FLX Titanium each resulting in 1 , 168 , 226 shotgun reads ( 505 , 023 , 982 nt ) and 519 , 989 paired-end reads ( 76 , 603 , 029 nt ) . In addition , a standard paired-end read library was prepared and sequenced in one channel Illumina HiSeq2000 ( 100 bp paired-end reads ) resulting in 264 , 907 , 616 raw reads ( 26 , 490 , 761 , 600 nt ) and 12 , 614 , 650 filtered and downsampled reads ( 1 , 261 , 465 , 000 nt ) . The 454 reads were separately assembled using Newbler ( 454 Life Sciences ) and Mira [88] and both assemblies were unified using minimus2 [89] . The Illumina reads were used to solve homopolymeric regions using Nesoni ( http://bioinformatics . net . au/software ) . This approach resulted in a total of 1 , 214 contigs ( ≥500 nt ) with a total of 41 , 405 , 106 nt and a N50 of 66 . 718 nt . Finally , the contigs were mapped on Newbler predicted scaffolds using MUMmer [90] resulting in 209 scaffolds with a total length of 33 . 6 Mb ( for statistics refer to Table 1 ) . The raw DNA-seq reads and the resulting genome assembly is available at EMBL under the study accession number PRJEB3978 ( http://www . ebi . ac . uk/ena/data/view/PRJEB3978 ) . Scaffolds of L . corymbifera were searched for repeats by Repbase and the server version of Censor [91] , [92] ( http://www . girinst . org/censor/index . php ) using the eukaryotic repeat database . The analysis was performed on the Ilumina reads with an algorithm described in the potato genome paper [31] . The algorithm was used to write a custom perl program . Based on the fastq data of the Illumina reads k-mers of 41 , 59 , 69 , and 79 nt were detected and analyzed . Component estimation was done manually in R . A local version of tRNAscan-SE v . 1 . 23 [93] with parameters –omlfrF was used for the detection of tRNAs . Ouput files are supported in the supplemental material ( http://www . rna . uni-jena . de/supplements/lichtheimia/index . html ) . With RNammer -S euk -m lsu , ssu , tsu -gff ( v . 2 . 1 ) [94] rRNAs were detected . The 1973 ncRNA classes currently available at RFAM ( v . 10 . 1 ) [95] were downloaded for homologous search . These classes were predicted with ( I ) BLAST ( v . 2 . 2 . 25 ) [96] with an E-value<10−4 ( II ) with infernal [97] using covariance models from RFAM and ( III ) by hand as indicated in main text . Genes discovered in the reads only were found with rnabob [98] in combination with various programs of the RNAViennaPackage v . 2 . 0 . 2 ( http://www . tbi . univie . ac . at/~ivo/RNA/ ) . All ncRNA genes are available at the supplemental material in gff and fasta format ( http://www . rna . uni-jena . de/supplements/lichtheimia/index . html ) . Additionally , sequence-structure-alignments for each RFAM-ncRNA class in stockholm format are provided . Motif search in promoter regions of polymerase III transcripts was performed with MEME ( v . 4 . 8 . 1 ) [99] , rnabob and by hand . Synteny analysis: for all of our identified ncRNA positions in L . corymbifera and R . oryzae , five direct upstream and downstream located genes and their function were extracted , according to protein-annotation files . Pairwise alignments of syntenic proteins with a ) -p blastn and b ) -p tblastn and a minimum E-value of E<10−4 were performed . For ncRNA-phylogeny reconstruction the best scored ncRNA per ncRNA family was joined , which was identified in all species , except 18S and 28S rRNA , and S . pombe used as outgroup . A multiple alignment was created by Mafft with the L-INS-i method , 1000 iterations as module in the EPoS framework for phylogenetic analysis [100] . Out of this alignment we constructed a Neighbour Joining Tree ( Kimura correction model , 1000 bootstrap replicates ) and Mr . Bayes ( v . 3 . 1 . 2; two runs with each four chains and 5 , 000 , 000 generations ) . Evidence-driven gene prediction was performed using AUGUSTUS v2 . 7 [37] using the gene models from Rhizopus oryzae prediction was supported by the incorporate pooled Illumina RNA-seq data from three biological replicates of three different physiological conditions ( control , hypoxia , iron depletion ) sequenced on Illumina HiSeq 2000 . After the raw RNA-Seq data were quality trimmed- using btrim [101] , the data were pooled and mapped using the splice-junction mapper tophat2 [102] . From this mapping data the AUGUSTUS protocol ( http://bioinf . uni-greifswald . de/bioinf/wiki/pmwiki . php ? n=IncorporatingRNAseq . Tophat ) was followed to create hints for gene structures in an iterative manner . Finally , the hints were incorporated during the AUGUSTUS predcition based on the Lichtheimia genome using the metaparameters of R . oryzae . For functional annotation predicted protein-coding genes were analyzed by BLASTp in BLAST2GO [38] with a minimum E-value of E≤10−25 and a HSP length cut-off of 33 amino acids . Conserved domains were identified using the InterProScan function of BLAST2GO and GO mapping was performed based on the BLAST and InterProScan results . Genome annotations are available at the ENA under the study accession number PRJEB3978 ( http://www . ebi . ac . uk/ena/data/view/PRJEB3978 ) . Paired-end RNA-seq data for three biological replicates of two infection-associated conditions ( i . e . , iron-depletion and hypoxia ) and a control treatment was obtained . L . corymbifera was grown on SUP agar [103] plates for 7 days at 37°C . Spores were washed off with sterile PBS , washed with PBS and counted using a Thoma chamber . Erlenmeyer flasks ( 500 ml ) containing 100 ml of chemical defined medium ( 1 . 7 g/l YNB w/o amino acids and ammonium sulphate , 20 g/l Glucose , 5 g/l ammonium sulphate , 50 mg/l arginine , 80 mg/l aspartic acid , 20 mg/l histidine , 50 mg/l isoleucine , 100 mg/l leucine , 50 mg/l lysine , 20 mg/l methionine , 50 mg/l phenylalanine , 100 mg/l threonine , 50 mg/l tryptophane , 50 mg/l tyrosine , 20 mg/l valine ) [60] were inoculated with 107 spores and grown for 16 h at 37°C under shaking . Afterwards ( i ) cultures were grown for additional 2 h under these conditions , ( ii ) the iron chelator bathophenanthrolinedisulfonic acid ( BPS , Sigma ) was added to a final concentration of 200 µm and cultures were incubated for additional 2 h under previous conditions or ( iii ) cultures were subjected to hypoxic conditions ( 1% oxygen , 5% CO2 ) and incubated for 2 h at 37°C under shaking . The mycelium was separated from the medium using a miracloth filter ( Millipore ) and immediately frozen in liquid nitrogen . For RNA isolation the mycelium was grounded using mortar and pestle under liquid nitrogen and total RNA was isolated using the RNAeasy Plant kit ( Quiagen ) according to the manufacturer's instructions . Sequencing was performed using Illumina HiSeq 2000 . Raw reads were quality-filtered using btrim [101] and mapped to the genome using tophat2 [102] ( parameters: –no-discordant –no-mixed –b2-very-sensitive –max-intron-length 5000 ) . Differentially expressed genes were identified with EdgeR [104] which also adjusted obtained p-Values for multiple testing . Transcripts with an absolute fold-change≥2 and an adjusted p-Value≤0 . 01 were considered differentially expressed . Results are available in Table S7 . The phylome , meaning the complete collection of phylogenetic trees for each gene in a genome , was reconstructed for the genome of L . corymbifera . 24 other fungal species were included in the reconstruction . A rough draft of the proteome of Mortierella alpina ( [26]; PUBMED ID:22174787 ) was predicted using AUGUSTUS [32] due to the lack of a publicly available proteome . The phylome was reconstructed using an automated pipeline previously described in [39] . Briefly , for each protein in the L . corymbifera genome a Smith-Waterman search was performed against the fungal proteome database . Results were filtered using an e-value cut-off E<1e−5 and a continuous overlapping region of 0 . 5 . At most 150 homologous sequences for each protein were accepted . Homologous sequences were then aligned using three different programs: MUSCLE v3 . 8 [105] , MAFFT v6 . 712b [106] , and kalign ( http://www . biomedcentral . com/1471-2105/6/298/ ) ] . Alignments were performed in forward and reverse direction ( i . e . using the Head or Tail approach [107] ) , and the 6 resulting alignments were combined with M-COFFEE [108] . This combined alignment was trimmed with trimAl v1 . 3 [109] ( consistency-score cut-off 0 . 1667 , gap-score cut-off 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 [110]; the likelihood of this topology was computed , allowing branch-length optimization , using 7 different models ( JTT , LG , WAG , Blosum62 , MtREV , VT and Dayhoff ) , as implemented in PhyML v3 . 0 [111]; the model best fitting the data , as determined by the AIC criterion [112] , was used to derive ML trees . Four rate categories were used and invariant positions were inferred from the data . Branch support was computed using an aLRT ( approximate likelihood ratio test ) based on a chi-square distribution . Resulting trees and alignments are stored in phylomeDB [39] ( http://phylomedb . org ) , with the phylomeID 245 . Trees were scanned using ETE v2 [46] . Orthologs between L . corymbifera and the other species included in the phylome were based on phylogenies obtained during phylome reconstruction . A species-overlap algorithm , as implemented in ETE v2 [46] , was used to infer orthology and paralogy relationships . Briefly the algorithm decides whether a node in a tree is a speciation of a duplication node depending on the overlap of the species branching from the node . Overlap between those species will indicate a duplication node . Otherwise a speciation node will be considered . The species tree was build using a concatenation method . 58 single-copy proteins that appeared in at least 21 of the 25 genomes were selected . After concatenation , the alignment was trimmed using trimAl [109] . Columns with more than 50% of gaps were removed . A conservation score of 50% of the alignment was used . The final alignment contained 46 , 793 positions . The tree was reconstructed using phyML [111] . LG model [113] was selected and a 4-categories GAMMA distribution was used . Bootstrap was obtained by creating 100 random sequences using SeqBoot from the phylip package . A tree was then reconstructed for each sequence and the consensus tree was inferred using phylip . All the nodes in the species tree had a bootstrap of 100 . Additionally a species tree based on the super-tree reconstruction program DupTree [42] was reconstructed . The input contained the 9 , 478 trees obtained during phylome reconstruction . Both species trees showed a similar topology . The only difference pertained to the position of S . pombe . In the concatenated tree it appeared grouped with S . cerevisiae while in the super-tree it appeared in its correct position at the base of Ascomycota . This difference was collapsed into a multifurcation for the tree in figure 1 . For the detection of conserved regions , all genomes were modeled as strings of integers . BLAST analyses [96] were performed for all proteins in the four mucoralean genomes all-against-all , with an E-value threshold of 0 . 1 . Homology families IDs were assigned to the protein-coding genes using GhostFam [114] with default parameters . Genomes were transformed into strings of gene IDs , which were then used as input for the reference gene cluster implementation in Gecko2 [43] , [44] . The two parameters for the algorithm were the minimum size of the reference cluster/hypothetical conserved region “s” and the maximal distance “δ” ( insertion or deletion of a gene ) . For every hypothetical gene cluster larger than s on the reference genome , all other genomes were tested for approximate occurrences of this reference gene cluster . The L . corymbifera genome was used as a reference genome and searched for gene clusters with parameters s = 3 ( minimum size of the reference gene cluster ) and δ = 0 ( number of insertions and deletions ) , s = 4/δ = 1 , s = 5/δ = 2 , s = 6/δ = 3 and s = 7/δ = 4 . Results of the different filter settings were combined and overlapping clusters were eliminated . Local rearrangements and duplications within the cluster occurrences were not punished . All regions that had approximate occurrences in at least one other genome were reported . If multiple occurrences did intersect , only the best scoring one was reported . To detect duplicated regions in the mucoralean species , each genome was analysed individually by using the single contigs as reference . As for the detection of conserved regions , the same homology assignment and parameters of s = 5 and δ = 2 were used . All regions with approximate occurrences in at least one other contig or the reference contig were reported , unless they intersected . Tandem duplications were defined by at least two genes assigned to the same GhostFam gene family and a maximum of three genes between the copies . Predicted transcripts of the genomes were separated in alternative splicing events by Astalavista [24] . Events of predicted transcripts that contain splice-junctions have been confirmed by the number of split-mappings that confirm each of the exon-exon junctions ( Table S8 ) . For a read to support a splice-junction , the left part of the read was required to be included in one exon , and the right part had to be included in the other exon of a splice junction , with the first/last position before/after the split matching exactly the position of the predicted intron . Genome data of Aspergillus fumigatus [115] , Aspergillus nidulans [116] , Batrachochydrium dendrobatidis , Cryptococcus neoformans , Encephalitozoon cuniculi [117] , Rhizopus oryzae [28] , Paracoccidioides brasiliensis , Schizosaccharomyces pombe [118] , Nosema ceranae [119] , Nematocida parisii [120] , Puccinia graminis [121] , Ustilago maydis and Coprinus cinerea [122] are genome sequencing projects of the Broad Institute of Harvard and MIT ( http://www . broadinstitute . org/ ) ( see Table S9 for detailed citations ) . Phycomyces blakesleeanus , Phanerochaete chrysosporium [123] , Laccaria bicolor [124] , Mucor circinelloides , Nematostella vectensis [125] , Monosiga brevicollis [126] and Serpula lacrymans [127] genomic data were obtained from Joint Genome Institute ( JGI ) . These sequence data were produced by the US Department of Energy Joint Genome Institute http://www . jgi . doe . gov/ in collaboration with the user community . The genomes of Homoloaphlyctis polyrhiza [128] and Mortierella alpina [26] were obtained from Genbank ( Hp: PRJNA68115; Ma: PRJNA41211 ) . The Neurospora crassa genome [129] was obtained from UniProt reference genomes . The Saccharomyces cerevisiae genome was obtained from Saccharomyces Genome database ( SGD ) ( see Table S9 ) [53] .
|
Lichtheimia species are ubiquitous saprophytic fungi , which cause life-threating infections in humans . In contrast to the mucoralean pathogen R . oryzae , Lichtheimia species belong to the ancient mucoralean lineages . We determined the genome of L . corymbifera ( formerly Mycocladus corymbifer ex Absidia corymbifera ) and found high dissimilarities between L . corymbifera and other sequenced mucoralean fungi in terms of gene families and syntenies . A highly elevated number of gene duplications and expansions was observed , which comprises virulence-associated genes like proteases , transporters and iron uptake genes but also transcription factors and genes involved in signal transduction . In contrast to R . oryzae , we did not find evidence for a recent whole genome duplication in Lichtheimia . However , gene duplications create functionally diverse paralogs in L . corymbifera , which are differentially expressed in virulence-related compared to standard conditions . In addition , new potential virulence factors could be identified which may play a role in the regulation of the adaptation to iron-limitation . The L . corymbifera genome and the phylome will advance further research and better understanding of virulence mechanisms of these medically important pathogens at the level of genome architecture and evolution .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[
"sequencing",
"techniques",
"organismal",
"evolution",
"medicine",
"and",
"health",
"sciences",
"genome",
"evolution",
"pathology",
"and",
"laboratory",
"medicine",
"microbiology",
"genome",
"sequencing",
"genomic",
"databases",
"phylogenetics",
"fungi",
"fungal",
"evolution",
"genome",
"analysis",
"microbial",
"evolution",
"molecular",
"biology",
"techniques",
"sequence",
"analysis",
"genome",
"complexity",
"mycology",
"medical",
"microbiology",
"microbial",
"pathogens",
"comparative",
"genomics",
"pathogenesis",
"evolutionary",
"systematics",
"molecular",
"biology",
"evolutionary",
"genetics",
"sequence",
"databases",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"evolutionary",
"biology",
"computational",
"biology",
"organisms"
] |
2014
|
Gene Expansion Shapes Genome Architecture in the Human Pathogen Lichtheimia corymbifera: An Evolutionary Genomics Analysis in the Ancient Terrestrial Mucorales (Mucoromycotina)
|
Toll-like receptors ( TLRs ) are sentinel receptors of the innate immune system . TLR4 detects bacterial lipopolysaccharide ( LPS ) and TLR5 detects bacterial flagellin . A common human nonsense polymorphism , TLR5:c . 1174C>T , results in a non-functional TLR5 protein . Individuals carrying this variant have decreased mortality from melioidosis , infection caused by the flagellated Gram-negative bacterium Burkholderia pseudomallei . Although impaired flagellin-dependent signaling in carriers of TLR5:c . 1174C>T is well established , this study tested the hypothesis that a functional effect of TLR5:c . 1174C>T is flagellin-independent and involves LPS-TLR4 pathways . Whole blood from two independent cohorts of individuals genotyped at TLR5:c . 1174C>T was stimulated with wild type or aflagellated B . pseudomallei or purified bacterial motifs followed by plasma cytokine measurements . Blood from individuals carrying the TLR5:c . 1174C>T variant produced less IL-6 and IL-10 in response to an aflagellated B . pseudomallei mutant and less IL-8 in response to purified B . pseudomallei LPS than blood from individuals without the variant . TLR5 expression in THP1 cells was silenced using siRNA; these cells were stimulated with LPS before cytokine levels in cell supernatants were quantified by ELISA . In these cells following LPS stimulation , silencing of TLR5 with siRNA reduced both TNF-α and IL-8 levels . These effects were not explained by differences in TLR4 mRNA expression or NF-κB or IRF activation . The effects of the common nonsense TLR5:c . 1174C>T polymorphism on the host inflammatory response to B . pseudomallei may not be restricted to flagellin-driven pathways . Moreover , TLR5 may modulate TLR4-dependent cytokine production . While these results may have broader implications for the role of TLR5 in the innate immune response in melioidosis and other conditions , further studies of the mechanisms underlying these observations are required .
Melioidosis is a tropical infection caused by the flagellated Gram-negative bacterium Burkholderia pseudomallei [1] . The disease commonly presents as pneumonia and sepsis in Southeast Asia and northern Australia . The mortality rate from infection with this bacterium can exceed 40% [2] . Hawn et al . first described a nonsense human Toll-like receptor ( TLR ) -5 polymorphism , TLR5:c . 1174C>T ( rs5744168 ) , that eliminates both the transmembrane and cytoplasmic signaling regions of the protein , rendering it unable to signal in response to flagellin [3] . Approximately 10% of white individuals of European ancestry [3] and Thai individuals [4] carry the minor T allele . While we have previously found that while the TLR5:c . 1174C>T polymorphism does not affect susceptibility to infection with melioidosis [4] , the variant is associated with significantly improved survival of individuals with culture-proven melioidosis [5 , 6] . Furthermore , we have studied the effect of the variant on cytokine expression induced by stimulation of whole blood ex vivo with B . pseudomallei . We have found that IL-6 and IL-10 levels were reduced in white North American carriers of the variant allele while IL-10 levels were reduced in Thai carriers [5] . Initially , these associations would seem to involve a mechanism including flagellin , as this is the bacterial motif that activates TLR5 . However , we have previously reported that lipopolysaccharide ( LPS ) , and not flagellin , is primarily responsible for the whole blood innate immune response to killed B . pseudomallei , highlighting the critical role of TLR4 in the host response of this organism [7] . Moreover , multiple TLR family members form heterodimers to facilitate innate immune signaling , and TLR5 in particular has been reported to bind TLR4 in transfected cells stimulated with flagellin [8] . We therefore considered the possibility that a TLR5/TLR4 signaling interaction may underlie the effect of the TLR5:c . 1174C>T polymorphism on inflammatory responses and survival in B . pseudomallei infection . In this study , we tested the hypothesis that TLR5 can modulate the flagellin-independent , TLR4-dependent innate immune response .
To construct a markerless in-frame deletion of B . pseudomallei flagellin , fliC ( BP1026b_I3555 ) , primers were designed to amplify approximately 800 bp regions up- and downstream of the gene including the first and last four codons of fliC , respectively . The up- and downstream fragments were joined by an overlap sequence using splicing by overlap extension as described previously [9] . The deletion construct was cloned into the suicide vector pMo130 , which carries a kanamycin resistance gene for the selection of transformants and a sacB gene as a counter-selectable marker [10] . Next , the plasmid was transferred from E . coli SM17 λpir to B . pseudomallei 1026b by mating . Transconjugants were selected on LB plates containing irgasan ( 25 μg/ml ) and kanamycin ( 300 μg/ml ) and streaked onto LB agar supplemented with 15% sucrose to select for transformants that resolved the merodiploid state . Colonies were screened for excision of the plasmid by replica plating on LB/kanamycin ( 500 μg/ml ) and LB plates . The deletion of fliC in kanamycin sensitive colonies was verified by PCR and swimming motility assays . B . pseudomallei 1026b and B . pseudomallei ΔfliC were grown , shaking , in LB broth for six hours in air at 37°C and heat-killed before use in in vitro assays [5] . To test B . pseudomallei ΔfliC , HEK293 cells were used as previously reported with some modifications [11] . Cells were cultured in a 96 well flat-bottomed tissue culture plate at 15–50 × 103 cells/well in DMEM plus 10% fetal bovine serum ( FBS ) and 1% L-glutamine at 37°C under 5% CO2 . The following day cells at 60–80% confluency were transiently transfected using PolyFect transfection reagent ( Qiagen , Valencia , CA ) . NF-κB-ELAM firefly luciferase 80 ng/well was transfected with β-actin-Renilla luciferase 8 ng/well , and huTLR5 ( cloned into pEF6 , a gift from Dr . Thomas R . Hawn [3] ) 175ng/well . Cells were immediately stimulated with IL-1β , flagellin from S . Typhimurium strain SL1344 deficient in flgM ( engineered to overproduce flagella , a gift from Kelly Smith , University of Washington [12] ) , B . pseudomallei 1026b or B . pseudomallei ΔfliC and incubated for 24 hours . Purity of the flagellin was confirmed by limulus amebocyte lysate assay and SDS-PAGE . Cells were then lysed with passive lysis buffer ( Promega , Madison , WI ) and NF-κB activation was determined in 10 μl of lysate by the ratio of firefly to Renilla luciferase light emission using the Dual Luciferase Reporter System ( Promega ) . For TLR4/TLR5 co-transfections , HEK293 cells stably transfected with MD-2 , CD14 , and TLR4 were obtained from Invivogen ( Catalog # 293-htlr4md2cd14 ) . Cells were cultured in a 96-well flat-bottomed tissue culture plate at 50 × 103 cells/well in cell growth media ( DMEM , 4 . 5 g/l glucose , 50 U/ml penicillin , 50 μg/ml streptomycin , 100 μg/ml Normocin , 10 μg/ml blasticidin , 50 μg/ml Hygromycin B Gold , plus 10% fetal bovine serum ( FBS ) and 1% L-glutamine ) . The following day cells were transiently transfected using PolyFect transfection reagent ( Qiagen , Valencia , CA ) and the following DNA: NF-κB-ELAM firefly luciferase 50 ng/well , HSV thymidine kinase-Renilla luciferase ( pRL-TK , Promega ) 5 ng/well , and huTLR5 ( cloned into pEF6 ) and empty pEF6 vector at 1–10 ng/well as indicated . The transfection complexes were allowed to incubate with the cells for 24 hours . The following morning , the media in the wells was aspirated and the cells were then stimulated with ultrapure Escherichia coli O111:B4 LPS ( List Biologicals , Campbell , CA ) or ultrapure S . Typhimurium flagellin ( Invivogen ) suspended in cell growth media . After four hours , cells were lysed with passive lysis buffer and NF-κB activation was determined as above . For ex vivo whole blood immunoassay studies , 380 μl of fresh whole blood in citrate mixed 1:1 with RPMI media was added to pre-prepared plates containing 20 μl of stimulants . Studies were performed at Harborview Medical Center ( HMC ) in Seattle , USA , and in Sunpasitthiprasong Hospital , Ubon Ratchathani , Thailand as previously described [5 , 13] . For the HMC study the stimulant analyzed was log phase heat-killed B . pseudomallei 1026b ΔfliC 2 . 5 x 106 CFU/ml . For the Sunpasitthiprasong Hospital study the stimulants analyzed were B . pseudomallei K96243 LPS 10 ng/ml , B . pseudomallei V688 LPS 10 ng/ml , B . pseudomallei 558 LPS 10 ng/ml , S . Typhimurium strain SL1344 deficient in flgM flagellin 500 ng/ml , Pam3CSK4 100 ng/ml ( Invivogen , San Diego , CA ) , and E . coli 0111:B4 LPS 10 ng/ml ( List Biological Laboratories , Campbell , CA ) . B . pseudomallei LPS preparations ( a gift from Bob Ernst , University of Maryland ) were generated using a hot phenol/water extraction method after growth of bacteria in lysogeny broth ( LB ) supplemented with 1 mM MgCl2 at 37°C [14] . Subsequently , LPS was treated with RNase A , DNase I and proteinase K to ensure purity from contaminating nucleic acids and proteins [15] . The LPS sample was additionally extracted to remove contaminating phospholipids [16] and TLR2 contaminating proteins [17] . We have previously demonstrated the lack of TLR2 activation by B . pseudomallei LPS prepared in this fashion [11] . Plates were incubated at 37°C on a shaking incubator under 5% CO2 for six hours before being spun down and plasma removed and frozen . For the HMC study , plasma cytokines were subsequently assayed in duplicate using an electrochemiluminescence imager ( Mesoscale Discovery ) . For the Sunpasitthiprasong Hospital study , cytokines were assayed in duplicate on a multiplex bead system ( Luminex , Austin , TX ) using reagents from R&D Systems . For each subject , a complete blood count with differential was performed in the clinical laboratory at the time of phlebotomy . Eight mL of whole blood was drawn from healthy subjects into BD vacutainer cell preparation tubes with sodium citrate ( BD , Franklin Lakes , NJ ) and centrifuged at 1500 RCF for 20 minutes . The PBMC layer was collected and washed twice with PBS , according to the manufacturer’s instructions . Monocytes were isolated from this preparation of PBMCs using the Miltenyi Biotec monocyte isolation kit II ( Auburn , CA ) . With this kit , non-monocytes were indirectly magnetically labeled using a cocktail of biotin-conjugated antibodies and anti-biotin microbeads . Enriched monocytes were collected after running the preparation on LS columns , retaining all of the non-monocytes in the columns . This procedure was completed according to the manufacturer’s instructions . Isolated monocytes were plated in RPMI media plus 10% FBS in a 96 well plate , with 50 , 000 monocytes per well . Monocytes were stimulated with ultrapure E . coli O111:B4 LPS ( List Biological Laboratories ) for four hours . RNA was isolated from monocytes using the Applied Biosystems Nucleic Acid Lysis Solution and the ABI Prism 6100 Nucleic Acid PrepStation ( Applied Biosystems , Grand Island , NY ) , according to the manufacturer’s instructions . RNA expression was determined using the SensiFAST Probe Lo-Rox One Step Kit ( Bioline , Taunton , MA ) . For each PCR reaction , 2ul RNA , 0 . 5 μl Taqman primers ( Life Technologies , Grand Island , NY ) , 5 μl SensiFAST mix , 0 . 1 μl reverse transcriptase , 0 . 2 μl RNAase inhibitor , and 2 . 2 μl water was added to each well of a 384 well plate . The reactions were completed using a Viia7 Real-Time PCR System ( Applied Biosystems ) , according to the thermocycling conditions suggested by Bioline . The following commercially available Taqman primers from Applied Biosystems were used for the PCR reactions: GAPDH HS02758991_g1 and TLR4 HS00152939_m1 . RNA was isolated from transfected THP1-Dual Cells using the Promega ReliaPrep RNA Cell Miniprep System ( Promega , WI ) , according to the manufacturer’s instructions . RNA expression was determined using the ABI High Capacity cDNA Reverse Transcriptase Kit ( Applied Biosystems ) and Bio-Rad SsoAdvanced Universal SYBR Green Supermix ( Bio-Rad , CA ) . The reactions were completed using a ViiA7 Real-Time PCR System ( Applied Biosystems ) , according to the thermocycling conditions suggested by Bio-Rad . The following commercially available primer assays from Bio-Rad were used for the PCR reactions: TLR5 ( qHsaCID0009328 ) , TLR4 ( qHsaCED0037607 ) , and UBC ( qHsaCED0023867 ) . For whole blood stimulation studies , DNA extraction and genotyping was performed as previously described [5] . For primary human monocyte studies , DNA was extracted from saliva using Oragene kits ( DNA Genotek ) and genotyping was performed using ABI TaqMan assays on an ABI Prism 7900 . Fasting blood samples for whole blood stimulation were obtained from healthy white participants in a Harborview Medical Center ( HMC ) inflammatory response research study [5] . Similarly , healthy Thai subjects donating blood at the blood donation center at Sunpasitthiprasong Hospital in 2010 were also recruited for investigation of inflammatory responses . Enrollment criteria and blood processing for both studies has been previously described [5 , 7 , 13] . Blood for monocyte isolation was obtained from healthy subjects of Southeast Asian ancestry at Harborview Medical Center from whom saliva was obtained and genotyped . Subjects were between the ages of 18 and 65 years , weighed between 100 and 350 pounds , did not smoke , did not have any chronic medical conditions , were not pregnant and had not given birth within the preceding nine months , did not have symptoms of infection within the past two weeks , had not received a vaccination in the past six weeks , and did not take any anti-inflammatory , antimicrobial , or prescription ( other than oral contraceptive ) medication . Subjects were asked to refrain from any heavy exercise or alcohol use for 24 hours and to fast overnight before the blood draw . The University of Washington Human Subjects Division Institutional Review Board; the Ethical Review Committee for Research in Human Subjects , Sunpasitthiprasong Hospital , Ubon Ratchathani , Thailand; and the Ethics Committee of the Faculty of Tropical Medicine , Mahidol University , Bangkok , Thailand approved the studies . All participating subjects were adults who gave written informed consent . The chromosomal DNA of Bp82 , a ΔpurM mutant of B . pseudomallei [18] , was isolated using the Promega Wizard Genomic DNA Purification Kit ( WI , USA ) . Primers 5’-TTTTGGATCCATGCTCGGAATCAACAGCAACATTAAC-3’ ( forward primer ) and 5’-TTTTGCGGCCGCTTATTGCAGGAGCTTCAGCACTTGC-3’ ( reverse primer ) [19] were designed to amplify the full-length flagellin gene . The resultant PCR amplified flagellin gene was cloned into the plasmid pCR 2 . 1-TOPO ( CA , USA ) according to the manufacturer’s protocol . This PCR construct was used to generate a N-terminal 6X-His-tag-flagellin protein under contract by NOVOprotein ( Summit , NJ , USA ) . The endotoxin content of the recombinant flagellin was verified to be <100 EU/mg . THP1-Dual Cells ( NF-κB-SEAP and IRF3-Lucia luciferase Reporter Monocytes , Invivogen , CA ) were differentiated over a period of 72 hours with Vitamin D3 ( final concentration 10 pM/ml ) and plated at 50 , 000 cells per well in media containing D3 ( 10 pM/ml ) . The next day , cells were transfected with Silencer Select siRNA ( Ambio , MA ) against TLR5 ( siRNA ID—s14197 ) or a Silencer Select Negative Control No . 1 siRNA ( # 4390843 ) at a final concentration of 5 nM using Lipofectamine RNAiMAX transfection Reagent ( Life Technologies , CA ) . After two days , some cells were lysed for RNA extraction and remaining cells were stimulated with either E . coli LPS at varying doses of recombinant B . pseudomallei flagellin 100 ng/ml as a positive control . The following day , supernatants were collected and assayed for SEAP using QUANTI-Blue detection reagent and IRF3 using QUANTI-Luc luciferase detection reagent according to the manufacturer’s instructions . Additional supernatants were also stored at -80°C until ready to assay . TNF-α and IL-8 concentrations were determined using DuoSet ELISA ( R&D , MN ) using the manufacturer’s protocol . Continuous in vitro data expected to follow a normal distribution are reported as mean ± standard deviations; comparisons between two groups were made using the t test . Non-normal data are reported as individual values and median; comparisons between two groups are made using the rank sum test . Cytokine values were normalized to monocyte counts and log10 transformed before analysis by linear regression , adjusting in the Sunpasitthiprasong cohort for age , sex , and batch . For gene expression , quantitative real-time RT-PCR Ct values observed for each sample were normalized to the reference gene values and reported as 2-ΔCt . Statistics were performed with Stata 14 . 2 ( College Station , TX ) . A two-sided P value ≤0 . 05 was considered significant .
We and others have shown that carriers of TLR5:c . 1174C>T have significantly impaired flagellin sensing [3 , 5] , and we have reported that B . pseudomallei-induced IL-10 and IL-6 concentrations in blood from healthy white North American subjects are lower in carriers of the variant while IL-10 levels are reduced in Thai carriers of the variant [5] . To determine whether the effect of this variant on blood cytokine production was flagellin-dependent , we used a markerless in-frame deletion method to create a B . pseudomallei mutant lacking fliC , the gene that encodes flagellin . To assess the ability of the ΔfliC mutant to induce TLR5-dependent signaling , we transfected TLR5 into HEK293 cells that only minimally express TLR5 [20] . We stimulated these cells with heat-killed wildtype B . pseudomallei 1026b or with the heat-killed ΔfliC mutant . We confirmed that the ΔfliC mutant did not induce any TLR5-dependent NF-κB activation with a luciferase assay ( Fig 1 ) . We then stimulated whole blood from healthy subjects with B . pseudomallei ΔfliC and compared IL-10 and IL-6 cytokine levels by TLR5:c . 1174C>T variant genotype . As was the case following stimulation of blood with wild type B . pseudomallei , carriers of the variant produced modestly but significantly attenuated IL-10 and IL-6 concentrations upon stimulation with the ΔfliC mutant ( Fig 2 ) . These results suggested that , besides its established role in modulation of flagellin sensing , the TLR5:c . 1174C>T variant may alter innate immune responses other than those induced by flagellin . We previously demonstrated the importance of B . pseudomallei LPS in the host response to the bacteria , implicating the LPS-TLR4 axis in melioidosis [7] . To test whether carriage of the TLR5:c . 1174C>T variant was associated with differential responses to TLR4-dependent stimuli , we measured cytokine and chemokine levels following whole blood stimulation with three different B . pseudomallei LPS preparations in a large cohort of healthy subjects from Thailand . As control ligands , we also assayed cytokine responses induced by E . coli LPS , by flagellin and by Pam3CSK4 , a TLR1 agonist . To limit any confounding by other TLR pathway variants , we restricted the analysis to individuals without genetic variation at TLR4:c . 896A>G and TIRAP:c . 558C>T . Both of these variants have been associated with susceptibility to infection [21–26] . IL-8 , TNF-α , IL-10 , and G-CSF concentrations in healthy subjects by TLR5:c . 1174C>T genotype are shown graphically in Fig 3 and the results of linear regression analysis of the association of genotype and cytokine concentration are shown in S1 Table . In unstimulated blood from individuals carrying the TLR5 variant , we observed no difference in released cytokine concentrations . As we have previously reported , in flagellin-stimulated blood from individuals with the variant , all four cytokine responses were uniformly lower [5] . In response to all three B . pseudomallei LPS , carriers of the variant generated modestly lower IL-8 levels ( p<0 . 05 for all , and all were significant assuming a false discovery rate of 0 . 10 for all tests performed ) . In addition , the magnitude of these effects , as assessed by the beta coefficient , was consistently less in LPS-stimulated blood than that in flagellin-stimulated blood . Pam3CSK4 also induced lower levels of IL-10 in carriers of the variant . Yet the consistent reduction in IL-8 concentrations induced by B . pseudomallei LPS stimulation of blood of carriers of the TLR5:c . 1174C>T variant provided supporting evidence that the variant may modulate elements of TLR4-dependent signaling . We next queried whether the TLR5:c . 1174C>T-dependent differences in B . pseudomallei LPS-induced cytokine production observed may be secondary to differences in TLR4 mRNA expression . We isolated peripheral blood monocytes from individuals with and without the variant and quantified TLR4 expression . We found that there was no variant-dependent difference in either baseline TLR4 expression or in TLR4 expression after stimulation with LPS for four hours ( Fig 4 ) . Therefore , it appeared that any effect of the variant on TLR4-dependent signaling is likely to occur down-stream of TLR4 transcription and mRNA processing . We then used a gene silencing approach to analyze the effect of TLR5 on LPS-induced cytokine responses . Using siRNA , we knocked down TLR5 expression in human monocytic THP1-Dual cells . We first confirmed that TLR5 gene expression was reduced but that TLR4 gene expression was unchanged ( Fig 5A ) . After stimulating these cells with LPS or with flagellin ( as a positive control ligand ) we measured cytokines in the cell supernatants . We found that TLR5 silencing impaired both TNF-α and IL-8 release ( Fig 5B & 5C ) in response to flagellin as well as in response to LPS . This observation further implicated TLR5 as a regulator of TLR4-driven responses to LPS . Transcription factors NF-κB and IRF3 mediate LPS-TLR4 signaling . We therefore quantified activation of these transcription factors in the same set of experiments using THP1-Dual cells that express SEAP-NF-κB and IRF-Lucia luciferase reporters . However , we did not detect significant differences in NF-κB or IRF activation after TLR5 silencing following LPS stimulation in this system ( Fig 5D & 5E ) . In additional experiments , we assessed NF-κB activation after transiently transfecting TLR5 into HEK293 cells stably transfected with TLR4/MD-2/CD14 . We first performed a series of pilot experiments to confirm the responsiveness of these cells to LPS and to flagellin based upon an NF-κB luciferase reporter assay ( Fig 6A & 6B ) . We next transfected these cells with increasing quantities of TLR5 , stimulated cells with 10 ng/mL LPS , and assessed NF-κB activation . The total amount of DNA transfected in each well was kept constant by adding empty vector as necessary . In this gain-of-function system , we found that addition of 1 ng of TLR5 tended to enhance TLR4-dependent NF-κB activity but at higher doses of TLR5 , this effect diminished ( Fig 6C ) . Together , these data do not clearly implicate NF-κB or IRF3 pathways in TLR5 modulation of TLR4-dependent signaling , suggesting that the effect observed on cytokine production may be via other pathways .
In this work we test the hypothesis that TLR5 modulates the flagellin-independent and TLR4-dependent immune response to B . pseudomallei . We show that selected cytokine responses from individuals carrying the TLR5:c . 1174C>T nonsense polymorphism are modestly but significantly impaired upon stimulation of blood with B . pseudomallei lacking flagellin or with B . pseudomallei LPS . This effect does not appear to be related to altered TLR4 gene expression in peripheral blood monocytes from carriers of the variant . In monocytic THP1 cells TLR5 silencing significantly reduces both TNF-α and IL-8 production in response to LPS without an impact on TLR4 expression or NF-κB or IRF activation . In transfected HEK cells , we did not observe a clear effect of co-transfection of TLR4 and TLR5 on NF-κB activation . Together , these data provide evidence of a modulating effect of TLR5 on the LPS-TLR4 axis but do not identify the underlying mechanism of this effect . Analysis of human genetic variation offers insights into the relevance of specific proteins in the clinical arena . Our observation of robustly enhanced survival in carriers of the nonsense polymorphism TLR5:c . 1174C>T infected with B . pseudomallei [5 , 6] prompted us to understand the mechanism underlying this association . As might be expected , TLR5:c . 1174C>T has a dramatic effect on flagellin-induced cytokine production , as demonstrated in previous work by Hawn et al . and by our group [3 , 5] . More surprisingly , we now report a flagellin-independent effect of TLR5:c . 1174C>T on whole blood cytokine production induced by B . pseudomallei ΔfliC or B . pseudomallei LPS ex vivo . While this LPS-dependent effect for the TLR5 variant is not as dramatic as the effect observed upon stimulation with flagellin ( based on degree of impairment of cytokine production and range of cytokines impacted ) , it is nonetheless detectable and significant . Although we have previously reported an association between TLR5:c . 1174C>T and survival in melioidosis , we have also established the primary importance of LPS in driving innate immune responses to killed B . pseudomallei [7] . In these studies , the LPS-TLR4 axis was responsible for ~80–90% of the B . pseudomallei-induced TNF-α release by human monocytes . Our current experiments demonstrating modulation of the LPS-TLR4 axis by TLR5:c . 1174C>T indicate that the observed association of TLR5:c . 1174C>T with survival in melioidosis may not be mediated by flagellin-sensing but may instead be mediated by LPS-sensing . As genetic association studies do not prove causation , we used molecular techniques to address the more general question of whether TLR5 alters TLR4-dependent signaling . We clearly show an effect of TLR5 on LPS-induced/TLR4-dependent TNF-α and IL-8 production after inhibiting TLR5 expression with siRNA in a monocytic cell line . As we have observed that TLR5 neither detects LPS nor modulates expression of TLR4 , CD14 , or MD-2 in this system , we speculate that there is an interaction between TLR4 and TLR5 that augments TLR4-dependent signaling . TLR5/TLR4 heteromerization has been demonstrated in transfected COS-1 cells by Mizel et al [8]; consequently , direct heterodimerization of these TLRs may be responsible for the effect we are observing . Alternatively , TLR5 may instead interact with one of the many accessory proteins in the TLR4 pathway to enhance signaling . Further , it is conceivable that the truncation of TLR5 protein in carriers of TLR5:c . 1174C>T may alter interactions with other innate immune molecules besides TLR4 . Whether this truncated protein is membrane-bound or becomes soluble is not clear . To clarify these potential mechanisms , investigation of the protein-protein interactions between human TLR4 and TLR5 ( wildtype and variant proteins ) is necessary in future studies . Additionally , we could not attribute the observed effects on cytokine production to differential activation of either NF-κB or IRF3 , transcription factors that mediate LPS-TLR4 signaling via MyD88-dependent and TRIF-dependent pathways , respectively [27] . Further investigations of other elements of MyD88-dependent pathways such as MAP kinases and transcription factor AP-1 are necessary to better elucidate this mechanism . Additionally , our data indicate that there may be differences in the TLR5:c . 1174C>T-dependent cytokine response between LPS from B . pseudomallei and LPS from E . coli , perhaps implicating distinct and as yet undefined sensing mechanisms and pathways . While our investigation focused on the interaction of the LPS-TLR4 and flagellin-TLR5 axes in evaluating possible mechanisms attributable to the TLR5:c . 1174C>T variant in melioidosis , several other mechanisms should be considered . Since we began our studies , other investigators have reported on a role for both human TLR2 and TLR4 in recognition of B . pseudomallei LPS [28] . This observation raises the possibility of a TLR2-mediated effect on our experiments , an issue that we did not investigate . We also note a modest but significant association between TLR5:c . 1174C>T genotype and IL-10 induced by TLR2/1 agonist PAM3CSK4 in our whole blood stimulation study . This may simply be due to chance but alternatively may reflect an interaction between TLR2/1 and TLR4 . Together , these observations provide corroborating evidence for a complex interplay between TLRs in recognition of bacterial ligands . It is also possible that the phenotypes associated with TLR5:c . 1174C>T in humans may reflect environmental differences relating to alterations in tonic signaling or microbiota development [29] . These possibilities deserve evaluation in future studies . Our finding that TLR5 modulates flagellin-independent , TLR4-dependent cytokine release in whole blood and in blood monocytic cells suggests a possible broadened biological significance of the TLR5:c . 1174C>T polymorphism , with potential implications in melioidosis and in other infections and inflammatory disorders . We acknowledge that the clinical magnitude of the effect may be small . Yet , the TLR5:c . 1174C>T polymorphism resulting in a non-functional TLR5 protein is fairly common ( 5% global minor allele frequency ) . Besides associations with outcomes from melioidosis , outcomes in several other diseases have been associated with this variant . For instance , in carriers of the variant , susceptibility to legionellosis ( infection with the Gram-negative flagellated bacterium Legionella pneumophila ) is increased [3] , the risk for Crohn’s disease and systemic lupus erythematosus is decreased , and the body mass index in cystic fibrosis is decreased [30–32] . Such associations underscore the importance of developing a better understanding of the role of TLR5 in innate immune pathways as a potential means of identifying novel therapeutics .
|
Toll-like receptors ( TLRs ) are important receptors of the innate immune system . TLR4 detects bacterial lipopolysaccharide ( LPS ) and TLR5 detects bacterial flagellin . A common human polymorphism in TLR5 encodes a shortened protein and blunts the immune response to flagellin . Individuals carrying this variant have decreased mortality from melioidosis , infection caused by the flagellated Gram-negative bacterium Burkholderia pseudomallei . The mechanism of protection is not known . We tested the hypothesis that the observed effect of the polymorphism is independent of flagellin and involves LPS-TLR4 pathways . We found that blood from individuals carrying the polymorphism produced lower levels of cytokines IL-6 and IL-10 in response to an aflagellated B . pseudomallei mutant and less IL-8 in response to purified B . pseudomallei LPS than blood from individuals without the variant . We further observed that in THP1 cells stimulated with LPS , silencing of TLR5 with siRNA reduced levels of both TNF-α and IL-8 . These effects were not explained by differences in TLR4 mRNA expression . We conclude that the effects of the TLR5 polymorphism on the host inflammatory response to B . pseudomallei may not be restricted to flagellin-driven pathways . These results provide insights into the role of TLR5 in the innate immune response in melioidosis and other conditions .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"bacteriology",
"blood",
"cells",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"luciferase",
"body",
"fluids",
"immune",
"physiology",
"cytokines",
"immune",
"cells",
"enzymes",
"melioidosis",
"immunology",
"microbiology",
"enzymology",
"immune",
"receptor",
"signaling",
"bacterial",
"diseases",
"developmental",
"biology",
"membrane",
"receptor",
"signaling",
"molecular",
"development",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"white",
"blood",
"cells",
"microbial",
"physiology",
"animal",
"cells",
"proteins",
"oxidoreductases",
"immune",
"system",
"toll-like",
"receptors",
"biochemistry",
"signal",
"transduction",
"bacterial",
"physiology",
"blood",
"cell",
"biology",
"anatomy",
"monocytes",
"physiology",
"flagellin",
"biology",
"and",
"life",
"sciences",
"immune",
"receptors",
"cellular",
"types",
"cell",
"signaling"
] |
2019
|
Flagellin-independent effects of a Toll-like receptor 5 polymorphism in the inflammatory response to Burkholderia pseudomallei
|
In this study , we developed a mouse model of type 2 diabetes mellitus ( T2DM ) using streptozotocin and nicotinamide and identified factors that increase susceptibility of T2DM mice to infection by Mycobacterium tuberculosis ( Mtb ) . All Mtb-infected T2DM mice and 40% of uninfected T2DM mice died within 10 months , whereas all control mice survived . In Mtb-infected mice , T2DM increased the bacterial burden and pro- and anti-inflammatory cytokine and chemokine production in the lungs relative to those in uninfected T2DM mice and infected control mice . Levels of IL-6 also increased . Anti-IL-6 monoclonal antibody treatment of Mtb-infected acute- and chronic-T2DM mice increased survival ( to 100% ) and reduced pro- and anti-inflammatory cytokine expression . CD11c+ cells were the major source of IL-6 in Mtb-infected T2DM mice . Pulmonary natural killer ( NK ) cells in Mtb-infected T2DM mice further increased IL-6 production by autologous CD11c+ cells through their activating receptors . Anti-NK1 . 1 antibody treatment of Mtb-infected acute-T2DM mice increased survival and reduced pro- and anti-inflammatory cytokine expression . Furthermore , IL-6 increased inflammatory cytokine production by T lymphocytes in pulmonary tuberculosis patients with T2DM . Overall , the results suggest that NK-CD11c+ cell interactions increase IL-6 production , which in turn drives the pathological immune response and mortality associated with Mtb infection in diabetic mice .
Mycobacterium tuberculosis ( Mtb ) infects one-third of the world’s population and causes almost 1 . 3 million deaths per year [1] . Approximately 90% of those infected have a latent tuberculosis infection and develop protective immunity to contain it; however , but 10% progressive to active tuberculosis ( TB ) disease months or years after infection [2] . The risk for progression to TB disease is increased by acquired factors including human immunodeficiency virus ( HIV ) infection , alcoholism , smoking , and diabetes [3] . Developing nations are epicenters of diabetes [4] . Diabetes mellitus alters innate and adaptive immune responses and increases the risk of developing active TB [5] . In type 2 diabetes mellitus ( T2DM ) patients , there is a reduced association between mycobacteria and monocytes; therefore , phagocytosis via complement receptors is compromised [6 , 7] . Mtb-infected diabetic mice show delayed priming of the adaptive immune response , which is necessary to restrict Mtb replication [8] . Hyperactive T-cell responses and increased Th1 and Th17 cytokine production are noted in TB patients with type 2 diabetes [9] . Limited information is available about experimental models used to study the effects of T2DM during Mtb infection . Spontaneous T2DM rodent models , such as GK/Jcl rats , have a higher bacterial load and increased immune pathology than non-diabetic Wistar rats after infection with a Mtb Kurono aerosol [10] . Furthermore , T2DM guinea pigs are highly susceptible to Mtb infection; even non-diabetic hyperglycemia exacerbates disease severity [11 , 12] . However , a detailed understanding of the protective immune responses in type 2 diabetic hosts during Mtb infection is essential if we are to develop an adequate prophylactic or therapeutic agent . In the current study , we employed an experimentally induced T2DM model in wild type C57BL/6 mice and investigated the immune response to Mtb infection . We found that natural killer ( NK ) and CD11c+ cell interactions in Mtb-infected T2DM mice led to increased IL-6 production , which drives the pathological immune response and increases mortality . We also found that IL-6 increases inflammatory cytokine production in pulmonary tuberculosis patients with T2DM .
A combination of streptozotocin ( STZ ) and nicotinamide ( NA ) induces T2DM in mice [8] . STZ ( 180 mg/kg of body weight ) and NA ( 60 mg/kg of body weight ) were administered intraperitoneally to C57BL/6 mice three times , with an interval of 10 days between doses . A schematic representation of T2DM induction is shown in Fig 1A . After 1 month , mice developed T2DM , as assessed by measurement of blood glucose and serum insulin levels . Blood glucose levels measured at monthly intervals for up to 8 months in STZ/NA-treated mice were consistently ≥250 mg/dl ( Fig 1B ) . Blood glucose levels in control mice were 60–125 mg/dl . To determine whether STZ/NA-treated mice developed insulin resistance , we next measured serum insulin levels . One and three months after T2DM induction , serum insulin levels in STZ/NA-treated mice were significantly higher than those in control mice ( Fig 1C ) . Six months after T2DM induction , serum insulin levels in STZ/NA-treated mice were 4-fold higher than those in control mice ( Fig 1C ) . Insulin resistance , a characteristic feature of T2DM , was confirmed by oral glucose tolerance test ( OGTT ) 6 months after STZ/NA injection ( Fig 1D ) . Furthermore , serum levels of cholesterol , triglyceride , and free fatty acids were elevated by 6 months after STZ/NA treatment ( Fig 1E–1G ) . Dyslipidemia is another characteristic of T2DM in humans and , combined with demonstrated insulin resistance , confirms the validity of our mouse T2DM model . We next investigated TB defense in T2DM mice by aerosol challenge with Mtb as shown in Fig 2A . One and three months post-infection ( p . i . ) , the lung bacterial burden was similar in T2DM and control mice ( Fig 2B ) . However , by 6 months p . i . , lung bacterial burden was significantly greater in T2DM mice compared to controls ( Fig 2B ) . A similar increase in the bacterial burden was observed in the spleens and livers of T2DM mice when compared with those of control mice ( data presented in Dryad Data Repository; doi:10 . 5061/dryad . qn42t ) . Alveolar macrophages are the first immune cells that Mtb encounters in the lung [13] . To determine whether the increased bacterial growth described above was due to altered antimicrobial function of these cells , we isolated alveolar macrophages from control and T2DM mice ( one , three and six months after T2DM induction ) and infected them with Mtb . The CFU were quantified after 5 days . Mtb growth was similar in the alveolar macrophages of control and T2DM mice after one and three months post induction of T2DM . However , control of Mtb growth was impaired in alveolar macrophages , six months after the induction of T2DM ( Fig 2C ) . We next determined the survival of uninfected control and T2DM mice and of Mtb-infected control and T2DM mice . By 10 months p . i . all Mtb-infected T2DM mice died , whereas only 40% of the uninfected T2DM mice and 6 . 6% of the Mtb-infected non-diabetic mice died ( Fig 2D ) . In contrast , all control mice survived . We next determined whether T2DM has any effect on pro- and anti-inflammatory responses following Mtb infection . Control and T2DM mice were infected with Mtb , and after 1 and 6 months the levels of various cytokines and chemokines were measured in lung homogenates by multiplex ( 23-plex ) ELISA . As shown in Fig 3A , there was a significant increase in both pro- and anti-inflammatory cytokines and chemokines in the Mtb-infected T2DM mice at 1 and 6 months p . i . , when compared with either uninfected T2DM mice or Mtb-infected control mice . However , the cytokine and chemokine levels in the lungs at 6 months p . i . were significantly higher than those at 1 month p . i . The levels of inflammatory cytokines ( IL-6 , IFN-γ , TNF-α , IL-1β ) and chemokines ( MCP-1 ) in whole-lung homogenates from T2DM mice were significantly higher than those in homogenates from Mtb-infected control mice or uninfected T2DM mice ( Fig 3A ) . In addition , we found that interleukin ( IL ) -1α , -5 , -9 , -12 [p40] , -12 [p70] , and -13 , and G-CSF , GM-CSF , KC , and MIP-1β , in the lung homogenates of Mtb-infected T2DM mice were significantly higher than those in Mtb-infected control mice and uninfected T2DM mice at 6 months p . i . ( data presented in the Dryad Data Repository; doi:10 . 5061/dryad . qn42t ) . We also examined expression of various pro- and anti-inflammatory cytokines in whole-lung tissue using real-time PCR . Similar to the ELISA data , we observed increased expression of TNF-α , IFN-γ , IL-6 , IL-1β , IL-21 , IL-23 , TGF-β , and IL-10 in Mtb-infected T2DM mice ( data presented in the Dryad Data Repository; doi:10 . 5061/dryad . qn42t ) . Histological analysis revealed significantly more inflammation throughout the lungs of Mtb-infected T2DM mice when compared with those of Mtb-infected control mice or uninfected T2DM mice ( Fig 3B ) . IL-6 is a pleotropic cytokine that regulates both pro- and anti-inflammatory cytokine production [14] , and it has both protective and pathogenic roles in diabetes [14] . We found that both pro- and anti-inflammatory cytokine production is dysregulated in Mtb-infected T2DM mice compared to control T2DM mice and Mtb-infected control mice . There are conflicting reports about the role of IL-6 in Mtb infection [15 , 16] . IL-6-deficient mice are susceptible to Mtb infection [15] , and IL-6 participates in the induction of type 1 protective T-cell responses after vaccination [17] . However , IL-6 is not required to generate specific immune responses to Mtb infection [18] . Thus , we next determined whether neutralizing IL-6 affects survival , cytokine production , or the bacterial burden in T2DM mice . Fig 4A shows a schematic representation of Mtb infection and anti-IL-6 mAb treatment in T2DM mice . One month after T2DM induction ( acute diabetes ) , mice were intranasally infected with 50–100 CFU of Mtb . At 6 months p . i . , the mice were treated with a neutralizing anti-IL-6 mAb , an isotype-matched control mAb , or PBS . As shown in Fig 4B , 65% ( p = 0 . 05 ) of Mtb-infected T2DM ( acutely diabetic ) mice that received the isotype-matched control mAb or PBS died within 2 months . By contrast , all mice that received the anti-IL-6 mAb survived . Anti-IL-6 mAb treatment also reduced the bacterial burden in the lungs ( Fig 4C ) , spleen ( 1 . 5 ± 0 . 8 × 104 vs . 10 . 5 ± 0 . 8 × 104 CFU; p = 0 . 0003 ) , and liver ( 1 . 75 ± 0 . 25 × 103 vs . 2 . 7 ± 0 . 25 103 CFU; p = 0 . 03 ) . Real-time PCR analysis of lung samples indicated that anti-IL-6 mAb treatment inhibited expression of IL-17 , TNF-α , IL-10 , and TGF-β ( Fig 4D ) when compared with that in mice treated with the isotype-matched control mAb or PBS . Histological examination of lung tissue indicated a similar degree of inflammation in PBS-treated and isotype-matched control antibody-treated mice with acute T2DM and infected with Mtb ( Fig 4E ) . By contrast , anti-IL-6 mAb treatment significantly reduced inflammation in the lungs of Mtb-infected acute T2DM mice ( Fig 4E ) . We next determined whether neutralizing IL-6 affected survival , cytokine production , or bacterial burden in mice with chronic T2DM ( mice were infected 6 months after T2DM induction ) . On the day of infection , mice received the anti-IL-6 mAb , the isotype-matched control mAb , or PBS ( Fig 5A ) . As shown in Fig 5B , 80% ( p = 0 . 05 ) of Mtb-infected T2DM mice ( chronically diabetic ) that received the isotype-matched control mAb or PBS died within 2 months . By contrast , all Mtb-infected chronic T2DM mice that received the anti-IL-6 mAb survived . Anti-IL-6 mAb treatment also reduced the bacterial burden in the lungs ( Fig 5C ) , spleen ( 1 . 5 ± 0 . 3 × 104 vs . 4 . 4 ± 0 . 8 × 104 CFU; p = 0 . 01 ) , and liver ( 0 . 6 ± 0 . 6 × 103 vs . 7 . 8 ± 1 . 4 × 103 CFU; p = 0 . 001 ) . Real-time PCR analysis of lung samples indicated that anti-IL-6 mAb treatment of chronically diabetic Mtb-infected mice was associated with reduced expression of IFN-γ , IL-17 , TNF-α , IL-10 , and TGF-β ( Fig 5D ) when compared with that in mice treated with the isotype-matched control mAb . Histological examination of the lungs suggested a similar degree of inflammation in PBS-treated and isotype-matched control antibody-treated mice with chronic T2DM harboring Mtb ( Fig 5E ) . By contrast , anti-IL-6 mAb treatment significantly reduced inflammation in the lungs of Mtb-infected mice with chronic T2DM ( Fig 5E ) . To confirm our finding that IL-6 levels in the lungs of Mtb-infected T2DM mice increased at 6 months p . i . , mice were euthanized and lung sections were examined for IL-6 expression by immunohistochemistry ( IHC ) . As shown in Fig 6A and 6B , the mean histology score ( H-score ) for IL-6 in Mtb-infected diabetic mice was significantly higher than that in Mtb-infected control mice and uninfected diabetic mice . To determine the cellular source of IL-6 in Mtb-infected T2DM mice , we first examined the leukocyte populations by flow cytometry . As shown in Table 1 , the number of CD11c+MHCII+CD103+ , CD11c+CD11b+MHCII+ , F4/80+CD64+MHCII+ , Ly6G+ neutrophils , and lymphocytes in the lungs of Mtb-infected T2DM mice at 1 month post-Mtb infection was significantly higher than that in Mtb-infected non-diabetic mice or uninfected T2DM . A similar increase was observed at 6 months p . i . We next examined the phenotype of IL-6 producing pulmonary cells at 1 and 6 months p . i . There were no significant differences in the absolute numbers of IL-6-producing Ly6G+ neutrophils , B220+IgM+ B cells , CD3+NK1 . 1-T cells , or CD3-NK1 . 1+ NK cells ( Fig 6C ) . However , the absolute number of IL-6-producing CD11c+MHCII+CD103+ and CD11c+CD11b+MHCII+ cells in the lungs of T2DM mice at 1 month after Mtb infection was significantly higher than that in the lungs of uninfected T2DM mice ( Fig 6C ) or Mtb-infected control mice ( Fig 6C ) . As shown in Fig 6C , a similar increase in IL-6-producing CD11c+ cells were noted in the lungs of T2DM mice at 6 months post-Mtb infection . Although there was an increased frequency of F4/80+CD64+MHCII+IL-6+ cells in the lungs of Mtb-infected T2DM mice compared with those of uninfected T2DM mice at 6 months p . i . ( Fig 6C ) , there was no significant difference between Mtb-infected non-diabetic control mice ( Fig 6C ) . Overall , these results suggest that CD11c+ cells are the major source of IL-6 in Mtb-infected T2DM mice . To further confirm the cellular source of IL-6 at 6 months p . i . , we isolated various cell populations from the pooled spleen , lymph node , and lung cell populations of Mtb-infected control and T2DM mice by magnetic sorting and measured IL-6 expression by real-time PCR . We found that CD11c+ cells are the major source of IL-6 ( data shown in the Dryad Data Repository; doi:10 . 5061/dryad . qn42t ) . The interaction between NK cells and macrophages is crucial for the initiation and amplification of early immune responses [19] . We found a significant increase in NK and CD11c+ cell numbers in the lungs of Mtb-infected T2DM mice ( Table 1 ) . To determine whether NK cells are involved in increased IL-6 production by CD11c+ cells , we first examined lung sections from Mtb-infected T2DM mice by confocal microscopy . Imaging results at 6 months p . i . indicated that more NK cells in Mtb-infected T2DM mice were in close proximity to IL-6-producing CD11c+ cells than in Mtb-infected control mice ( Fig 7A and S1 Fig ) . More importantly , the result indicates that the marked increase in IL-6 production occurs in the region in which both NK cells and CD11c+ cells interact . We further determined whether the NK and CD11c+ cell interaction increases IL-6 production by lung mononuclear cells in Mtb-infected T2DM mice . Six months after Mtb infection , mononuclear cells were isolated from the lungs of T2DM and non-diabetic control mice and some cell populations were depleted of NK cells by magnetic separation . Lung mononuclear cells and NK cell-depleted lung mononuclear cells were cultured with γ-irradiated Mtb H37Rv ( γ-Mtb ) . After 48 h , IL-6 levels in the culture supernatants were measured by ELISA and the phenotype of IL-6-producing cells was identified by flow cytometry . Stimulation with γ-Mtb significantly enhanced IL-6 production by pulmonary mononuclear cells from Mtb-infected T2DM mice ( Fig 7B ) when compared with those from Mtb-infected control mice ( Fig 7B ) . However , depletion of NK cells from Mtb-infected T2DM pulmonary mononuclear cells led to a significant reduction in IL-6 levels ( Fig 7B ) . Furthermore , we found that the frequency of IL-6+CD11c+MHCII+ and IL-6+CD11b+MHCII+ cells ( Fig 7B ) increased significantly after culture of Mtb-infected T2DM pulmonary mononuclear cells with γ-Mtb . However , depletion of NK cells from Mtb-infected T2DM pulmonary mononuclear cells resulted in a significant reduction in the frequency of IL-6+CD11c+MHCII+ cells ( Fig 7B ) . By contrast , depletion of NK cells had no effect on the frequency of IL-6+CD11b+MHCII+ cells ( Fig 7B ) . These results further confirm that CD11c+ cells are a major source of IL-6 and that NK cells from Mtb-infected T2DM mice increase IL-6 production by CD11c+MHCII+ cells . NK cell-activating receptors play an important role in the development of diabetes [20] . We next examined expression of NK cell-activating receptors in Mtb-infected mice by flow cytometry . Lung CD3-NKp46+ NK cells from Mtb-infected T2DM mice expressed higher levels of NKG2D ( Fig 7C ) and DNAM-1 ( Fig 7C ) than those from Mtb-infected control mice . We then examined the possible role of these activating receptors in stimulating CD11c+ cells to produce IL-6 . At 6 months p . i . , lung mononuclear cells from Mtb-infected T2DM mice were cultured with γ-Mtb in the presence of blocking NKG2D or DNAM-1 mAbs or isotype-matched control antibodies . The frequency of IL-6-expressing CD11c+MHCII+ cells ( Fig 7D ) increased significantly after culture of Mtb-infected T2DM pulmonary mononuclear cells with γ-Mtb in the presence or absence of the isotype-matched control antibodies . Blocking the NKG2D ( Fig 7D ) or DNAM-1 ( Fig 7D ) interaction with CD11c+ cells led to a significant reduction in the frequency of IL-6+CD11c+ cells . Similarly , IL-6 levels in the culture supernatants of cells cultured with blocking NKG2D ( Fig 7D ) or DNAM-1 mAbs ( Fig 7D ) decreased significantly . To further confirm the above findings , NK cells and CD11c+ cells were isolated from pooled splenic , lymph node , and lung cells from Mtb-infected control and T2DM mice by magnetic selection . Autologous NK cells and CD11c+ cells were cultured together at a ratio of 1:4 ( 1 NK and 4 CD11c+ ) in the presence or absence of γ-Mtb and with or without the isotype control or NKG2D or DNAM-1 blocking antibodies . After 48 h , the culture supernatants were collected and IL-6 levels were measured by ELISA . Culture of Mtb-infected control or Mtb-infected T2DM mouse NK cells with autologous CD11c+ cells in the absence of γ-Mtb did not induce IL-6 production . Culture of Mtb-infected control mouse NK cells with autologous CD11c+ cells in the presence of γ-Mtb resulted in 251 . 5 ± 65 . 1 pg/ml IL-6; this increased to 556 . 9 ± 52 . 5 pg/ml ( p = 0 . 02 ) in NK cells and CD11c+ cells from Mtb-infected T2DM mice ( Fig 7E ) . This increase in IL-6 production by CD11c+ cells was inhibited by anti-DNAM-1 and anti-NKG2D blocking antibodies ( Fig 7E ) . The above result indicated that the interaction between NK and CD11c+ cells increases IL-6 production in Mtb-infected T2DM mice . Therefore , we next determined whether depletion of NK cells with a NK1 . 1 antibody affected survival , cytokine production , or the bacterial burden in acute T2DM mice . One month after STZ/NA treatment , acutely diabetic mice were infected with50–100 CFU Mtb . At 6 months p . i . , mice were treated with an anti-NK1 . 1 mAb , an isotype-matched control mAb , or PBS ( Fig 8A ) . As shown in Fig 8B , 65% ( p<0 . 05 ) of Mtb-infected T2DM mice that received the isotype-matched control mAb or PBS died within 2 months . By contrast , all Mtb-infected T2DM mice that received the anti-NK1 . 1 mAb survived . Anti-NK1 . 1 mAb treatment also reduced the bacterial burden in the lungs ( Fig 8C ) , spleen ( 1 . 5 ± 0 . 86 × 104 vs . 10 . 5 ± 0 . 86 × 104 CFU; p = 0 . 0003 ) , and liver ( 1 . 75 ± 0 . 25×103 vs . 2 . 75 ± 0 . 25 × 103 CFU; p = 0 . 03 ) by a marginal , but statistically significant , amount . Real-time PCR analysis of lung samples indicated that anti-NK1 . 1 mAb treatment of acutely diabetic Mtb-infected mice was associated with significantly lower levels of IL-6 , TNF-α , IL-10 , and TGF-β ( Fig 8D ) expression than those observed in mice treated with the isotype-matched control mAb . Histological examination of lung tissue indicated a similar degree of inflammation in PBS-treated and isotype-matched control antibody-treated mice with acute T2DM mice and Mtb ( Fig 8E ) . By contrast , anti-NK1 . 1 mAb treatment significantly reduced inflammation in the lungs of Mtb-infected mice with acute T2DM ( Fig 8E ) . To determine the relevance of the above findings with respect to human Mtb infection , we obtained blood from pulmonary tuberculosis patients with or without T2DM and first determined the frequency of pro-inflammatory cytokine-producing T cells by flow cytometry . The frequency of IFN-γ- ( 1 . 5-fold , p = 0 . 026 , Fig 9A ) and IL-2- ( 2 . 1-fold , p = 0 . 002 , Fig 9A ) producing cells was significantly higher in the blood of diabetic than in that of non-diabetic pulmonary tuberculosis patients . By contrast , there were no significant differences in the frequency of TNF-α- and IL-17-producing cells between the two groups . We also cultured whole blood in the presence of 10 μg/ml purified protein derivative ( PPD ) . After 18 h , the frequency of IFN-γ- , IL-2- , TNF-α- , and IL-17-producing cells was determined by flow cytometry . As shown in Fig 9A and 9B , PPD significantly induced expression of IFN-γ , IL-2 , TNF-α , and IL-17A . The frequency of IFN-γ- ( 1 . 7-fold , p = 0 . 0002; Fig 9A ) , IL-2- ( 2 . 1-fold , p = 0 . 0001; Fig 9A and 9B ) , TNF-α- ( 1 . 6-fold , p = 0 . 0005; Fig 9A ) , and IL-17A- ( 2-fold , p = 0 . 0004; Fig 9A ) producing cells was significantly higher in diabetic than in non-diabetic pulmonary TB patients . We also examined whether neutralizing the IL-6 receptor affected PPD-induced changes in the frequency of IFN-γ- , IL-2- , TNF-α- , and IL-17A-producing cells in pulmonary TB patients with T2DM . As shown in Fig 9B , the anti-IL-6 antibody significantly reduced the frequency of IFN-γ- ( 1 . 7-fold , p = 0 . 0005 ) , IL-2- ( 2 . 2-fold , p = 0 . 009 ) , TNF-α- ( 6 . 6-fold , p = 0 . 0005 ) , and IL-17A- ( 3 . 3-fold , p = 0 . 0005 ) producing cells when compared with the isotype-matched control antibody . However , the anti-IL-6 antibody had no effect on the frequency of cytokine-producing cells in healthy volunteers ( Fig 9C ) .
In this study , we investigated the immune response of mice to Mtb infection following the induction of T2DM . Diabetic mice were found to have increased lung bacterial burden and mortality compared to non-diabetic controls . Alveolar macrophages from T2DM mice were more permissive to Mtb growth ex vivo compared to non-diabetic controls , indicating an impairment of innate antimicrobial function . Multiplex cytokine and chemokine data and real-time PCR analysis of Mtb-infected T2DM lungs also demonstrated significantly higher expression of genes encoding pro- and anti-inflammatory cytokines than in lungs from uninfected T2DM and infected non-diabetic control mice . Neutralization of IL-6 increased the survival of all Mtb-infected T2DM mice , reduced the bacterial burden , and reduced cytokine production . We found that CD11c+ cells were the major source of IL-6 in Mtb-infected T2DM mice . IL-6 production by CD11c+ cells was further enhanced by NK cells . We also found that IL-6 enhances inflammatory cytokine production in pulmonary tuberculosis patients with T2DM . Limited information is available about protective immune responses in type 2 diabetic hosts during Mtb infection . Our results suggest that the NK-CD11c+ cell interaction increases IL-6 production , which drives the pathological immune response and reduces survival of Mtb-infected T2DM mice . Chemically induced type 1 diabetes ( T1DM ) models are widely used in research , and mice with STZ-induced insulin deficiency develop susceptibility to TB [8] . Approximately 90% of people living with diabetes have T2DM , making it by far the most prevalent form of diabetes in TB patients [21] . It is therefore important to employ T2DM models for mechanistic studies of this dual burden . Different approaches have been used to model T2DM in animals , including the combination of high-fat diet with low to intermediate doses of STZ [22] . Some approaches require relatively long periods of time to exhibit all of the major features of the disease and some fail to replicate persistent hyperglycemia [23] . This is an important limitation since the vascular and renal complications of diabetes only develop after prolonged hyperglycemia , and related mechanisms may drive at least some features of diabetic immunopathy [24] . In the current study , we induced T2DM in mice using STZ and NA , which resulted in sustained hyperglycemia for up to 8 months . After 6 months , T2DM mice showed significantly elevated blood cholesterol and triglyceride levels . These findings suggest that our model can be used to investigate TB defenses in the setting of acute or chronic T2DM . Experimental Mtb infection has been investigated in other animal models of T2DM . Sugawara et al . [10] reported that GK/Jcl rats , which spontaneously develop T2DM , have a higher bacterial load and more severe immune pathology than non-diabetic Wistar rats at 5–12 weeks after infection with Mtb Kurono . Expression of mRNA encoding several cytokines , including IFN-γ , TNF-β , and IL-1β , was higher in Wistar rats at 1 and 3 weeks p . i . , but higher in GK/Jcl rats by 12 weeks p . i . Podell et al . [11] reported increased TB susceptibility in guinea pigs with T2DM induced by low dose STZ plus a high-fat , high-sucrose diet . The guinea pig TB phenotype was characterized by an increased bacterial burden , more severe immune pathology , increased cytokine expression , and increased mortality . The results of TB studies using the rat and guinea pig T2DM models , the mouse T2DM model presented here , and the mouse T1DM model [8] differ in some aspects; however , all show impaired control of Mtb replication , more severe immune pathology , and increased expression of multiple cytokines . TB in diabetic people is associated with increased sputum smear positivity at the time of diagnosis ( a surrogate marker for higher bacterial burden ) , increased radiographic severity of disease , increased mortality ( reflecting increased immune pathology ) , and increased expression of several nominally protective cytokines [8 , 9 , 25 , 26] . These similarities support the relevance of our T2DM mouse model to the interaction between TB and diabetes in people . We found that C57BL/6 mice with STZ/NA-induced T2DM exhibited increased expression of pro- and anti-inflammatory cytokine genes , including IL-6 , after Mtb infection . IL-6 is a pleiotropic cytokine that has both protective and pathogenic roles in diabetes [14] . There are conflicting reports about the role of IL-6 in Mtb infection [15 , 16] . IL-6 contributes to vaccine-induced protective immunity in mice [17] , and IL-6 knockout mice are highly susceptible to Mtb [15] . By contrast , IL-6 produced by macrophages infected with Mtb in vitro selectively inhibits macrophage responses to IFNγ , thereby contributing to the survival of mycobacteria [27] . Correspondingly , IL-6 neutralization increases IFN-γ-mediated killing of intracellular Mtb by inducing autophagy [28] . In human TB patients , IL-6 is implicated in the pathogenesis of the immune reconstitution inflammatory syndrome [29 , 30] We found that in vivo neutralization of IL-6 conferred a survival benefit and was associated with a reduced bacterial burden in the lung and pro-inflammatory cytokine expression in Mtb-infected mice with acute or chronic T2DM; however , it did not alter the hyperglycemic status of the mice . We also found that CD4+ cells from TB patients with T2DM produced significantly elevated levels of Th1 and Th17 cytokines , and that this was inhibited by neutralizing IL-6 . A variety of hematopoietic and non-hematopoietic cells produce IL-6 , and we found that CD11c+ cells were the major source of IL-6 in Mtb-infected T2DM mice . IL-6 production was enhanced by the interaction between NKG2D and DNAM-1 and the corresponding ligands on CD11c+ cells . These results suggest that IL-6 produced during the NK-CD11c+ interaction may be a key factor that drives the damaging immune pathology in diabetic hosts infected by TB . Of note , increased expression of activation markers by unstimulated myeloid cells from diabetic individuals has been documented [31] . NK cells are prominent components of the innate immune system and play a central role in resistance to microbial pathogens . NK cells protect against viruses , parasites , and bacteria , including Mtb , by destroying infected cells and secreting cytokines that shape the adaptive immune response [32–34] . NK cells interact with antigen-presenting cells and T cells , and are involved in one or more stages of immune-mediated attack . Abnormalities in the frequency and activity of NK cells have been described both in animal models and in patients with diabetes . By contrast , depletion of NK cells prevents the development of diabetes [35 , 36] . In TB patients with T2DM , altered CD8+ and NK cell function leads to enhanced pathology [37] . In conclusion , we found that hyperactive NK cells interact with CD11c+ cells to amplify the IL-6-mediated inflammatory immune response in TB . Our data suggest that NK cell-mediated IL-6 production by CD11c+ cells is responsible for driving hyperinflammation and increased mortality in T2DM mice infected with Mtb . The mechanism that underlies NK cell hyperactivation in T2DM mice remains unknown , but NK cell activation was recently reported to be an upstream event in T2DM pathogenesis [38] . The NK-CD11c+ axis and the IL-6 pathway may be promising new targets for host-directed therapies aimed at reducing the severity of immune pathology , which drives morbidity and mortality in those infected by TB .
Specific pathogen-free female wild-type C57BL/6 mice ( 4 to 6 weeks old ) were purchased from Jackson Laboratory and housed at the animal facility at the University of Texas Health Science Center at Tyler . All animal experiments were approved by the Institutional Animal Care and Use Committee of the University of Texas Health Science Center at Tyler . Blood was obtained from 20 healthy controls , 20 pulmonary tuberculosis patients , and 20 pulmonary tuberculosis patients with type T2DM . All subjects were HIV-seronegative with culture-proven pulmonary tuberculosis who had received anti-tuberculosis therapy for < 1 week . Acid-fast stains of sputum samples were positive for all patients . Type 2 diabetes patients had HbA1c levels > 6 . 5% and random blood glucose levels > 200 mg/dl . All human studies were approved by the Institutional Review Board of the National Institute of Research in Tuberculosis ( NCT01154959 ) , Chennai , India , and informed written consent was obtained from all participants . All animal studies were approved by the Institutional Animal Care and Use Committee of the University of Texas Health Science Center at Tyler ( Protocol #533 ) . All animal procedures involving the care and use of mice were undertaken in accordance with the guidelines of the NIH/OLAW ( Office of Laboratory Animal Welfare ) . PE-conjugated anti-IL-6 ( eBioscience ) , FITC-conjugated anti-CD3 ( Tonbo Biosciences ) , PE-conjugated anti-NKp46 ( BioLegend ) , PE-cy7 anti-NKG2D ( eBioscience ) , and APC-anti-DNAM-1 ( eBioscience ) were used for flow cytometry . Antibodies used for the in vivo neutralization experiments were purchased from BioXcell ( mouse anti-IL-6 [MP5-20F3] , anti-NK1 . 1 , and isotype controls [rat IgG1 and mouse IgG2a antibodies] ) . The NKG2D and DNAM-1 blocking antibodies were obtained from eBioscience . STZ and NA were obtained from Sigma Chemicals . Anti-CD11c , anti-IL-6 , anti-NK1 . 1 , secondary antibodies ( goat anti-hamster IgG-Alexa 568 , donkey anti-rat-Alexa 488 , and goat anti-rabbit-Alexa 647 ) , and DAPI were obtained from Life Technologies and used for confocal microscopy . γ-irradiated Mtb H37Rv ( γ-Mtb ) was obtained from BEI Resources . Highly purified mouse recombinant IL-6 with a specific activity of 1 108 units/mg was purchased from BioLegend ( Bedford , MA ) . T2DM was induced by combined administration of STZ and NA . STZ was dissolved in a 50 mM citric acid buffer and administered ( 180 mg/kg of body weight ) intraperitoneally three times , with an interval of 10 days between doses . NA was dissolved in saline and administered intraperitoneally ( 60 mg/kg of body weight ) 15 min before STZ . Mice were fasted for 16 h before the STZ and NA injections . Blood glucose was measured using a glucometer at weekly intervals for up to 8 months . Mice were considered diabetic if their blood glucose was > 250 mg/dl . Control mouse blood glucose levels were always between 80 and 100 mg/dl . Serum insulin levels in fasting ( 16 h ) control and diabetic mice were measured using a Mercodia Ultrasensitive Insulin ELISA Kit ( Mercodia AB Uppsala , Sweden ) . Serum free fatty acids , cholesterol , and triglyceride levels were measured using either a fluorometric or colorimetric assay ( Cayman Chemicals , USA ) , according to the manufacturer’s instructions . OGTTs were performed in control and diabetic mice after fasting ( 16 h ) . A glucose solution ( 2 . 0 g/kg ) was given orally . Blood glucose concentrations were measured 30 min before and 15 , 30 , 60 , and 120 min after administration . Murine AMs were isolated from control and T2DM mice by bronchoalveolar lavage at 1 and 6 months post-induction of diabetes . Briefly , mice were euthanized by CO2 asphyxiation . The trachea was then cannulated following a midline neck incision , and the lungs were lavaged five times with 1 . 0 ml of ice cold PBS . Alveolar cells were separated from the lavage fluid by centrifugation at 1800 RPM for 10 min . Alveolar cells were plated ( on plastic ) to permit adherence of alveolar macrophages and subsequent removal of non-adherent NKT and T lymphocytes by washing three times with normal PBS . Adherent cells were resuspended in RPMI-1640 . Highly purified AMs were used to determine whether reduced growth of Mtb was due to dysfunctional AMs . Alveolar cells were plated in 96-well tissue culture plates at a density of 2 ×105/100 μl/well , incubated for 24 h at 37°C in 5% CO2 , and washed three times with antibiotic-free RPMI-1640 . Around 98% of the cells expressed CD11c , as determined by flow cytometry . AMs were infected with Mtb H37Rv at a MOI of 1:2 . 5 ( 2 . 5 Mtb to 1 macrophage ) . This MOI was based on the viability of AMs at different MOIs for up to 7 days p . i . More than 90% of AMs were viable at this MOI . Cells were incubated for 2 h at 37°C in a humidified 5% CO2 atmosphere , washed to remove extracellular bacilli , and cultured in RPMI 1640 containing 10% heat-inactivated human serum . To quantify the intracellular growth of Mtb H37Rv , infected AMs were cultured for 5 days after which the supernatant was aspirated and AMs were lysed . Bacterial suspensions in cell lysates were ultrasonically dispersed , serially diluted , and plated in triplicate on 7H10 agar . The number of colonies was counted after 3 weeks . Before infecting mice with Mtb H37Rv , bacteria were grown in liquid medium to the mid-log phase and then frozen in aliquots at -70°C . Bacterial counts were determined by plating on 7H10 agar supplemented with oleic albumin dextrose catalase ( OADC ) . For infection , bacterial stocks were diluted in 10 ml of normal saline ( to 0 . 5 ×106 CFU [colony forming units]/ml , 1 ×106 CFU/ml , 2 ×106 CFU/ml , and 4 × 106 CFU/ml ) and placed in a nebulizer within an aerosol exposure chamber custom made by the University of Wisconsin . In preliminary studies , groups of three mice were exposed to the aerosol at each concentration for 15 min . After 24 h , mice were euthanized and homogenized lungs were plated on 7H10 agar plates supplemented with OADC . CFUs were counted after 14–22 days of incubation at 37°C . The concentration that deposited ~75–100 bacteria in the lung during aerosol infection was used for further studies . For some experiments , mice were treated with neutralizing anti-IL-6 antibodies . For the first set of experiments , conducted 1 month after the induction of T2DM , mice were challenged with aerosolized Mtb . At 6 months p . i . , mice received 0 . 3 mg of anti-IL-6 mAb ( BioXcell ) or isotype-matched control Ab ( rat IgG1 ) intravenously every 4 days for up to 2 months . For the next set of experiments , conducted at 6 months after the induction of T2DM , mice were infected with Mtb H37Rv . On day “0” of infection , mice received 0 . 3 mg of anti-IL-6 or isotype control Ab intravenously every 4 days for up to 2 months . One month after the induction of T2DM , mice were infected with aerosolized Mtb . On day “0” of infection , mice received 0 . 3 mg of anti-NK1 . 1 mAb or isotype control Ab intravenously every 4 days for up to 1 month . Previously , we used the same anti-NK1 . 1 ( PK136 ) antibody to deplete 95% of the NK1 . 1 cells [34] . Lungs from Mtb-infected mice were mechanically homogenized and filtered through a 70 μm cell strainer . Cells were washed twice , and mononuclear cells were isolated from lung single-cell suspensions using a one-step gradient separation method ( GE Healthcare ) , according to the manufacturer’s instructions . Some mononuclear cell populations were depleted of NK cells by positive selection using antibody-labeled magnetic beads ( Miltenyi Biotech ) , according to the manufacturer’s instructions . Lung mononuclear cells and NK cell-depleted mononuclear cells were seeded in 12-well culture plates ( 1 × 106 cells per well ) and cultured with γ-Mtb . After 48 h , cells were harvested and IL-6 expression was measured by intracellular flow cytometry . Supernatants were also collected to measure IL-6 levels by ELISA . Single-cell suspensions of pooled lung , lymph node , and murine splenocytes were prepared . CD11c+ cells were isolated by positive selection with magnetic beads ( Miltenyi Biotec ) conjugated to anti-CD11c; positively selected cells comprised > 96% CD11c+ cells , as measured by flow cytometry . NK cells were isolated by negative selection using kits obtained from Miltenyi Biotec . Isolated cells comprised > 97% CD3-NK1 . 1+ cells , as measured by flow cytometry . After the mice were euthanized , the lungs were perfused with 5 ml of PBS via the right ventricle . Lungs were mechanically homogenized and passed through a 70 μm cell strainer . The remaining red blood cells were lysed using BD Pharm Lyse ( BD Biosciences ) . Surface staining to identify leukocyte populations was then performed . For IL-6 intracellular staining , cells ( 106/ml ) were suspended in RPMI 1640 containing 10% FBS and brefeldin A ( 5 μg/ml ) , placed in 24-well culture plates , and stimulated with LPS ( 1 μg/ml ) , and incubated for a further 4 h at 37°C to allow intracellular accumulation of cytokines . Cells were then permeabilized with 0 . 1% saponin and stained for intracellular IL-6 . The cells were washed , resuspended in FACS buffer , and analyzed by flow cytometry using a FACS Calibur flow cytometer . Whole blood was diluted 1:1 with RPMI-1640 medium and cultured in 12-well plates ( 0 . 5 × 106 cells/well ) in RPMI 1640 containing penicillin/streptomycin ( 100 U/100 mg/ml ) , L-glutamine ( 2 mM ) , and HEPES ( 10 mM ) ( Invitrogen , Carlsbad , CA ) in the presence or absence of a PPD ( 10 μg/ml ) at 37°C in a humidified 5% CO2 atmosphere . In some cases , an IL-6R neutralizing Ab ( 2 . 5 μg/ml ) was added to the culture . Brefeldin A ( 10 μg/ml ) was added to the cultures 2 h before the termination of the cultures ( total culture time , 6 h ) . Cells were washed , and red blood cells were lysed with lysis buffer . The cells were fixed using Cytofix/Cytoperm buffer ( BD Biosciences ) and cryopreserved at -80°C . Intracellular staining for IFN-γ , TNF-α , IL-2 , and IL-17A was performed as previously described [9] . Mouse multiplex ELISA kits ( 23-Plex kits , Bio-Rad ) were used to measure chemokine and cytokine levels , according to the manufacturer’s instructions . Total RNA was extracted from lung leukocytes or lung tissue as described previously [39] . Total RNA was reverse transcribed using the Clone AMV First-Strand cDNA synthesis kit ( Life Technologies ) . Real-time PCR was performed using the Quantitect SYBR Green PCR kit ( Qiagen ) in a sealed 96-well microtiter plate ( Applied Biosystems ) on a spectrofluorometric thermal cycler ( 7700 PRISM; Applied Biosystems ) . PCR reactions were performed in triplicate as follows: 95°C for 10 min , followed by 45 cycles of 95°C for 15 s , 60°C for 30 s , and 72°C for 30 s . All samples were normalized to the amount of β-actin/GAPDH transcript present in each sample . The primers used in the study are listed in Table 2 . Lungs were inflated and fixed in 10% neutral buffered formalin ( v/v ) for 24 h . Tissue sections were stained with hematoxylin and eosin . A semi-quantitative analysis was performed using a score from 0 ( no inflammation ) to 4 ( severe inflammation ) for each of the following criteria: alveolar wall inflammation , alveoli destruction , leukocyte infiltration , and perivascular inflammation . Immunostaining of thin paraffin-fixed lung sections was performed using antibodies against IL-6 according to the manufacturer’s instructions ( Novus Biologicals , USA ) . Unstained sections of formalin-fixed lung tissue from paraffin blocks were first deparaffinized and subjected to antigen retrieval in a citrate buffer at 95°C , as previously described . Endogenous peroxidase activity was blocked by addition of 3% H2O2 in methanol . Slides were incubated in 3% BSA in TBS for 2 min , after which primary antibodies were added at predetermined dilutions in TBS-Tween + 1% BSA ( 1:100 ) for 1 h at 25°C . Sections were then washed three times in TBS-T for 15 min . The IL-6 antigen was detected by IHC and the DAB ( DAKO ) chromogen , as previously described . Lung inflammation [40 , 41] and immunohistochemical readouts were independently assessed by two investigators as previously described [42] . The H-score was determined according to the method described by Pirker et al . Briefly , the percentage of cells with different staining intensities was determined by visual assessment and assigned a score ( 1+ for light staining , 2+ for intermediate staining , and 3+ for dark staining ) using the ImageJ IHC profiler . The H-score was calculated using the formula 1 × ( % of 1 + cells ) + 2 × ( % of 2 + cells ) + 3 × ( % of 3 + cells ) . Confocal microscopy was performed to colocalize IL-6-producing CD11c+ and NK1 . 1+ cells in lung sections . Nonspecific binding was blocked with 1% goat serum in PBS for 30 min . The slides were then incubated at 4°C overnight with hamster monoclonal anti-CD11c ( Abcam ) , rabbit polyclonal anti-NK 1 . 1 ( Bioss ) , and rat monoclonal anti-IL-6 ( Novus Bio ) antibodies . Subsequently , the slides were washed thoroughly using 1 × PBS . Then , cells were stained with the respective secondary antibodies ( goat anti-hamster IgG-Alexa 568 , goat anti-rabbit-Alexa 647 , or donkey anti-rat-Alexa 488; Life Technologies ) , washed with PBS , and mounted with Prolong Gold antifade reagent containing DAPI ( Life Technologies , USA ) . The slides were examined and analyzed under a laser-scanning confocal microscope ( Zeiss LSM 510 Meta laser-scanning confocal microscope ) . The efficacy of in vivo IL-6 neutralization was assessed in a IL-6 bioassay using the 8G2 cell line ( an IL-6-dependent murine hybridoma cell line LS132 . 8G2 ) instead of the B9 cell line as previously described [43] . Data analyses were performed using GRAPHPAD PRISM ( GraphPad Software , Inc . , La Jolla , CA ) . The results are expressed as the mean ± SE . For normally distributed data , comparisons between groups were performed using a paired or unpaired t-test and ANOVA , as appropriate . Statistically significant differences between two clinical groups were analyzed using the non-parametric Mann–Whitney U-test . Data are deposited in the Dryad Data Repository: ( doi:10 . 5061/dryad . qn42t ) [44] .
|
In the current study , we employed an experimentally induced type 2 diabetes mellitus ( T2DM ) model in wild type C57BL/6 mice and investigated the immune response to Mycobacterium tuberculosis ( Mtb ) infection . We found that natural killer ( NK ) and CD11c+ cell interactions in Mtb-infected T2DM mice led to increased IL-6 production , which drives the pathological immune response and reduces survival of Mtb-infected T2DM mice . We also found that IL-6 increases inflammatory cytokine production in pulmonary tuberculosis patients with T2DM . The NK-CD11c+ axis and the IL-6 pathway may be promising new targets for host-directed therapies aimed at reducing the severity of immune pathology , which drives morbidity and mortality in those infected by tuberculosis ( TB ) . The study demonstrates for the first time that NK-CD11c+ cell interactions increase IL-6-mediated inflammation and reduce survival in T2DM mice infected with Mtb . The NK-CD11c+ cell axis appears to be a promising new target for reducing inflammation and mortality in tuberculosis patients with type 2 diabetes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"blood",
"cells",
"flow",
"cytometry",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"cytokines",
"immune",
"cells",
"immunology",
"tropical",
"diseases",
"diabetes",
"mellitus",
"animal",
"models",
"bacterial",
"diseases",
"developmental",
"biology",
"endocrine",
"disorders",
"model",
"organisms",
"molecular",
"development",
"bacteria",
"type",
"2",
"diabetes",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"spectrum",
"analysis",
"techniques",
"white",
"blood",
"cells",
"animal",
"cells",
"tuberculosis",
"endocrinology",
"mouse",
"models",
"actinobacteria",
"metabolic",
"disorders",
"spectrophotometry",
"immune",
"system",
"cytophotometry",
"cell",
"biology",
"mycobacterium",
"tuberculosis",
"nk",
"cells",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"organisms"
] |
2016
|
NK-CD11c+ Cell Crosstalk in Diabetes Enhances IL-6-Mediated Inflammation during Mycobacterium tuberculosis Infection
|
Nipah virus ( NiV ) is a paramyxovirus that infects host cells through the coordinated efforts of two envelope glycoproteins . The G glycoprotein attaches to cell receptors , triggering the fusion ( F ) glycoprotein to execute membrane fusion . Here we report the first crystal structure of the pre-fusion form of the NiV-F glycoprotein ectodomain . Interestingly this structure also revealed a hexamer-of-trimers encircling a central axis . Electron tomography of Nipah virus-like particles supported the hexameric pre-fusion model , and biochemical analyses supported the hexamer-of-trimers F assembly in solution . Importantly , structure-assisted site-directed mutagenesis of the interfaces between F trimers highlighted the functional relevance of the hexameric assembly . Shown here , in both cell-cell fusion and virus-cell fusion systems , our results suggested that this hexamer-of-trimers assembly was important during fusion pore formation . We propose that this assembly would stabilize the pre-fusion F conformation prior to cell attachment and facilitate the coordinated transition to a post-fusion conformation of all six F trimers upon triggering of a single trimer . Together , our data reveal a novel and functional pre-fusion architecture of a paramyxoviral fusion glycoprotein .
Henipavirus , a relatively recently recognized viral genus in the family Paramyxoviridae , comprises likely over 20 species , including three established species: Hendra ( HeV ) , Nipah ( NiV ) and Cedar ( CedPV ) viruses , with HeV and NiV well-recognized as highly pathogenic agents for both humans and animals [1–6] . Henipaviruses have two surface spike glycoproteins . The G glycoprotein attaches to cell surface receptors , and upon receptor binding triggers the F glycoprotein to execute virus-host cell membrane fusion , facilitating viral entry [7–9] . The host cell receptor proteins employed by the henipaviruses are B-class ephrin molecules [10–12] . The henipavirus F glycoprotein is a trimeric class I transmembrane glycoprotein synthesized as a precursor F0 that undergoes post-translational cleavage by host cell cathepsin-L within the endosomal compartment , yielding the fusogenic F1 and F2 subunits held together by a disulfide bond [13–16] . Crystal structures of other paramyxovirus F glycoproteins in both pre-fusion and post-fusion forms have been reported , supporting a model of the F glycoprotein undergoing a transition from a metastable pre-fusion state to a more thermodynamically stable post-fusion state upon activation [17–23] . This transition brings together the viral envelope and host cell membrane to facilitate membrane fusion and viral entry . Additionally , the same viral glycoproteins facilitate viral spread from infected to naïve cells by a similar cell-cell fusion mechanism ( syncytia formation ) . However , many of the molecular details of the membrane fusion process remain elusive . Paramyxovirus envelope-host cell membrane fusion likely shares common features with other types of viral and cellular membrane fusion processes , such as influenza virus entry , and synaptic vesicle fusion within neuronal cells . While influenza entry is mediated by viral glycoprotein trimers , synaptic vesicle fusion is triggered by SNARE ( soluble N-ethylmaleimide–sensitive factor attachment glycoprotein receptor ) molecules that functionally resemble the F glycoprotein oligomers . It has been suggested that at least three hemagglutinin trimers are required for influenza virus entry [24 , 25] . It has also been demonstrated that at least three copies of SNAREpins are required for keeping the nascent fusion pore open long enough to ensure efficient neurotransmitter release [26 , 27] . However , cooperation of multiple fusion proteins ( F glycoproteins ) in the paramyxovirus entry process has not yet been demonstrated . Here we report a 3 . 4 Å crystal structure of the NiV-F glycoprotein in its pre-fusion form . Interestingly , the structure reveals a hexameric pre-fusion assembly consisting of six copies of F glycoprotein trimers encircling a central axis . Electron tomography and structure-based studies in the context of cell-cell fusion and virus-entry further support the functional relevance of the hexamer-of-trimers assembly . Our findings suggest a cooperative F glycoprotein trimer activation model in the NiV entry process , providing insight into the viral-cell and cell-cell membrane fusion mechanisms .
The extracellular region of the NiV-F glycoprotein ( residues 1–488 ) was expressed and purified as described previously [28 , 29] . A GCNt helical bundle motif was fused to the C-terminus of the F glycoprotein to stabilize it in the pre-fusion conformation . After successful crystallization and diffraction data collection , the crystal structure of this construct was determined at 3 . 4 Å resolution by molecular replacement , using the PIV5 pre-fusion F glycoprotein structure ( PDB ID 2B9B ) as a search model [20] , and was refined to Rwork/Rfree of 23 . 0/24 . 8% ( for details see Materials and Methods ) . The NiV-F trimer has a “tree-like” overall shape , with the three copies of the F glycoprotein twined around a central axis ( Axis-T ) that is parallel to the C-terminal helical bundle ( Fig 1 ) . The fusion peptide ( FP ) , residing in the N-terminal segment ( residues 110–122 ) of the F1 subunit , is docked into a groove formed by the F1 subunit of a neighboring F molecule within the trimer . Furthermore , the C-terminus of F2 and N-terminus of the FP fold into a β-hairpin ( S1-S2 ) , that forms a continuous β-sheet with β-strands S3-S6 in the F1 subunit , thus securing the position of the FP and stabilizing the pre-fusion state ( Fig 1 , inset ) . The cathepsin-L cleavage site ( R109-L110 ) , located on the tip of the β-hairpin S1-S2 , is easily accessible for cleavage in the F trimer due to its surface exposure and structural flexibility . The C-terminus of the NiV-F construct ( after D482 ) and the GCN tag are not visible in our structure , suggesting that these regions are disordered . The structure of the NiV-F glycoprotein most closely resembles the pre-fusion structure of the PIV5 F [20] glycoprotein and the two can be superimposed with Rmsd of 2 . 5 Å between 1160 Cα atoms ( Fig A in S1 Text ) . This similarity is in line with the 29% identity in their primary sequences . This similarity also suggests that these F glycoproteins undergo a similar pre- to post-fusion transition . Interestingly , the β-hairpin ( S1-S2 ) adopts a conformation similar to that of the cleaved activated ( CA ) -PIV5 pre-fusion F glycoprotein [30] , rather than of the uncleaved form [20] ( Fig A in S1 Text , inset ) . The NiV-F glycoprotein was expressed in human ( HEK293 ) cells , and N-linked carbohydrate moieties were modeled in the known four utilized N-glycosylated sites [31–33] N99 , N414 , N465 & N485 ( Fig 1 ) . NiV-F crystallized in the H3 space group , with two F trimers in each crystallographic asymmetric unit ( AU ) , which , together with the other four counterparts in the two neighboring AUs , form a hexagonal arrangement of six F trimers surrounding a central axis ( Fig 2 ) . The Axis-Ts of the six F glycoprotein trimers form alternative 45° and 135° angles to the hexagonal axis ( Axis-H ) . The C-terminal helix-bundles of the six trimeric F glycoproteins are all in the same face of the hexameric assembly , with three of them pointing inwards , and the other three pointing outwards . Such an arrangement renders the maximum contact area between neighboring F trimers due to their “tree-like” shape and surface glycan arrangements . The observed hexameric assembly is generated via two distinct F-trimer/F-trimer interfaces as indicated on Fig 2 , burying approximately 1200 Å2 and 1100 Å2 surface area , respectively , on each interacting F trimer . Interface 1 ( Fig 2C , inset ) is mainly between the S1-S6 β-sheet and a hydrophobic patch on the surface of the adjacent F trimer . Intriguingly , the core of interface 1 is formed by insertion of the tip of β-hairpin S1-S2 into a hydrophobic cavity . Specifically , the side chain of R109 is embedded into a groove defined by residues P52 , L53 , Y248 , L256 , F282 , P283 and I284 of the adjacent F trimer . R109 is further stabilized by a hydrogen bond between its main chain nitrogen and the hydroxyl group of Y248 . The positive charge of R109 is neutralized by the neighboring D107 . Residues L108 , L110 and A111 , located on the tip of β-hairpin S1-S2 , together with Q393 and G398 , located in β-strand S5 and S6 , contribute additional hydrophobic interactions . Interface 2 ( Fig 2A , inset ) is composed of two similar contact regions involving multiple hydrogen bonds and van der Waals’ interactions . In each of these regions , β-sheet S1-S6 on one F trimer interacts with the surface of the adjacent F trimer . The tip of β-hairpin S1-S2 is not visible in the electron density at interface 2 , presumably because of its flexibility . The two interfaces in the hexameric assembly would stabilize the pre-fusion F conformation by burying β-sheet S1-S6 , and preventing separation of the F1 and F2 subunits . We next investigated the oligomeric state of NiV-F in solution . Gel-filtration assays indicated that at low concentrations ( ~1–2 mg/ml ) the NiV-sF protein exists as a simple trimer in solution [28] . However , previous cryo-EM studies documented that the paramyxovirus F glycoprotein was densely packed in distinct patches on the virus surface [34] , which represent an enriched , high protein concentration 2-D environment . Therefore , purified soluble NiV-F was concentrated to simulate physiological conditions and evaluated oligomer formation using a crosslinking assay . The results are presented in Fig 3A , indicating that higher-order oligomers of NiV-sF , including the apparent molecular weight equivalent of hexamers-of-trimers ( top arrow ) , are indeed formed in solution at high concentrations ( ~20–45 mg/ml ) . No oligomers larger than hexamers-of-trimers were observed . Additionally , although the majority of oligomers were not hexamers-of-trimers , but lower-order oligomers , it is likely that cross-linking is not sensitive enough to capture every single hexamer-of-trimers . Furthermore , examination of the cross-linked species by negative stain EM revealed distinct oligomeric forms including hexamers-of-trimers ( Fig 3B ) , which were the predominant oligomeric form ( see Fig B in S1 Text ) . These oligomers appear somewhat non-uniform and varying in appearance , suggesting that the presence of the lipid membrane and additional steric constraints along the envelope surface may be required for stabilization of the hexamer-of-trimer architecture . To examine this hypothesis we employed electron tomography ( ET ) to visualize the arrangement of NiV-F in a native membrane environment ( Fig 3C ) . In agreement with both the crystal structure and EM visualization of cross-linked NiV-F , the majority of F trimers were observed as hexamers-of-trimers ( in addition to single F trimers–see Fig B in S1 Text ) on viral-like particles ( VLPs ) with full-length F . Moreover , interactions between hexamers-of-trimers were observed , with one trimer being part of more than one hexamer-of-trimers , sometimes resembling a soccer ball arrangement . Although one cannot exclude artifacts due to dehydration during ET and/or heavy metal deposition during negative staining of the VLPs , the shapes and sizes of the observed NiV-F spikes and spike assemblies are fully consistent with the structures of the pre-fusion F trimers , the crosslinking results , and the hexamers-of-trimers observed by X-ray crystallography . Next , to evaluate whether there is any functional relevance for the NiV-F hexamer-of-trimer assembly , structure-based targeted mutations were designed to either destabilize ( L53D and V108D ) or stabilize ( R109L and Q393L ) the hexameric interfaces . Residues L53 , V108 and R109 are among the key components in the formation of the hexameric interface 1 hydrophobic core ( Fig 2C , inset ) ; while residue Q393 , located in β-sheet S5 ( insets of Fig 2A and 2C ) , is involved in forming both interfaces 1 and 2 . L53D , for example , would abrogate the hydrophobic groove in interface 1 , while V108D would not only decrease the local hydrophobicity , but also create a repulsive electric force with D252 , and weaken interface 1 . On the other hand , the R109L substitution would favor embedding the side chain of this residue in the hydrophobic groove of the adjacent F trimer . Similarly , Q393L would not only enhance the hydrophobicity of interface 1 , but would also form favorable van der Waals interactions with L53 and I284 of the adjacent F trimer in interface 2 . These four F mutants were evaluated first in a HeLa-USU/HEK293T cell-cell fusion assay ( Fig 4A and 4B ) . Remarkably , the mutations predicted to impair the hexameric interfaces reduced the cell-cell fusion efficiency ( P<0 . 01 ) , while the mutations predicted to stabilize the hexamer-of-trimer interfaces tended to enhance cell-cell fusion efficiency , in comparison to wild-type ( wt ) NiV-F . The normalized fusion levels shown in Fig 4A already take into account glycoprotein cell surface expression levels . Polyclonal anti-F specific antibody and conformation-specific monoclonal antibodies ( mAbs 5B3 , 5E5 , 12B2 ) were used to confirm that these mutations in the hexameric interfaces did not affect normal F processing and the overall folding and structure of the F glycoprotein ( Fig 4B and 4C ) . Residue L53 is also a critical residue within the epitope recognized by mAb 5B3 ( Chan and Broder , manuscript in preparation ) , as revealed by its defective binding to the 5B3 mAb ( Fig 4C ) . To confirm the importance of the hexamer-of-trimer assembly and interfaces for virus-host cell membrane fusion , the NiV-F mutants were next tested in the context of virion entry . Specifically , a well characterized pseudotyped virus entry assay [31 , 35] was used to compare the entry efficiencies of viral particles containing either wt or mutant NiV-F over several logs of viral input . The results ( Fig 5 ) were fully consistent with the results derived from the cell-cell fusion assay . It should be noted that while R109L showed the same viral entry levels as WT F , the overall level of incorporation of R109L into VSV particles was significantly lower than WT F levels , suggesting that this mutant indeed supports higher entry efficiency than the WT F glycoprotein .
A synergetic NiV-F activation and a cooperative fusion-pore opening model can be deduced from our results as depicted in Fig 6 . Significant conformational changes and movement of the FP region are thought to occur during F protein transition from pre-fusion to post-fusion states . The β-sheet S1-S6 structure ( insets of Figs 1 , 2A , 2C and S1 Text ) partially sequesters the FP , therefore stabilizing the pre-fusion F conformation . Disruption of this β-sheet would free the FP and facilitate the energetically favorable transition to a post-fusion conformation , providing an efficient way to trigger F protein activation . In the hexameric F assembly , β-sheet S1-S6 is buried in the hexameric interface and , therefore , the pre-fusion F conformation is stabilized . On the other hand , disturbance or dissociation of the hexameric assembly would apply mechanical force on the interface-forming β-sheet S1-S6 triggering F protein activation . The residues on the tip of β-hairpin S1-S2 , as well as the hexameric interface-forming residues on β-strands S3-S6 , are therefore crucial for the regulation of NiV-F activation and fusion efficiency in this model , and we refer to these sites as “priming sites” . Each F-trimer contains 3 priming sites–two of which are buried within the hexamer interfaces , while the third is available for interaction with ( and priming by ) the NiV-G/ephrin complex ( Fig 6 ) . Consistent with this model , several of the residues in this “priming site” have been previously shown to mediate the viral fusion-attachment ( F-G ) protein interactions in the PIV5 virus [36] . Upon viral attachment , the ephrin-mediated re-arrangements of NiV-G would exert a triggering disturbance at an exposed “priming site” ( presumably via direct G-F priming-site interactions ) . This triggers the F glycoprotein trimer to undergo a conformational change from a pre-fusion to a post-fusion configuration . Upon activation by a NiV-G/ephrin complex , the contacted F trimer would not only destabilize the pre-fusion state in its interacting neighbors , but could also directly facilitate an alteration in their “priming sites , ” triggering a conformation switch in the rest of the hexamer-of-trimer assembly . Thus , a single ephrin/NiV-G/F trimer interaction would result in the synergetic switch from a pre-fusion to a post-fusion conformation in all six F trimers within the assembly ( Fig 6 ) . The resulting eighteen copies of the post-fusion F protein ( forming eighteen six-helix bundles ) is more likely to pull the two membranes together than triggering of a single F trimer alone . The combination of biochemical and microscopical data ( Fig 3 ) , and functional cell-cell fusion ( Fig 4 ) , and viral entry ( Fig 5 ) data , is consistent with this model . The necessity for multiple fusion glycoproteins acting together to affect membrane fusion has been most clearly demonstrated for the SNAREpins ( composed of v- and t- SNAREs ) during the neurotransmitter cargo-releasing fusion process in synapses [26 , 27] . Several lines of evidence suggest similar requirements for various enveloped viruses , including influenza , HIV , baculovirus , Semliki Forest virus , and vesicular stomatitis virus [24 , 25 , 37–43] . Multiple copies of fusion glycoproteins around the entry site would provide a radial force around the newly fused lipid bilayers that cooperatively stabilizes and opens up the nascent fusion pore , allowing efficient virus entry into the host cell . The core fusion machinery for NiV entry consists of a single glycoprotein , F , which localizes at the viral envelope prior to viral entry . Furthermore , the attachment and the fusion process are exerted by two separate glycoproteins ( G and F , respectively ) . Therefore , the oligomeric assemblies of F glycoproteins and their synergetic activation triggered by the ephrin/G complexes provide an efficient way to ensure the simultaneous availability of multiple NiV-F helix bundles at the contact zone where G glycoproteins are attracted to the host cell-surface ephrins . Noteworthy , the cathepsin cleavage site ( R109/L110 ) is accessible in the trimer ( Fig 1 ) , but fairly buried in interface 1 of the hexameric assembly ( Fig 2 ) . So how/when does the cathepsin cleavage occur ? F glycoproteins can exist as trimers on the cell surface and on virions in both cleaved and uncleaved forms [44 , 45] . Following cathepsin-L cleavage F could shift primarily to the hexamer-of-trimer assembly . Alternatively however , cleavage of F may occur even in the hexamer-of-trimer structural assembly , as the cleavage site is not completely blocked in all hexameric interfaces . In addition , it is likely that R109 and L110 are not the only determinants of F cleavage , as at least F mutant R109 is still cleaved ( Figs 4C and 5B ) . It is clear that F trimers traffic to the plasma membrane before cleavage and then are cycled back into the cell where cleavage occurs , and then recycled back to the cell surface , and both cleaved and uncleaved F exist inside and on the surface of expressing cells [14 , 35 , 46] . It is also possible that both cleaved and uncleaved F monomers co-exist within a trimer and/or within the hexamer-of-trimer assembly . Furthermore , it is also possible that a hexameric structure is not static and that trimers may form and disassemble from higher order oligomeric forms allowing for cleavage to occur . Further experimentation is required to unequivocally establish this mechanistic model . Although evidence in this study suggest that a hexamer-of-trimers assembly is a functional form for the F protein , it remains to be determined whether this is the primary assembly form of F in the presence of G . Furthermore , studies of other paramyxoviruses indicate that the transmembrane regions of the F proteins are critical to stabilize the pre-fusion structure , often requiring a stabilizing GCN4 domain [20 , 30] . Consistently , our pre-fusion structure and hexamer-of-trimer assembly observed for NiV-F was a result of addition of a GCN4t domain to replace the F transmembrane and cytoplasmic tail regions . It remains to be determined whether the same structural features will be observed for the full-length F proteins embedded in a cellular membrane . Technical challenges prevent such studies from being performed under current available technologies . However , the prevalence of F hexamers-of-trimers in VLPs containing full-length F without a GCN4t domain , in combination with our functional studies , support our current F pre-fusion and hexamer-of-trimers structural results . In addition , it remains to be determined whether this relatively simple mechanistic model applies to other enveloped viruses . It should be noted that various oligomeric viral fusion-protein assemblies , e . g . containing 3–10 copies , could serve the same purpose in other viral systems by potentially stabilizing the pre-fusion conformation and by facilitating a cooperative transition to the post-fusion conformation in all glycoproteins within the assembly upon transition of a single fusion protein .
The construction , expression and purification of the GCNt stabilized pre-fusion soluble version of the NiV-F glycoprotein ( NiV-sFGCNt ) had been detailed previously [28 , 29] . Briefly , The predicted transmembrane ( TM ) anchor domain ( residues 488–510 ) and the C-terminal cytoplasmic tail ( CT ) domain ( residues 511–546 ) of the NiV- F were replaced by the GCNt motif ( MKQIEDKIEEILSKIYHIENEIARIKKLIGE ) [47] in heptad phase followed by a Factor Xa protease cleavage site ( IEGR ) and the S-peptide tag . A 293T stable cell line expressing the NiV-sFGCNt was generated . Preparation of the sF glycoprotein from 293T stable cell lines was carried out using serum-free culture conditions and employing a combination of S-protein agarose ( EMD Biosciences ) affinity column and size exclusion chromatography using a HiLoad 16/60 Superdex 200 preparative grade gel filtration column ( GE Healthcare ) steps to isolate pure NiV-sFGCNt trimer . Crystals of NiV-sFGCNt formed after two weeks in 2μl hanging drops by vapor diffusion against reservoir solution ( in a 1:1 protein-to-reservoir-solution ratio ) containing 0 . 1 M HEPES pH 7 . 3 , 1% Jeffamine ED-2001 and 1 . 2 M Sodium Malonate . For cryo-protection , crystals were soaked step-wise in reservoir solution starting from 5% up to 25% glycerol and were flash-frozen in liquid nitrogen . Diffraction data were collected at beamline NE-CAT ID-24 of the Advanced Photon Source at Argonne National Laboratory and processed with HKL2000 program package[48] . The structure was determined by molecular replacement with PDB ID 2B9B ( PIV5 pre-fusion F protein ) as the searching model and MOLREP program in CCP4 Suite [49] . The structure was refined carefully with grouped B-factor refinement ( two B-factors per residue ) , and non-crystallographic symmetry ( NCS ) restrains in CNS1 . 2 [50] . NCS restrains were applied to the six chains of NiV-F in one asymmetric unit . The electron density maps for model building were improved by B-factor sharpening with a value of -70 Å2 . All model building was performed in Coot [51] . The final structure model was checked by the program PROCHECK [52] . Statistics of data collection and refinement are listed in Table A in S1 Text . Noteworthy , all conserved Cys residues aligned between the NiV and PIV5 F structures . Purified NiV-sFGCNt trimer was concentrated to 50 mg/mL using the Corning Spin-X UF 500μL Centrifugal Concentrator ( Corning Inc ) . Glutaraldehyde was added to the concentrated material to various final concentrations ranging from 0 . 005% to 0 . 1% yielding protein concentrations between 40–45 mg/mL . 5μg of cross-linked products were analyzed on Blue Native PAGE ( Invitrogen ) with Coomassie staining . To separate different oligomeric species of cross linked NiV-sFGCNt trimer , 100μg of the 50 mg/mL sF was cross-linked with 0 . 08% gluteraldehyde yielding a final protein concentration of 42 mg/mL and the cross linked material was then subjected to 10–25% sucrose gradient ultracentrifugation and fractionated . To obtain the gradient , 6 ml of 10% sucrose was underlaid with 6 ml of 25% sucrose in a polyallomer 14- by 95-mm tube . A linear sucrose gradient was generated using the Biocomp Gradient Master ( Biocomp , Frederickton , NB , Canada ) at an angle of 81 . 5° for 2 min 19 seconds at a speed of 14k rpm . The cross linked material was then overlaid on top of the gradient . The gradient was centrifuged at 40 , 000 rpm for 20 h at 4°C using an SW40 rotor ( Beckman Coulter , Inc . ) . Fractions of ~200 μl each were collected from the bottom to the top of the gradient using a Beckman fraction recovery system and automated fraction collector . To analyze the fractions , 10 μl of each collected fraction was resolved on 3–12% BN-PAGE ( Invitrogen ) followed by western blotting . F oligomers were probed by monoclonal mouse anti-F specific antibody . Fractions with higher oligomers were pooled , desalted , concentrated and analyzed on Blue Native PAGE using Coomassie staining . Samples were prepared using conventional negative staining protocols [53] . Briefly , 3 μL of sample was pipetted onto a glow-discharged carbon-coated grid and stained with 1% ( w/v ) uranyl formate . Imaging was performed at room temperature with a Morgagni 268 ( D ) transmission electron microscope ( FEI Company ) at 100kV at a magnification of 30 , 416x . NiV-F VLPs were produced by expressing NiV-F on 293T cells and then collecting , clearing , and purifying the cell supernatants through 20% sucrose , as previously established [54] . It has also been established that NiV-F can autonomously assemble and produce budding of NiV-F VLPs [55] . 3 μl VLP sample was applied on carbon covered copper grids . After one minute the fluid was absorbed with filter paper and 3 μL urinyl acetate ( UA ) ( 2 . 5% ) was applied for 40 seconds . Excess UA was absorbed and the grid was left to dry for 2 minutes before being transferred to a Gatan-626 specimen holder . A Tecnai F20 transmission electron microscope ( 200kV ) ( FEI Company , OR ) equipped with a field mission gun ( FEG ) , computer-controlled compustage , a TIETZ F415MP 16 megapixel CCD camera was used to obtain a total of 141 tomographic images at tilts from -70 to +70 degrees . Tilt series were collected at 29 , 000x magnification with an applied defocus of -3 μm with FEI’s Batchtomography package . Tilt series alignment and tomographic reconstruction were performed by IMOD [56] . Fusion between NiV F and G glycoprotein-expressing effector cells and permissive target cells was measured by a β-Galactosidase ( β-Gal ) assay that was previously described [57] . Briefly , plasmids encoding WT NiV F or each mutant of F and NiV G at a 1:1 ratio or control/mock transfection were transfected into HeLa-USU effector cells . The transfected cells were then infected with vaccinia virus-encoding T7 RNA polymerase the following day . HEK293T cells served as receptor-positive target cells were also infected with the E . coli Lac Z-encoding reporter vaccinia virus . Cells were infected at an MOI of 10 and incubated at 31°C overnight . Cell-cell fusion reactions were conducted by incubating the target and effector cell mixtures at a ratio of 1:1 in 96-well plates at 37°C . Cytosine arabinoside ( 40μg/ml ) was added to the fusion reaction mixture to reduce nonspecific β-Gal production . Nonidet P-40 was added to 0 . 5% final concentration at 2 . 5 hrs , and aliquots of the lysates were assayed for β-Gal at room temperature with the substrate chlorophenol red–D-galactopyranoside ( Roche ) . Assays were performed in triplicate , and fusion results were calculated and expressed as rates of β-Gal activity ( change in optical density at 570 nm per minute x 1 , 000 ) in VersaMAX microplate reader ( Molecular Devices , Sunyvale , CA ) . Equal amount of leftover envelope glycoprotein expressing effector cells from each fusion reaction was used for evaluation of total F and G expression by immunoprecipitation and cell surface F expression by flow cytometry . For immunoprecipitation , cells were lysed and clarified by centrifugation . The lysates were then subjected to rabbit polyclonal anti-F or -G specific antisera and protein G Sepharose precipitation followed by SDS PAGE and western blot analysis . The blots were probed with F or G specific murine mAbs . For flow cytometry , the envelope expressing cells were washed once with PBS and then incubated with pre-fusion specific F mAb followed by incubation with fluorescein isothiocyanate ( FITC ) -conjugated polyclonal anti mouse Ab ( Cell Signaling Technology , Inc . , Danvers , MA ) . All incubations were in PBS with 3% goat serum and on ice for 1 h . Samples were washed three times with cold PBS between incubations and before being fixed in 1 . 6% paraformaldehyde and then analyzed on a Beckman Coulter Epics XL flow cytometer . The individual cell fusion reactions mediated by each mutant were converted to percentages of WT fusion activity and normalized with cell surface expression of WT F and each F mutants . VSV-rLuc pseudotype viruses , harboring NiV-F and–G on their surface , were produced as previously described [31] . All NiV-F ( C-terminal AU1-tag ) and–G ( C-terminal HA-tag ) constructs , utilized for the production of the pseudotyped viruses , were cloned in the pcDNA3 . 1 vector . Viral copy numbers were quantified by real-time PCR [31] and 10-fold serial dilutions of each viral prep were tested on Vero cells for viral entry using a renilla luciferase detection system ( Pierce ) [58] . Relative light units ( RLUs ) were plotted against the viral genome copy number per milliliter and analyzed by linear regression using GraphPad Prism as previously described [58] . All molecular representations were produced with PyMOL ( Delano Scientific LLC ) . Figures were prepared using Adobe Illustrator , Adobe Photoshop .
|
Paramyxoviruses infect host cells through the coordinated functions of two envelope glycoproteins . The G glycoprotein attaches to cell receptors triggering the fusion ( F ) glycoprotein to execute membrane fusion . The crystal structure of the NiV-F protein has not been reported . Additionally , many molecular details of the virus-cell fusion process remain elusive , including how the higher-energy pre-fusion conformation state of the F glycoprotein is stabilized , or how many copies of the F glycoprotein are required for fusion . This manuscript reports the pre-fusion crystal structure of NiV-F glycoprotein , and a functional hexamer-of-trimers assembly , with six F trimers encircling a central axis . Multidisciplinary data suggested that this assembly plays a role in the stability of the pre-fusion F conformation prior to cell attachment and F-triggering to a post-fusion conformation . Thus this assembly may coordinate this transition in all six F trimers upon triggering of a single trimer during membrane fusion pore formation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Crystal Structure of the Pre-fusion Nipah Virus Fusion Glycoprotein Reveals a Novel Hexamer-of-Trimers Assembly
|
Synaptic plasticity is considered to play a crucial role in the experience-dependent self-organization of local cortical networks . In the absence of sensory stimuli , cerebral cortex exhibits spontaneous membrane potential transitions between an UP and a DOWN state . To reveal how cortical networks develop spontaneous activity , or conversely , how spontaneous activity structures cortical networks , we analyze the self-organization of a recurrent network model of excitatory and inhibitory neurons , which is realistic enough to replicate UP–DOWN states , with spike-timing-dependent plasticity ( STDP ) . The individual neurons in the self-organized network exhibit a variety of temporal patterns in the two-state transitions . In addition , the model develops a feed-forward network-like structure that produces a diverse repertoire of precise sequences of the UP state . Our model shows that the self-organized activity well resembles the spontaneous activity of cortical networks if STDP is accompanied by the pruning of weak synapses . These results suggest that the two-state membrane potential transitions play an active role in structuring local cortical circuits .
Cortical networks show complex dynamics of intrinsic activity when sensory inputs are absent . Whether this spontaneous activity is a mere idling state of the brain or , rather , an internal state that actively engages in brain functions remains unclear . Recent experimental studies have revealed an important characteristic of the intrinsic dynamics of cortical neurons . In vivo and in vitro cortical pyramidal neurons exhibit spontaneous transitions of the membrane potentials between a depolarizing UP state and a resting DOWN state [1]–[14] . Results of a multi-intracellular recording study showed that the onsets of the UP state spread from a local focus and that activations originating from multiple cortical sites are almost synchronous [15] . Other experiments revealed a repeated activation of neurons in neocortical slices with a diverse repertoire of precisely-timed temporal sequences [16]–[19] . Since blockade of glutermatergic synaptic transmissions eliminated the membrane potential transitions , recurrent synaptic input is considered to be crucial for these transitions . These results indicate that the spontaneous cortical activity is not merely a collection of irregular neuronal firing , but is organized into the spatiotemporal patterns that possibly reflect the structure of local cortical networks . We may raise several questions regarding this issue . How can recurrent cortical networks maintain the two-state membrane potential transitions ? Does the self-organized spontaneous activity exhibit precise temporal sequences ? What is the likely structure of the local cortical networks self-organized through the two-state transition ? Does the two-state transition exert a significant impact on the network structure ? To address these issues , we constructed a recurrent network model of pyramidal neurons and fast-spiking interneurons with synapses between pyramidal neurons modifiable by spike-timing-dependent plasticity ( STDP ) [20] . As observed in cortical networks [17] , the minority of the model's pyramidal neurons displays autonomous membrane potential transitions in the absence of synaptic input . We show that , driven by the autonomous activity , the network self-organizes repeated epochs of UP-state propagation that exhibit irregular durations and intervals . This irregular network activity well resembles the experimentally-observed cortical activity if we prune weak excitatory synapses . Furthermore , transitions from the DOWN state to the UP state ( UP transitions ) exhibit precisely-timed sequences in the self-organized network . Self-organization of spontaneous two-state transitions was studied partially in our previous model [21] . A novel finding in the present paper is a repeated activation of temporal sequences during network UP states , in which UP transitions propagate through the entire network . While such sequences were not remarkable in the previous model consisting of homogeneous neuronal populations , the inhomogeneous intrinsic properties of the present model induce neuron-dependent excitabilities that promote sequential neuronal activation . We demonstrate the role of the self-organized synapses and the nontrivial interactions between the self-organizing process and the two-state transitions in generating the repeated sequences . In addition , the non-homogeneous intrinsic properties create a wide variety of firing patterns , as observed in experiments .
Pyramidal neurons were numbered in the ascending order of the excitability , with neuron #512 having the highest excitability . In this model , the neuron-dependent excitability of each neuron was adjusted by the density of H-current . Some arguments favorable to the use of H-current are given in the discussion . In highly excitable ( HiE ) neurons , H-current is slowly activated in the DOWN state and eventually depolarizes the membrane potential to cause an UP transition . H-current is rapidly inactivated in the UP state , and a hyperpolarizing current in turn grows by slow activation of Ca2+-dependent potassium current , so the neurons may return to the DOWN state . Thus , the HiE neurons display autonomous two-state transitions through the intrinsic mechanism even without synaptic inputs . In the present study , the excitatory neurons with indices >400 were classified as the HiE neurons , which in reality may correspond to the layer 5 pyramidal neurons that initiate the spread of UP state [3] . Self-organization of recurrent synapses proceeded as in our previous network models with [21] and without [22] two-state membrane potential transitions . The pyramidal-to-pyramidal connections were initially all-to-all and had equal weights , i . e . , the half maximum conductance . Due to the activity regulation by STDP , the average weight was reduced to approximately half of the initial value ( Figure 1A , left ) , and the weights of self-organized recurrent synapses developed a bimodal distribution ( Figure 1A , right ) [23] , [24] . The average weight and bimodal profile remained unchanged once the network reached a stationary state . The self-organized synapses exhibited relatively weak competition in the recurrent network , with the synaptic weights continuously distributed from the minimum to the maximum [21] , [22] . Reverberating synaptic inputs induce input-output correlations in individual neurons , and presumably contribute to the strengthening of the weak synapses . Spontaneous activation of HiE neurons propagated synaptically to the other neurons with low excitability ( LoE neurons ) . Driven by recurrent synapses , activities of the neurons were strongly synchronized in the beginning of self-organization . During the initial stage transient , short-term depression at the pyramidal-to-pyramidal AMPA synapses prevented too rapid increases in firing rate of each neuron , thus preventing too rapid decreases in the maximum synaptic conductance by STDP . Thus , short-term depression made it much easier to maintain the persistent network activity . As recurrent synapses settled down on the steady distribution , pyramidal neurons started to exhibit spatiotemporal activity patterns representing the spread of UP transitions to the entire network ( network UP state ) ( Figure 1B ) . Recurrent excitation maintained the network UP state , until the negative feedback effects by the activation of Ca2+-dependent potassium currents , inhibitory interneurons and synaptic depression would terminate it . Effects of the various inhibitory feedbacks on self-organization were quantitatively studied in Figure S1A . In general , a weaker inhibition results in higher firing rates of excitatory neurons during network UP states . Since the synapses are significantly modified by STDP during these states , the average synaptic weight in the steady state was decreased ( as STDP is LTD-dominant ) . However , an overly weak inhibition failed to produce epochs of the DOWN state in the spontaneous network activity ( see the slowly decaying curves in some of the leftmost panels ) . Such a continuous UP state made the synaptic weights too weak to maintain the persistent network activity ( asterisks in the rightmost panels ) . Figure S1B schematically summarizes the resultant patterns of network activity . If the network activity could survive the initial-stage down regulation by LTD , a steady state with a moderate firing rate was robustly obtained for additive STDP [23] with different timing windows or different LTD/LTP area ratios ( Figure S1C ) . In contrast , the steady state with a low firing rate was not achieved by other rules of STDP that do not exhibit activity regulation , e . g . , a multiplicative STDP rule [25] ( Figure S1D ) . In such a case , all neurons in the network displayed very high-frequency ( >150 Hz ) tonic firings , in which searching for a structure in network activity is meaningless . The result is also consistent with that of a recent study of self-organization without the two-state transitions using a variant of multiplicative rule [26] . The model neurons displayed a broad range of firing patterns depending on their temporal positions in the activity spread . The HiE neurons exhibited brief UP states with a few spikes , whereas the LoE neurons displayed long-lasting UP states with bursts of spikes ( Figure 1C , leftmost column ) . The different activity patterns resulted in quite different profiles of the bi-modal membrane-potential distributions ( Figure 1C , middle column ) . The distributions exhibited clearly distinct bimodal peaks in the LoE neurons , whereas such peaks were obscure in the HiE neurons . These results are consistent with the experimentally observed variation of membrane potential distributions [1]–[5] , [14] , [16] . We note that the distributions of the synaptic weights show quite different profiles in the two neuron types ( Figure 1C , rightmost column ) . In HiE neurons , the distribution was strongly biased towards weak synapses due to the asymmetric flow of neuronal activity from HiE to LoE neurons . Why do pyramidal neurons display this variation of firing patterns ? In particular , why do LoE neurons exhibit highly irregular firing patterns when they are driven by the near-regular firing of HiE neurons ? In our model , the complex interactions between recurrent inhibition and the inward-rectifier K+ current IAR enhance the irregular firing of LoE neurons ( Figure 1D ) . A LoE neuron may exhibit UP transitions when it receives volleys of excitatory synaptic input . However , such input does not necessarily elicit a prolonged spike generation . A strong excitatory input is often followed by a strong inhibitory synaptic input , which may quickly hyperpolarize the membrane potential to activate IAR . Then , the neuron may briefly stop firing , or even make a DOWN transition . Thus , the inhibitory regulation , at least partly , causes irregular firing of the driven neurons . To get an insight into the network dynamics , we show the structure of the self-organized neuronal wiring in Figure 2A . Here , we selected every 8 neurons from LoE to HiE neurons ( i . e . , 64 neurons in total ) and arranged them anti-clockwise along a ring in the ascending order of the excitability , with the least excitable neuron at the three o'clock position . Hereafter , we call the information flow directed from HiE to LoE neurons “feed-forward” . Strong , modest and weak synaptic connections are separately shown , and red or blue lines indicate feed-forward or feedback connections , respectively . The figures show that the neuronal wiring is highly asymmetric , i . e . , most of the strongest projections are feed-forward whereas almost all weak synapses are feedback . However , some feedback projections are also strong , and they are dense especially among HiE neurons . The activities of these neurons were modulated by the slow intrinsic rhythms , so the relative times of their firing were varied in repeated network UP states . This is why STDP does not eliminate the feedback connections among them . Nevertheless , the feed-forward–dominant organization of the self-organized network is apparent from the average weights of the synapses terminated on or sent from the individual neurons ( Figure 2B ) . To see whether STDP regulates the activity of recurrent network , we shuffled all the self-organized synapses across the entire population of excitatory neurons . This manipulation keeps the distribution of synaptic weights unchanged over the whole network . However , it mixes up the synapses among the neurons and changes the weight distribution on each neuron . As a result , the synaptic mechanism to regulate neuronal activity was destroyed and the firing rate of excitatory neurons were significantly increased ( Figure 2C ) . Moreover , it eliminated the hyperpolarizing DOWN state ( hence , the spontaneous two-state transitions ) from the network activity . These results indicate that the additive STDP regulates the activity of the recurrent network , and that such a regulation is crucial for the spontaneous membrane potential fluctuations . The propagation of neuronal activity in the self-organized network may resemble that observed in cortical networks in vivo [27] . However , synchronized network UP states in the self-organized activity occurred in narrower time windows and in more regular temporal patterns than those in the cortical activity . These discrepancies were robustly seen in the spontaneous two-state transitions obtained in our simulations . Interestingly , however , eliminating weak excitatory synapses makes the self-organized network activity better resemble the in vivo cortical activity . In fact , the elimination of the weakest 40% of excitatory synapses , which involved a large fraction of the feedback connections ( Figure 2D , right ) , generated irregular activity patterns quite similar to the experimentally observed ones ( Figure 2D , left ) . Such an elimination of the recurrent synapses significantly reduced the frequency and amplitude of inhibitory feedback synaptic current ( Figure 2E ) , and allowed each neuron to stay in the UP state for longer periods of time . However , a further elimination of the synapses ( 60% elimination ) spoiled the propagation of UP states to the far downstream neurons ( Figure 2F ) . An optimal degree of the elimination exists . We have shown that the self-organized network has a primarily feed-forward neuronal wiring . We tested whether this near–feed-forward structure generates temporal sequences of UP transitions , since such sequence has been reported in cortical networks [16] . To this end , we detected the onset times of the UP state in each neuron by monitoring transient changes in the calcium concentration ( see Materials and Methods ) . By using a template matching method ( timing jitters<1 ms: see Materials and Methods ) , we looked for such precise sequences that consisted of more than two UP transitions and repeated in more than one UP-state epoch . In Figure 3A , we depicted examples of those sequences that were repeated in successive network UP states . The relative temporal relationships between different sequences ( e . g . , the red and green ones ) in general changed in the repetition . Nevertheless , the relative timing of UP transitions within each sequence little jittered . To confirm the statistical significance of the precise sequences , we constructed a set of N independent non-stationary Poisson event sequences ( N is the network size ) . As shown previously , the individual neurons display quite different temporal activation patterns depending on their relative positions in the UP-state propagation . Moreover , the driven LoE neurons participate in sequences more often than the driving HiE neurons ( Figure S2A ) , presumably reflecting the fact that the driven neurons fire in narrower temporal windows in the repeated network UP states ( Figure S2B ) . In constructing the non-stationary Poisson sequences , we attempted to preserve such a sub-structure of network activity , since it may be crucial for generating the repeated sequences . To this end , we divided the excitatory neural population into eight subgroups of equal sizes and calculated the UP-transition rates in each subgroup in successive time windows sufficiently shorter than the typical duration of network UP states . Then , in each subgroup we set the time evolution of the population event rate of the Poisson sequences equal to thus calculated time evolution of UP-transition rate . Thus , the non-stationary Poisson events preserve the spatial and temporal structure of UP transitions in larger scales , while randomizing the fine spatiotemporal structure of the events ( Figure S2C ) . We found that the number of temporal sequences in the network simulations is significantly larger than that expected by chance ( p<0 . 001 ) ( Figure 3B ) . Actually , the neuronal wiring self-organizing through STDP underlies the sequence generation in the present recurrent network . To show this , we shuffled the synaptic connections in a completely random manner . This manipulation destroys the weight distributions on the individual neurons , while preserving the distribution of the synaptic weights over the entire network . Shuffling synaptic connections eliminated most of the DOWN states , and therefore greatly reduced both the number of UP transitions and that of recurrent patterns ( Figure 3C ) . In order to rescue the DOWN states , we rescaled all the excitatory synapses by a factor less than unity . This manipulation recovered the number of UP transitions to the original level , but not that of temporal sequences . The results strongly indicate that the specific structure of self-organized neuronal wiring underlies the repeated sequences . To further elucidate the role of the self-organized synapses , we examined whether a recurrent network might generate temporal sequences without STDP . We connected pyramidal neurons randomly by excitatory synapses of identical strength , and adjusted the synaptic strength such that the network could exhibit spontaneous two-state transitions at a low rate similar to that of the self-organized network activity . The connectivity of the resultant random recurrent excitation was about 20% . The number of precise sequences was significantly smaller in the random network than in the self-organized one ( Figure S3 ) . These results confirm that the precise temporal sequences are a consequence of STDP . Excitatory neurons were activated nearly in a sequential manner in each network UP state . However , the order of activation was not strictly fixed across different epochs of the network UP state , but rather jittered from epoch to epoch ( Figure 3D ) . Moreover , unlike in experiments [16] , the sequences did not generally repeat similar temporal patterns of the subthreshold membrane-potential fluctuations ( Figure 3E ) . The results seem to reflect the fact that the present self-organized network is not a purely feed-forward network ( Figure 2A ) , which would generate only small jitters in the activation order and repeat similar fluctuating patterns of the membrane potential ( or postsynaptic current ) . It is noted that the propagation of UP transitions in cortical neurons was shown to exhibit large timing jitters during slow-wave sleep [15] . If the network is not driven by HiE neurons , the depressing effect of STDP would eventually terminate spontaneous neuronal firing during self-organization . Thus , an apparent role of the membrane potential transitions in the self-organizing process is to maintain the spontaneous activity . Do they also play an active role in sequence generation ? We studied this intriguing question by testing whether precise temporal sequences can self-organize without clearly-distinct two-state transitions . To reduce the voltage differences between the two membrane-potential states , we lifted the hyperpolarizing membrane potential of the DOWN state by reducing the maximum conductance of the inward rectifier K+ current IAR . Cortical neurons show a similar suppression of the DOWN state typically when animals are awake or in rapid-eye-movement sleep [4] . However , the mechanism of this change remains unknown , and reducing IAR is to be regarded as an ad hoc method to suppress the membrane potential transitions . The suppression of the DOWN state changed the membrane potential distributions from bimodal to unimodal ( data not shown ) and scattered action potentials uniformly over temporal domain ( Figure 4A ) without much changing the average firing rate ( Figure 4B , dashed ) . We then tested whether the recurrent network with the reduced IAR could self-organize temporal sequences . STDP failed to develop sufficiently strong synaptic connections , and the feed-forward network structure was obscure compared with the previous case ( Figure 4C , upper raw ) . Consequently , the generation of precise sequences was significantly impaired ( Figure 4D ) . The reason for this reduction of sequences may be understood as follows . The excitatory neurons are sequentially activated during network UP states according to their different excitabilities . The loss of clearly-distinct DOWN states , which was caused by the reduced IAR , suppresses the neuron-dependence of excitability and the essential difference between HiE and LoE neurons . Therefore , such a loss eliminates the sequential neuronal activation during network UP states . In addition , clearly-distinct network DOWN states reset network activity to prepare for the sequential activation . Thus , the disappearance of DOWN states prevents the development of a feed-forward–dominant network structure . We further tested the active role of clearly separated UP and DOWN states by reducing the connection probability or the maximum conductance of the interneuron-to-pyramidal GABAergic synapses . Here , IAR was reset to the original magnitude . As in the case of reduced IAR , the weak inhibitory feedback eliminated , partially or perfectly , the distinct DOWN states ( Figure 4B ) . As a result , the weak inhibition impaired the self-organization of strong synapses ( Figure 4C , middle and lower rows ) and a nontrivial structure of the synaptic matrix ( Figure 5 ) . Therefore , the self-organized networks failed to generate precise temporal sequences ( Figure 4D ) .
Blocking AMPA and NMDA synaptic currents terminates the temporally-correlated spontaneous spike sequences in cortical slices [17] . However , layer 5 pyramidal neurons with large apical dendrites exhibit spontaneous regular firing even after these receptors are blocked . In addition , layer 5 neurons initiate the spread of UP transitions in cortical slices [3] . The autonomous activity can be diminished by blocking the persistent sodium current ( NaP ) and H-current . Since H-current is abundant at the distal sites of apical dendrites in pyramidal neurons [41]–[46] , our pyramidal neuron model included these ionic currents in the dendritic compartment . The minority of the model neurons having especially rich concentrations of H-current exhibited self-sustained membrane potential transitions , which contributed to setting the entire network to spontaneous activity ( Figure 1B ) . Since LTD is the dominant component of STDP [47] , it rapidly reduces the weights of recurrent synapses during the self-organization . Thus , without the self-sustained activity , the network would easily fall into a permanent quiescent state . Previous models of UP and DOWN states hypothesized very strong recurrent connections to induce and maintain UP states , and a potassium-dependent intrinsic mechanism [48] or short-term synaptic depression [49] to terminate the UP state . However , due to the powerful down regulation by STDP , recurrent excitatory connections in our model could not remain strong enough to maintain network activity without the intrinsic drive by the HiE neurons . It is likely that under the continuous influence of STDP the persistence of spontaneous activity requires an intrinsic mechanism to initiate the UP state . As mentioned above , here the H-current serves for this role . To turn off the UP state , this model employed a potassium-dependent intrinsic mechanism , short-term synaptic depression and recurrent inhibition ( Figure 4B ) . In principle , we could induce DOWN transitions solely with any one of these mechanisms . In such cases , however , it was difficult to keep the values of the related parameters in moderate ranges . In addition , the network UP states exhibited approximately fixed temporal lengths and regular periodicity ( data not shown ) . The highly irregular repetition of network UP states , such as observed in experiments [27] , appeared if the network recruited the multiple mechanisms of DOWN transitions . H-current is generally considered to be crucial for the integration of synaptic inputs [41] , [45] . In addition , some experiments suggested that the blockade of H-current enhances , rather than suppresses , the two-state membrane potential transitions [50] . Whether this current engages in the maintenance of the two-state transitions seems open for future studies . We may replace H-current with some other mechanisms of excitability , such as neuron-dependent resting membrane potentials . In this case , however , the pacemaker activity of the HiE neurons would not be accompanied by slow subthreshold oscillations . Whether the autonomous neuronal firing observed experimentally displays the subthreshold oscillations requires a further experimental clarification . We have shown that an additive STDP rule self-organizes such a neuronal wiring that is primarily feed-forward , that is , most synaptic connections are formed to propagate activity from the self-activated HiE neurons to the driven LoE ones . However , the network also develops a non-negligible amount of relatively weak feedback synaptic connections . We have shown that pruning the weak excitatory synapses , most of which are feedback connections , broadens the epochs of network UP state and increases the epoch-to-epoch variability in the duration of network UP states ( Figure 2D ) , as seen in cortical networks [27] . In fact , this prolongation of network UP states is caused by a resultant decrease in recurrent inhibition . These results may suggest that some physiological mechanism eliminates or silences overly weak cortical synapses during self-organization . Note that STDP does not describe such an elimination of synapses . It is intriguing to experimentally test whether such a pruning of cortical synapses follows STDP . Our model predicts that the membrane potential distributions display clear bimodality for the LoE neurons located the downstream of information flow , whereas the bimodality is less obvious in HiE neurons ( Figure 1C ) . The membrane potential transitions in cortical neurons do display a wide variety of temporal profiles [1]–[5] . It seems intriguing to examine whether the experimentally observed pattern of the variations is consistent with our categorization of neurons based on their excitability . We have shown that the propagation of two-state membrane potential transitions plays an active role for self-organizing precise temporal sequences ( Figure 4D ) . As previously mentioned , HiE neurons provide a powerful synaptic drive on other neurons during self-organization . This synaptic input would activate LoE neurons in the temporal order determined by the resting levels of their DOWN states: the lower the resting potential , the slower the activation of that neuron . The resting level depends significantly on the density of H-current in this model . Then , STDP would strengthen or weaken the synaptic connections between LoE neurons according to the relative order of their activations . Ikegaya et al . attributed the generation of precise temporal sequences to the repetition of the membrane-potential fluctuations that reflect synaptic inputs . However , Mokeichev et al . showed that such repeated membrane-potential fluctuations do not appear more often than the chance level expected from their power spectrum structure , thus rejecting the hypothesis by Ikegaya et al . that the repetition of the membrane-potential fluctuations underlies the precise sequences [51] . In our model , the repeated sequences were rarely accompanied by the repetition of the membrane-potential fluctuations ( Figure 3E ) . This was presumably due to the lack of an obvious feed-forward network structure in our model ( note that synaptic connections still play a crucial role in the sequence repetition: see Figure 3C and Figure S3B ) . Thus , our model suggests that the repetition of the membrane-potential fluctuations is not necessarily required for the generation of sequences . This result seems consistent with that of a statistical analysis by Mokeichev et al . However , our results also imply that the repeated membrane-potential fluctuations do not always provide a good statistical measure for the repeated sequences . In a broad range of parameter values , the self-organization with UP and DOWN states led the present recurrent network to repetition of network UP states , in which neuronal activations are broadly synchronized ( Figure 1 ) . Moreover , pruning the weak synapses made the temporal pattern of the network UP states similar to those observed in vivo ( Figure 2D , 2E ) . By contrast , the in vitro cortical activity displayed no obvious synchronous activation patterns [16] . Whether the present model may replicate the in vitro cortical activity is open for further studies . In vivo cortical neurons typically display the two-state membrane potential fluctuations when subjects exhibit a slow-wave oscillation state [4] , [52]–[54] . Boosting the slow oscillations during non-REM sleep improves the ability of declarative memory [55] , and the removal of slow-wave sleep significantly disrupts subject's memory [52] , [53] . The present results suggest that the two-state transitions may assist local cortical networks in encoding temporal sequences to enhance memory consolidation during sleep .
Mathematical details of the model are given in Supporting Information ( Text S1 ) . We modified the two-compartment model of pyramidal neurons which was previously introduced for describing the propagation of spontaneous neuronal activity in cortical networks [48] . Thus , the pyramidal neuron is modeled aswhere Cm = 1 mF/cm2 , As = 0 . 015 mm2 , Ad = 0 . 035 mm2 and gsd = 1 . 75 mS . The somatic compartment involves a voltage-gated sodium current ( INa ) , a delayed-rectifier potassium current ( IDR ) , a leak current ( IL ) , a transient potassium ( IA ) , and a slowly inactivating potassium current ( IKS ) . The dendritic compartment contains a high-threshold calcium ( ICa ) , a calcium dependent potassium ( IKCa ) , a non-inactivating persistent sodium ( INaP ) , and a non-inactivating inward rectifier potassium current ( IAR ) that is activated at hyperpolarization . In experiments , a minority of pyramidal neurons exhibited autonomous membrane potential transitions even after the blockade of excitatory synaptic transmissions . A hyperpolarization-activated cation current ( H current , Ih ) and a persistent sodium current have been suggested as the source of this autonomous activity [17] . We included H-current in the dendritic compartment , as it is abundant at the distal dendritic sites of the neocortical and hippocampal CA1 pyramidal neurons [41]–[46] . H-current plays an active role in generating rhythmic firing of thalamocortical relay neurons [56] and globus pallidus neurons [57] . The voltage-dependent kinetics of this current follows those of a thalamocortical relay neuron model [58] . We set the values of parameters and the kinetics of the various ionic currents as given in [48] , except for the following: the reversal potential of leak current EL = −65 mV; values of the maximum conductance were scaled by 1 . 2 and 0 . 5 for an inward-rectifier and a calcium dependent potassium current , respectively; values of the maximum conductance of Ih were distributed across neurons according to a Gaussian distribution with a mean of 0 . 004 mS/cm2 and a variance of 0 . 02 ( mS/cm2 ) 2 . Cell-dependent strength of the H-current determines different excitability for individual neurons . H-current-rich neurons show spontaneous rhythmic firing at a low rate ( <1 Hz ) even without synaptic input . All pyramidal neurons received a weak background input represented by a Gaussian white noise with a diffusion constant of 3 . 0 mV2/ms . Following [21] , [22] , fast-spiking GABAergic interneurons were modeled as Our network model has 512 two-compartmental pyramidal neurons and 128 inhibitory interneurons and ∼500 , 000 synapses including plastic ones for most of the simulations presented here . Pyramidal-to-pyramidal synaptic input is mediated by AMPA and NMDA receptor-mediated glutamatergic synaptic currents , and the pyramidal-to-interneuron synaptic current is mediated by AMPA receptor-mediated synaptic current . Interneuron-to-pyramidal and interneuron-to-interneuron synaptic transmissions , however , are mediated by GABAA receptors . The AMPA and NMDA glutamatergic synapses are located on the dendritic compartments , whereas the GABAergic synapses are located on the somatic components of pyramidal neurons . The gating variables s ( t ) of the AMPA and GABAA synapses are increased by 1 . 0 at the arrival of a pre-synaptic spike and then decay following a first-order kinetics with a decay time constant of τdec = 5 ms . The NMDA synaptic current obeys a double exponential function with a raising time constant of τraise = 10 ms and a decay time constant of τdec = 50 ms [59] . In addition , the NMDA synaptic current is gated by a voltage-dependent gating variable [59] . Reversal potentials of synaptic currents were set as EAMPA , NMDA = 0 mV , EGABA = 80 mV , and the values of maximum conductance as AMPA synapses at excitatory-to-excitatory connections are modifiable by STDP depending on the relative times Δt between an EPSP and a postsynaptic spike [20] , [23] , [60]:A synapse is strengthened or weakened if the interval from an EPSP to a neighboring postsynaptic action potential is positive or negative , respectively . We employed this additive form of STDP to induce competition between synapses , which has received some experimental support [61] . This rule was chosen to obtain persistent network activity at a moderate firing rate and repeated sequences . However , other study reported a multiplicative STDP rule at cortical synapses [25] . The strength of NMDA synapse between an excitatory neuron pair is rescaled in proportion to that of the AMPA synapse between the neuron pair [62] , [63] . The values of the parameters were set as A+ = 0 . 005 , A− = 0 . 00525 , and τ+ = τ− = 20 ms . The area law ( A+τ+<A−τ− ) induces a competition between the synapses [23] . The pyramidal-to-pyramidal synapses also exhibit short-term synaptic depression [64]–[66] . The mathematical description of depressing synapses follows that given in [66] . Each synapse is multiplied by a depression factor at every presynaptic spike , while each synapse recovers from the depression in the absence of presynaptic spikes . Depression factors are 0 . 99 and recovery time constants were distributed over the excitatory neuron population according to Gaussian distributions with a mean of 700 ms and a variation of 50 ms2 . The connectivity of pyramidal-to-interneuron and interneuron-to-interneuron synaptic connections is 10% , and the interneuron-to-pyramidal synaptic connections have a connectivity of 30% . These types of synaptic connections are not modifiable by STDP . Simulation software was written in C and ran on Pentium 4 3 . 0 GHz×8CPU PCs using parallel computing by the MPI programming techniques . Sequence detection . We marked the UP transitions by the times at which the relative calcium influx defined by Δ[CA2+]/ ( [Ca2+]+ε ) exceeded 2 . 0 ( ε = 0 . 01 µM ) . We used a template matching method to detect the precise sequences repeated more than two times . We selected a reference excitatory neuron , say n1 , and picked up the first UP transition in this neuron at time t1 as a reference event . Then , we searched for all UP transitions that occurred in other neurons after t1 to construct an event vector , [t1 t2 … n1 n2 …] . We constructed similar event vectors taking every UP transition in neuron n1 as the reference event . We compared all possible pairs of thus obtained event vectors as follows . Whenever the same neuron , say nk , appeared in the two vectors , we examined whether the time differences tk−t1 are equal in the two vectors with a precision of ±0 . 5 ms ( criteria 1 ) . If so , we kept the chain n1−nk and repeated the same procedure until we reached the end of the vectors , adding every neuron satisfying criteria 1 to the chain . If the final chain contained more than two neurons ( including n1 itself ) , we regarded this chain as a precise sequence ( criteria 2 ) . We repeated the above procedure for all possible choices of the reference neuron . It is noted that this algorithm might count some sequences in multiple times . However , these redundant counts were negligibly small since the majority of precise sequences contained only three UP transitions in the present simulations ( thus , multiple counting was rejected by criteria 2 ) .
|
Information processing by the brain relies crucially on neuronal circuits . Therefore , clarifying the structure of the brain circuitry is a crucial step towards understanding how the brain processes information . In particular , the cerebral cortex occupies a large portion of the brain in primates and humans , so the organization of local cortical networks is essential for the emergence of higher cognitive functions . However , the complex structure and computations of local cortical networks remain largely unknown . In this study , we investigate the neuronal wiring and activity self-organizing with synaptic plasticity in a model of local cortical networks . Synaptic plasticity describes how synapses between neuron pairs are modified according to activities of the individual pairs . The irregular activity self-organizing in our model surprisingly resembles the spontaneous cortical activity observed during sleep . Moreover , such an autonomous activity contains a diverse repertoire of precisely timed temporal sequences . Whether local cortical networks produce such precise temporal sequences is currently debated in neuroscience . The self-organization of temporal sequences in the sleep-like state suggests that they may play an active role in learning sensory experiences and motor skills , for which sleep is known to be crucial .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"neuroscience",
"neuroscience/theoretical",
"neuroscience"
] |
2008
|
Structure of Spontaneous UP and DOWN Transitions Self-Organizing in a Cortical Network Model
|
In tsetse flies , nutrients for intrauterine larval development are synthesized by the modified accessory gland ( milk gland ) and provided in mother's milk during lactation . Interference with at least two milk proteins has been shown to extend larval development and reduce fecundity . The goal of this study was to perform a comprehensive characterization of tsetse milk proteins using lactation-specific transcriptome/milk proteome analyses and to define functional role ( s ) for the milk proteins during lactation . Differential analysis of RNA-seq data from lactating and dry ( non-lactating ) females revealed enrichment of transcripts coding for protein synthesis machinery , lipid metabolism and secretory proteins during lactation . Among the genes induced during lactation were those encoding the previously identified milk proteins ( milk gland proteins 1–3 , transferrin and acid sphingomyelinase 1 ) and seven new genes ( mgp4–10 ) . The genes encoding mgp2–10 are organized on a 40 kb syntenic block in the tsetse genome , have similar exon-intron arrangements , and share regions of amino acid sequence similarity . Expression of mgp2–10 is female-specific and high during milk secretion . While knockdown of a single mgp failed to reduce fecundity , simultaneous knockdown of multiple variants reduced milk protein levels and lowered fecundity . The genomic localization , gene structure similarities , and functional redundancy of MGP2–10 suggest that they constitute a novel highly divergent protein family . Our data indicates that MGP2–10 function both as the primary amino acid resource for the developing larva and in the maintenance of milk homeostasis , similar to the function of the mammalian casein family of milk proteins . This study underscores the dynamic nature of the lactation cycle and identifies a novel family of lactation-specific proteins , unique to Glossina sp . , that are essential to larval development . The specificity of MGP2–10 to tsetse and their critical role during lactation suggests that these proteins may be an excellent target for tsetse-specific population control approaches .
Tsetse reproductive biology is unusual among insects . Female tsetse give birth to a fully mature third instar larva ( viviparity ) after an extended intrauterine gestation . This reproductive strategy limits the capacity of tsetse mothers to only 8–10 offspring per lifetime [1] . To accommodate intrauterine larval development , the morphology and physiology of the female tsetse reproductive organs have undergone extensive modification . The reproductive tract has been expanded into a uterus to serve as a safe harbor for developing larvae . Ovarian development alternates between the right and left ovaries to produce a single oocyte during each gonotrophic cycle . The female accessory gland has been modified and expanded to provide milk that is secreted into the uterus and consumed by the developing larva [1] . The distinctive aspects of tsetse viviparity represent significant reproductive bottlenecks that could be exploited for population control . Furthermore , identification of factors specific to milk production could lead to development of novel tsetse-specific compounds that interfere with larval development and induce abortion ( abortifacients ) without impacting non-target insects . The nutritional components of tsetse milk consist mainly of proteins and lipids emulsified in an aqueous base [2] . In total , 6–10 mg of nutrients ( combined with 10 mg of water ) are transferred to the larva in the milk during intrauterine development . Few studies have examined regulation of tsetse milk production , including an investigation of structural changes in the milk gland , radioisotope studies of nutrient movement within the mother during lactation , and direct examination of specific milk proteins [1] , [3]–[9] . To date , six milk proteins have been characterized , including Transferrin [7] , [10] , a lipocalin ( Milk Gland Protein1 , MGP1 [6] , [11] ) , two unknown milk proteins ( MGP2–3; [12] ) , Acid Sphingomyelinase 1 ( aSMase1; [9] ) and Peptidoglycan Recognition Protein-LB ( PGRP-LB , [13] ) . Furthermore , we recently showed that lipid metabolism is governed by the cooperative activity of insulin and juvenile hormone signaling pathways during the pregnancy cycle [14] . However , the full suite of proteins present in the milk and underlying mechanisms for their regulation during tsetse lactation and pregnancy have yet to be determined . In this study , a satellite paper to our report on the whole genome sequence of the tsetse species Glossina morsitans morsitans [15] , we used differential RNA-seq analyses to compare transcript abundance in females carrying an intrauterine larva ( lactating ) with females 24–48 hours after parturition ( non-lactating or dry ) . The lactation period occurs during larvigenesis , while the dry period occurs over the course of oogenesis and embryogenesis [14] . In addition to transcriptome analysis , protein constituents of tsetse milk were identified through LC/MS/MS analyses of gut contents from nursing larvae . We describe the expression profile , the predicted structure based on an in silico approach , and the microsyntenic genomic organization of nine tsetse-specific milk proteins ( MGP2–10 ) that we propose represent a highly divergent , novel protein family . siRNA-based knockdown analysis was employed to examine the functional roles of the MGP2–10 proteins during tsetse reproduction . Since MGP2–10 are tsetse-specific and have substantial influence over tsetse fecundity , we discuss their potential for exploitation in novel population reduction approaches . Lastly , we discuss our findings in light of milk secretions described from other lactating organisms .
To understand the major products of lactation and factors that may be responsible for regulating their expression , we analyzed two RNA-seq libraries . The first library represents lactating females carrying an early third instar intrauterine larva , while the second represents dry females collected approximately 48 hours post-parturition , at which time they had an early embryo developing in the uterus and lactation has yet to commence . In total , over 42 million reads from each of the two libraries were recovered ( Table 1 ) . Overall read quality was high for both sample sets based on FastQC analysis . Removal of contaminating tsetse symbiont ( Wigglesworthia , Sodalis and Wolbachia ) specific sequences and cleanup resulted in a 4% reduction in the total number of sequences identified in lactating flies and a 5 . 1% reduction for dry flies , respectively ( Table 1 ) . De novo assembly of the two datasets by Abyss [16] , [17] and Trinity [18] generated 42 , 935 contigs that were subsequently identified according to multiple search parameters ( Table S2 ) . There were 34 , 674 contigs at least 200 bp in length , with the longest contig measuring 24 , 573 bp in size ( Fig . S1 ) . Distribution of reads per contig was comparable between the two datasets with the exception that there was a greater number of highly expressed genes in lactating flies ( Fig . S2 ) . Comparative analyses of contigs revealed that most were more highly expressed in dry flies compared to their lactating counterparts . A total of 297 contigs ( 2 . 1% ) with at least 50 mapped transcriptome reads showed elevated expression in lactating versus dry flies ( Fig . 1a; Table S2 ) . Only 1311 were expressed at statistically different levels between the two datasets , with 48 contigs ( 4 . 2% ) more highly expressed in lactating flies ( Fig . 1b; Table S3 ) . Classification of the lactation-expressed contigs based on specific metabolic and structural functions revealed enrichment for lipid metabolism , transport and storage , protein synthesis , secreted proteins , and those of unknown function ( Fig . 2a ) . Examination of the most highly expressed contigs in lactating flies revealed known milk protein genes along with novel transcripts not previously associated with lactation . The known milk protein genes ( mgp1 , mgp2–3 , tsf and asmase1 ) were expressed at least 10-fold higher in lactating versus dry flies ( Fig . 3a–c; Table S2 , S3 ) . Of particular interest was the discovery of a group of seven new genes similar to the previously identified mgp2–3 genes that were upregulated during pregnancy ( mgp4–10; Fig . 3b; Table S2 ) . Transferrin and aSMase1 , involved in iron transport and sphingolipid metabolism respectively , were the only other proteins that were highly expressed and showed increased transcript abundance in lactating flies ( Fig . 3c; Table S2 ) . Recently , the immunoregulatory Peptidoglycan Recognition Protein LB ( PGRP-LB ) was also detected in tsetse milk [13] . Based on this analysis , PGRP-LB expression patterns are different from those of other lactation associated-proteins , as its expression did not increase throughout pregnancy ( Table S2 ) . In addition to the genes described above , specific ribosomal RNAs were significantly elevated in lactating flies ( Fig . 3d; Fig . S3 ) , and may account for the overall increase in contigs coding for genes involved in protein synthesis ( Table S2 ) . Confirmation of transcript abundance during lactation was achieved by qPCR analysis of the mgp1–10 , 28S , transferrin , pgrp-lb and asmase1 genes ( Table 2; Text S1 ) . The majority of the contigs ( 94% ) were more abundant in dry ( non-lactating ) compared to lactating flies ( Fig . 1a; Table S2 ) . Multiple gene families were highly expressed in dry flies ( Fig . 2a , Table S2 ) . Contigs encoding heat shock proteins and antioxidant enzymes were increased in dry flies , indicating that dry flies may be better suited than their lactating counterparts to respond to stress and environmental insult ( Fig . S4 ) . In particular , qPCR analysis validated the transcriptome data for Cu/Zn superoxide dismutase and catalase , which encode proteins that remove reactive oxygen species to prevent damage ( Table 2 ) . Lipid metabolism contigs were more abundant in dry flies with the exception of Brummer lipase , which was only two-fold higher than in lactating flies ( Fig . S4; Table S2 ) . Expression of tsetse yolk proteins , yolk protein 1–3 ( yp1–3 ) was also higher in dry flies , reflecting the yolk protein synthesis that occurs between bouts of lactation ( Fig . S4; Table S2 ) . Contigs identified as trypsin showed decreased transcript abundance in lactating flies ( Table S4 ) . Given that the transcriptome analysis was from whole females , this finding likely correlates with lactating females' smaller bloodmeals that result from limitations on abdominal space imposed by developing intrauterine larva [5] . These results suggest that many processes in tsetse mothers are down regulated during lactation ( = higher in dry flies ) , when the female devotes energy and resources to synthesize milk-associated proteins to nourish the developing larva . We conducted a secondary RNA-seq analysis after removing reads that mapped directly to the twelve most abundant lactation-specific genes ( asmase1 , mgp1–10 and transferrin ) . Removing these reads yielded only a 2 . 6% reduction in the dry fly dataset , but a drastic reduction of 47 . 2% was observed in the lactating fly dataset ( Fig . 4a ) . This difference suggests that lactating flies invest over 47% of their total transcriptional activity toward producing the main protein constituents of tsetse milk ( Fig . 4a ) . Each milk-specific gene accounted for 1 . 4–6 . 7% of the total read count in lactating flies , with the most reads mapping to mgp10 and transferrin ( Fig . 4a ) . This removal resulted in a total of 2238 genes that were more highly expressed in lactating flies , but with only 151 that were significantly elevated relative to dry flies ( Fig . 4b , c ) . Assignment by metabolic category resulted in a more balanced distribution of highly expressed contigs in lactating and dry flies ( Fig . 5a ) . This second analysis revealed a few additional contigs whose expression increased during lactation; their expression was previously overshadowed by highly expressed milk-specific genes ( Table S5; Table S6 ) . These included dawdle ( an activin signaling molecule ) , glyoxylate/hydroxypyruvate reductase ( an enzyme that converts glycerate to hydroxypyruvate ) , choline-phosphate cytidylyltransferase ( an enzyme in the Kennedy pathway that catalyzes choline phosphate to CDP-choline ) and multiple ribosomal proteins ( Fig . 5b , Tables S5 , Table S6 ) . Using LC/MS/MS analyses on the gut contents of feeding larva , we identified 155 proteins that may be constituents of tsetse milk . Most of these proteins have a low exponentially modified protein abundance index ( empai ) value and are likely present in milk at low levels or may be products from the larval gut ( Table S7 ) . Most of the highly abundant proteins identified in tsetse milk were products of genes identified as highly expressed during lactation in our transcriptomics analysis , including MGP1–10 , Transferrin and aSMase1 ( Table 2 ) . Previously , PGRP was documented in tsetse milk [13] and we confirmed the presence of this immune protein in the milk proteome ( Table 2 ) . In addition to the transcriptionally-abundant proteins , the milk proteome identified three other abundant proteins ( empai >1 . 2; Table 2 ) . These three proteins include a sterol binding protein ( Niemann-Pick C-2g , NPC2G ) , Ubiquitin Associated and SH3 Domain Containing A ( UBASH3A , a protein belonging to the T-cell ubiquitin ligand , TULA , family [19] ) , and a putative tsetse protein with unknown function ( GmfB8 ) . Transcript levels for NPC2G , GmfB8-like protein and UBASH3A were measured in the milk gland/fat body fraction during and after lactation and in the larval gut to confirm whether these are generated by the milk gland or if they are products of the gut ( Fig . 6 ) . Both npc2g and gmfb8-like protein expression were detected at high levels in the larval gut . Transcript level for UBASH3A was higher in the milk gland/fat body , showing an expression profile similar to PGRP-LB ( Fig . 6 ) , suggesting that this is likely a low abundance protein generated by the milk gland during lactation . These results provided further validation for our transcriptome-based identification of milk protein genes as actual secreted products in tsetse milk . In addition , we recovered potential milk proteins that are not under extensive transcriptional regulation during lactation . BLASTx searches of the NCBI nucleotide collection failed to recover orthologous sequences to the MGP2–10 from any organism . Partial gene sequences encoding MGPs were identified from two other tsetse species , Glossina fuscipes fuscipes ( MGP2 , 5 ) and Glossina pallidipes ( MGP3 , 4 ) , using RT-PCR with degenerate primers ( Fig . S5 , S6 ) . Mining of sequence data from recent EST projects on the flesh fly , Sarcophaga crassipalpis [20] , [21] , revealed one sequence with marginal sequence similarity with the tsetse MGP2–10 ( Fig . S5 , S6 ) . The average number of amino acids for MGP2–10 was 179 . 3 ( range 170–191 , Table 3 ) with a predicted molecular weight of 21 . 4 kD ( range 20 . 4–22 . 4 kD; Table 3 ) . The average isoelectric point for MGP2–10 was 6 . 2 ( range 5 . 9–6 . 5; Table 3 ) . The aliphatic index , or the relative volume of a protein occupied by aliphatic side chains ( alanine , valine , isoleucine , and leucine ) that is indicative of the stability of globular proteins [22] , is moderate for MGP2–10 , ranging from 64–86 ( Table 2 ) . There are no predicted glycosylation sites on MGP2–10 , but there are at least four predicted phosphorylation sites for each MGP ( Table 3 ) . Amino acid alignments of MGP2–10 identified a conserved secretory peptide and three conserved regions with 68–100% and 12–100% nucleotide and amino acid similarity , respectively ( Fig . 7a , b; Fig . S5 ) . Phylogenetic analysis shows MGP2 and MGP4 as recently duplicated paralogs sharing 92% amino acid similarity ( Fig . 8a; Fig . S6 a , b ) . Mapping of mgp2–10 coding sequences to genomic scaffolds revealed that these genes localize to a 40 kb microsyntenic region ( Fig . 8a ) . The phylogeny for mgp2–10 splits these genes into two distinct groups , one consisting of mgp2 , 4 , 5 , 6 , 9 , 10 and the other of mgp3 , 7 , 8 . When the phylogeny is mapped against the genome location of mgp2–10 , mgp2 , 4 , 5 , 6 , 9 , 10 localized with a region surrounded by mgp3 , 7 , 8 ( Fig . 8a ) . All mgp genes share a conserved exon-intron structure ( Fig . 8a ) , despite showing varying levels of amino acid sequence similarity amongst them ( Fig . 8b; Fig . S5 ) . Our results indicate that MGP2–10 proteins are likely specific to Glossina , but it remains to be seen if evolutionarily-related sequences may exist in other closely-related viviparous genera ( i . e . bat flies and sheep ked; data not available for these species ) . The sequence obtained from the flesh fly may represent either a class of proteins distinct from the tsetse MGP family , or could be highly divergent ancestral sequence to MGP2–10 genes found in tsetse . Based on amino acid composition , the novel tsetse milk proteins provide all essential amino acids necessary for larval growth and development ( Table S8 ) . Protein structure predictions for the novel MGPs were generated by four individual programs . Structural predictions were ab initio as no homologous protein structures were available . The ab initio structure predictions from the four programs revealed that MGPs usually form multiple α-helices ( 6–10 per protein ) . However , no functional insights were provided by the I-TASSER program , since MGPs lack structural similarity to other characterized proteins ( Fig . S7 ) . Of particular interest , these proteins contain high percentage of hydrophobic amino acids ( this study , [12] ) , including the hydrophobic secretory peptide that was identified in tsetse milk , indicating that this region is not always cleaved during secretion . Examination of amino acid alignments identified several regions of moderate conservation across MGP2–10 from G . morsitans . To assess the relative selective pressures acting on these paralogous genes , we performed several computational analyses of nonsynonymous-to-synonymous substitution ratios ( dN/dS ) along the coding sequences for these genes . This type of analysis is usually conducted on orthologous genes in different species or on multiple alleles within a species , but utilization of this on paralogous genes could provide insight into regions critical to their function . When dN/dS substantially exceeds 1 , evidence for positive selection ( = adaptive evolution ) is inferred . In contrast , dN/dS = 1 implies neutral evolution , while dN/dS values closer to zero provide evidence for negative or purifying selection . Sequences were translated and multiple alignments were performed in ClustalX [23] , followed by optimization in BioEdit [24] or MEGA 4/5 [25] , [26] . We reverse-translated amino acid sequences to obtain codon alignments as input sequences for dN/dS analyses under both PARRIS [27] and FEL ( Fixed Effects Likelihood; [28] ) analyses in DataMonkey ( www . DataMonkey . org; [29] , [30] ) , a web-based implementation of the HyPhy algorithm [31] . PARRIS allows detection of positive selection across an entire coding sequence , while the FEL method is suitable for detecting positive or negative selection in a site-specific manner in small ( 10–15 sequences ) datasets [28] . Under the PARRIS algorithm , we found no evidence for positive selection across the coding sequence of MGP2–10 , suggesting that no residues in these proteins are targets of adaptive evolution . FEL analysis likewise showed no evidence for individual codons subject to positive selection . In contrast , while the preponderance of residues in MGP2–10 are apparently undergoing neutral evolution , FEL analysis indicates that the identified N-terminal secretory signal peptide is largely subject to purifying selection ( Fig . 7a , b ) , suggesting that this region is indispensible for protein function or appropriate intra/intercellular transport . A minority of residues , largely dispersed throughout the C-terminal half of MGP2–10 are additionally subject to negative selection as evidenced by dN/dS ratios significantly less than 1 ( p = 0 . 05 , Fig . 7a , b ) . A role for amino acids under purifying selection outside of the secretory region is unknown . A majority of these conserved sites are proline ( 33 . 3% ) and phenylalanine ( 26 . 6% ) residues , suggesting these amino acids may be critical for MGP2–10 folding and/or function . Previous examination of mammalian milk proteins revealed that discrete , specific sections of each gene are subject to neutral , negative or positive selection [32] . Using a similar MEGA-based analysis to specifically investigate MGP2–10 from G . morsitans , the secretory peptide and conserved region 1 appear to be largely under negative selection ( Fig . 7b ) . The first variable region has a high dN/dS and is likely subject to neutral or positive selection ( Fig . 7b ) , but additional MGP sequences need to be recovered from other Glossina sp . to more confidently determine site- or region-specific selective pressure across MGP coding sequences . Overall , these analyses indicate that only the secretory peptide and the first conserved region are likely subject to purifying selection , but additional analysis will be necessary once full-length MGP genes are recovered from other members of Glossina to establish regions of selection . Using RT-PCR analysis to determine the tissue specificity of MGP expression , we found that expression of mgp2–10 is specific to the female fat body/milk gland ( Fig . 9a ) . Temporal expression profiles obtained for these genes showed that mgp2–10 transcripts increase dramatically during larvigenesis and then rapidly decline within 24–48 h following parturition ( Fig . 9b ) . This temporal and spatial expression profile is consistent with other characterized milk proteins , including mgp1 ( Fig . 9a , [6] ) , asmase1 [9] and transferrin [7] . The temporal expression profiles for the MGP genes we identified from the two other tsetse species ( gpmgp3 , 4 from G . pallidipes and gfmgp2 , 5 from G . fuscipes ) were similar to those observed for mgp2–10 in G . morsitans . Transcript abundance was lower in teneral females and in females with developing intrauterine embryos , becoming progressively greater through larvigenesis ( Fig . S8 ) . In contrast , the MGP-like sequence discovered from the flesh fly , another brachyceran that exhibits larviposition but does not nourish the developing larva , was not expressed in this manner; we observed no differences in MGP-like gene expression between male and female flesh flies ( nonpregnant vs . pregnant minus larval expression level; Fig . S8 ) . Thus , even though a gene with moderate sequence similarity was identified in S . crassipalpis , its expression profile is incongruent with tsetse MGPs . Injection of siRNAs targeting mgp5 , 7–9 significantly reduced corresponding target transcripts in lactating flies ( Fig . 10a ) . Differences in the knockdown efficiency are likely due to the combined effects of technical variation and slight natural variation in the pregnancy cycle . Suppression of individual transcripts ( even the highly expressed mgp7 ) had no effect on the number of pupae produced per female , the length of pregnancy , or the incidence of pupal emergence ( Fig . 10b–d ) . This suggests functional redundancy among the multiple mgp paralogs , which are all expressed in a similar spatio-temporal manner during pregnancy . Simultaneous suppression of two MGPs ( i . e . 5 and 7 ) reduced the number of pupae deposited per female by 10–15% and extended the duration of pregnancy by 2–3 d , but no difference was observed in the incidence of adult eclosion ( Fig . 10b–d ) . When mgp5 , 7–9 were co-suppressed , fecundity was reduced by nearly 70% and in cases where mothers produced viable progeny , pregnancy was extended by 6–8 d ( Fig . 10b–d ) . Together , these results suggest that the paralogous mgps share a critical function in tsetse reproduction . Bradford assay of the milk protein content indicated that simultaneous knockdown of mgp5 , 7–9 reduced overall milk protein by nearly 22% ( 0 . 43±0 . 06 mg protein/5 µl milk ) compared to siGFP treated controls ( 0 . 56±0 . 04 mg protein/5 µl milk ) . We hypothesized that these proteins may serve a function in maintaining milk lipid stability . To explore this possibility , we assessed the stability of milk emulsification after MGP knockdown utilizing an emulsion stabilization assay . Knockdown of mgp5 , 7–9 resulted in an increased rate of separation of the aqueous and lipid fractions of the milk by over two-fold ( Fig . 10d , e ) . These findings suggest that the novel MGPs are not only an important amino acid/protein resource for the developing larva , but function to stabilize lipids within tsetse milk , allowing fat to remain homogenously distributed .
Several of our major findings from this transcriptome analysis included evidence for a substantial shift during tsetse milk production toward secreted proteins , genes involved in lipid metabolism , transporter genes , and specific genes coding for protein synthesis machinery . Structural studies have previously shown that there are extensive arrays of endoplasmic reticulum ( ER ) that develop in actively secreting milk gland cells , and then degenerate during an involution period following parturition [33] . A similar pattern of increased ER production in the milk gland was documented in another tsetse fly species , Glossina austeni [34] and in Melophagus ovinus , a viviparous sheep ked [35] . The rate of protein synthesis is elevated in the milk gland during periods of lactation , in accordance with an increase in ER [1] , [36]–[38] . Increase of milk gland associated ER to allow for production of lactation-associated proteins is likely the reason for high abundance of transcripts for genes involved in protein synthesis during tsetse milk production . The products of mgp1–10 likely constitutes over 95% of the protein content in tsetse milk ( this study , [6] , [12] ) , accounting for the increase in contigs for secreted proteins . Transferrin and aSMase1 account for the high read abundance of contigs associated with transporters and lipid metabolism , respectively . The extreme elevation of asmase1 , mgp1–10 and transferrin transcript levels during lactation indicate that the expression of these 12 genes which constitutes less than 0 . 0005% of the contigs from the de novo library , represents over 45% of the RNA-seq library . In contrast , in dry flies , these same gene transcripts represent less than 2 . 6% of the library . A heavy investment in specific genes during reproduction is not uncommon [39]–[43] , but in most studies the effect was measured directly within a specific organ , rather than the entire organism . As an example , the lactating wallaby invests over 50% of transcript abundance in the mammary gland to the production of protein bound for transfer in the milk [39] . In addition , most milk production transcripts , specifically those directly incorporated into milk , show drastic changes throughout lactation in mice and bovine within the mammary gland [44]–[46] . For tsetse , heavy transcript investment according to total Illumina read levels is based on the entire female fly rather than the isolated milk gland organ . This high investment is not surprising since at least 4–5 mg ( 20–25% of the total mass of the mother ) of proteins are secreted by the milk gland during the 4–5 day period of lactation , representing 40–50% of the nutritional content of the milk [2] . Thus , female tsetse may be uniquely adapted to generate milk with its entire resources devoted to the transcription of milk proteins at the expense of other biological processes during lactation . Increases in transferrin , asmase1 and mgp1 expression during lactation were expected since these proteins are recognized as major constituents of tsetse milk [7] . The role for Transferrin in tsetse milk has yet to be determined [7] , though speculation suggests transferrin may serve as a source of iron as well as for immune development/protection [7] . Regarding other milk proteins , knockdown of asmase1 in lactating females reduces fecundity and severely impacts progeny fitness [9] . Biochemical studies have revealed that secreted aSMase1 is inactive and conversion to the biologically active form , which allows sphingomyelin digestion , occurs upon encountering the acidic conditions of the larval gut [9] . As a lipocalin , MGP1 likely carries a critical unknown hydrophobic ligand in the milk [6] , and has been documented to be critical for tsetse fecundity [6] . Lipids , specifically triacylglycerides , constitute the other major nutritional components in tsetse milk [2] , [47] . Recent studies have shown that Brummer ( Bmm ) lipase- and adipokinetic hormone ( AKH ) -mediated lipolysis are both critical for mobilizing lipids during tsetse reproduction [47] . Our transcriptome data indicate that only a single lipase , bmm , is increased during lactation , while expression of most other lipid metabolism genes are suppressed or expressed at levels equivalent to those seen in dry flies . Such minimal transcript variation is perhaps not surprising in light of recent studies on insect lipolysis , which reveal that most control occurs at the post-translational level , either through insulin signaling or through other factors that interact with the surface of the lipid droplet [48]–[51] . An argument for post-translational regulation is further supported by our recent study showing that insulin and juvenile hormone signaling both act to coordinate lipid metabolism in tsetse mothers through transcriptional regulation of select lipolytic/lipogenic genes including midway and bmm , while other such genes associated the lipid metabolism show little variation [14] . Further , a reasonable explanation for the lack of an observed increase in lipolysis genes is that these genes typically increase prior to lactation ( late embryogenesis/early larvigenesis [14] ) , while our lactating sample was collected during the peak of lactation occurring during the last stages of larvigenesis . In addition , perilipin1 and perilipin2 transcripts are elevated in dry versus lactating flies and perhaps these proteins , which interact with lipid droplets , are necessary to accommodate the drastic increases in fat body volume that occurs during the involution periods that separate lactation cycles . In general , our results regarding expression of lipid metabolism genes support our previous studies that bmm-mediated lipolysis plays a critical role in regulating lipid homeostasis during pregnancy [14] , [47] . Removal of reads mapping to the 12 abundant milk protein genes during RNA-seq analysis allowed for the identification of three additional genes that were enriched and highly expressed during milk production . Dawdle is an activin signaling molecule that has been linked to synaptic growth at the neuromuscular junction [52] and immunity [53] in Drosophila . As a member of transforming growth factor beta ( TGF β ) superfamily of growth factors , activin may be signaling growth of a specific tissue , possibly the milk gland , during lactation . In addition to the role in Drosophila , activin is key in regulating growth of the mammary gland during lactation in multiple mammals and has a critical role in breast cancer [54]–[56] . The increased levels of glyoxylate/hydroxypyruvate reductase , grhpr , may provide additional substrates to maintain homeostasis of proline , the main nutrient source in tsetse hemolymph [57]–[59] , as milk production requires a massive amino acid investment [1] . Finally , expression of choline-phosphate cytidylyltransferase , cct , has been linked to changes in lipid droplet size [60] , and this enzyme may be playing a role either in the fat body during the rapid lipolysis associated with tsetse lactation [47] , or in the generation of fat globules for incorporation into the milk . Alternatively or in combination , CCT could be critical for the allocation of choline and choline-derivatives into milk during lactation . The provision of choline is essential for proper organismal growth and development [61] , [62] . Our transcriptome data revealed that the majority of genes are expressed at higher levels in dry versus lactating flies . This difference is likely due to the fact that transcript levels for most genes are reduced in lactating flies at the expense of generating lactation-specific proteins . In dry flies , transcript elevation for genes associated with digestive processes likely corresponds to the increased bloodmeal size in flies not harboring an intrauterine larva [1] . Elevated transcripts for genes coding for heat shock proteins suggest that dry flies may be better suited for stress tolerance than their lactating counterparts . In addition , proteins involved in oocyte development are elevated in dry flies , likely since oocyte development is nearly complete before lactation begins [1] . Thus , the transcript profile diversity in dry flies is more robust , featuring a more global/representative expression of genes , compared to the rather specific gene set expressed in lactating flies . Recent studies focusing on MGP2 and MGP3 failed to verify their transfer to the nursing larvae since antisera were not available [12] . The proteomic analysis performed here confirms that these highly expressed genes synthesize milk proteins that are indeed transferred to the intrauterine larva . The proteomics data also confirm that Transferrin , MGP2–10 , and aSMase1 are the primary protein components of the tsetse milk [6] , [7] , [9] . In addition our proteomic analysis also identified UBASH3A as a component of the tsetse milk . UBASH3A is a member of the TULA protein family and contains ubiquitin-associated ( UBA ) and Src-homology 3 ( SH3 ) domains along with a histidine phosphatase domain [19] , [63] , [64] . A potent regulator of cellular function documented in most metazoan species [19] , [64] , UBASH3A is critical for regulation of T-cell proliferation and other aspects of the mammalian immune response , specifically for suppressing immune cell proliferation . The role of insect UBASH3A has not yet been determined but the presence of UBASH3A in tsetse milk suggests that it may play a role in modulating the immune system of the mother or progeny to allow intrauterine larval development . Along with potentially modulating mother-offspring immune relationship , tsetse's milk secretions also provide a route for the transmission of tsetse's microbial symbionts ( Wigglesworthia and Sodalis , [65] , [66] ) and host immune responses may need to be regulated for symbiotic homeostasis . Our prior studies had shown that the presence of PGRP-LB in the milk is critical for symbiont transfer and overall offspring fitness [13] and the presence of UBASH3A may play a similar role in host-symbiont dynamics during the bacteria transfer within the milk . The ability to transfer symbionts to allow for maintenance of the microbiome in the offspring has been documented to be critical for tsetse immune maturation [67] and the development of the peritrophic matrix development [68] . Many other proteins were observed at lower levels; these low abundance proteins may be critical for larval development . Due to the recovery of milk from within the larval gut contents , we cannot rule out the possibility that these proteins could be products of the larval alimentary canal . Studies devoted to each low abundance peptide will be necessary to determine if it is a product of tsetse milk . We identified seven new milk gland proteins , MGP4–10 , that are similar to MGP2–3 . MGP2–10 each contains a conserved secretory signal and multiple sites throughout three moderately-to-highly conserved regions with several residues under apparent strong purifying selection . Structural analysis of these MGPs failed to provide functional insights , but did reveal that these proteins are likely globular , consisting of multiple α-helices . Further study is necessary to conclusively determine the structures of these novel proteins . The coordinated high expression levels observed for mgp2–10 during lactation and reduced expression after parturition indicate that these proteins are under similar transcriptional regulation and that they may also serve as a source of proteins for larval nutrition [this study , 12] . Indeed , milk protein content was reduced by 20–25% when mgp5 , 7–9 transcripts were suppressed by 60–70% , suggesting that MGPs , based on total transcript abundance , likely account for 70–75% of the total protein content of tsetse milk . The MGP2–10 proteins also contain all amino acids , supporting the notion that they function as a complete protein resource for the developing larva . Furthermore , multiple phosphorylation sites associated with each protein suggests that MGP2–10 may also serve as a source of phosphate in tsetse milk . The lack of predicted glycosylation sites on MGP2–10 is not surprising since carbohydrate levels are extremely low in tsetse milk [2] . Previous studies have shown that low molecular weight proteins interact with lipids in tsetse milk [2] . This prompted us to investigate a potential role of MGPs for stabilization of milk-borne lipids . Here , we show that RNA interference of mgp7–9 results in acceleration of lipid separation from the aqueous phase of milk . This suggests that MGP7–9 ( and likely the other MGPs ) may represent the previously-documented , unidentified low molecular weight proteins associated with tsetse milk lipids [2] . MGP2–10 have a high proportion of hydrophobic amino acids [this study , 12] , which may enable these proteins to interact with milk lipids . Thus , it appears that these newly-identified proteins are critical for maintenance of proper lipid/water dynamics in tsetse milk . Similarities among MGP2–10 suggest that these proteins represent a highly divergent lactation-specific protein family from tsetse flies . These genes are localized to a single 40 kb chromosomal loci , have similar gene structures and their phylogeny correlates with their chromosomal organization indicating that mgp2–10 may have expanded by multiple gene duplication events from a common ancestor . It is possible that ancestral duplication events yielded two separate groups which may have been subsequently expanded as a result of unequal genetic crossing-over with the mgp2 , 4 , 5 , 6 , 9 , 10 being encoded on the antisense strand . Predicted three-dimensional structures between MGP2–10 is similar including multiple α-helices and a globular protein tertiary arrangement . mgp2–10 are under nearly identical transcriptional regulation showing increased expression during tsetse fly lactation and rapid decline during involution . These proteins also exhibit functional redundancy as a source of secreted amino acids in the milk and in sustaining lipid-protein homeostasis within the aqueous milk base . Although MGP2–10 have varying levels of amino acid similarities ( 18–91% ) , there are conserved regions they share outside of the secretory peptide . Specifically , 23 sites are under purifying selection ( 8 in the secretory peptide coding sequence and 15 dispersed throughout the remaining portions of the sequence ) , and these are likely critical to the functional role of MGP2–10 during tsetse lactation . Collectively , similarities between MGP2–10 indicate that these proteins constitute a novel family in tsetse similar to other highly divergent protein families , including caseins [69] , [70] , aquaporins/major intrinsic proteins [71] , [72] , odorant binding proteins [73] , [74] and small heat shock proteins [75] . Our previous work demonstrated that several mechanisms underlying tsetse lactation parallel characteristics of mammalian lactation . First , both systems have highly specialized lactating cells that cycle through periods of high productivity during lactation to low activity following involution [76] , [77] . Second , there are multiple , functionally analogous proteins involved in tsetse and mammalian lactation [78] , [79] . These proteins include a lipocalin ( MGP1 vs . β-lactoglobulin [6] , [38] , [45] , [80] ) , an iron-transfer protein ( Transferrin vs . Lactoferrin [7] , [10] ) , SMase in milk or the gut contents of feeding progeny [9] , [81]–[83] and various immunity proteins ( PGRP and UBASH3A vs . multiple mammalian immunity proteins , [this study] , [ 45] , [78] , [84 , 85] . Third , the lipid content transferred to the developing offspring is similar during lactation in both systems . Fourth and finally , microbial symbionts are transferred from the mother to the developing offspring in both tsetse and mammals [66] , [86]–[88] . There are however a few noteworthy differences between tsetse and mammalian lactation , such as the abundance of calcium transport proteins in mammalian not found in tsetse milk [this study] , [76] , [79 , 89] . This difference is unsurprising , since insects do not require large amounts of calcium for their chitin-based exoskeleton . In addition , tsetse milk contains a lower carbohydrate content than mammalian milk [76] , [90] , indicating that tsetse flies rely solely on lipids and protein for growth and development , rather than a combination of sugar/lipids/protein as in the mammalian case . Such reduced reliance on sugar is also unsurprising as tsetse flies have little to no detectable levels of glucose within their bodies and use proline as their circulating hemolymph resource , rather than a glucose-based substrate such as trehalose [1] , [57] . Mammalian genomes contain no orthologous sequences to the nine novel tsetse MGPs . However , MGPs might function analogously to caseins in mammalian milk . Caseins are the major amino acid and calcium source for the mammalian neonate [65] , [70] , [91] . While MGPs do not carry calcium , they do , like caseins , represent a major amino acid resource in the milk [39] , [46] , [69] , [92] . The presence of multiple phosphorlyation sites in MGPs suggests that this novel protein family may also act in tsetse milk as a source of phosphate as do caseins in mammalian milk [69] , [70] . Furthermore , caseins are amphipathic molecules that form micelles , which interact directly with lipids both in vivo and in vitro [69] , [93] . According to our results , MGPs likewise interact with lipids to promote stability of lipid emulsions in the aqueous tsetse milk . To determine if MGP2–10 have amphipathic structural properties like caseins , direct protein structural studies , rather than protein modeling , will be necessary . In addition , expansion of the casein and MGP gene families has occurred for both mammals and tsetse within specialized regions of their genomes . This indicates that expansion of these protein families ( MGPs and caseins ) is advantageous for provisioning the necessary nutrients in both tsetse and mammalian milk , respectively [this study] , [ 69 , 92] . Finally , members of the MGP and casein families show substantial divergence in sequence similarity [this study] , [69 , 78] , which is a characteristic of proteins that are mainly nutritional components of milk . Proteins involved in mechanics of lactation , i . e . milk fat globule formation or have an enzymatic function , are typically more conserved within and between organisms [78] . These similarities further support the idea that MGPs perform an analogous role to mammalian caseins in tsetse milk . Few studies have examined the effects of casein knockdown/knockout in mammals . In mice , knockout lines have been developed for α- , β- and κ-casein [94]–[96] , and in goats there are naturally occurring deficiencies in α-casein [97] . Knockdown phenotypes differ dramatically , depending on the casein variant targeted . The knockout mutant for β-casein in mice [95] and null αS1-casein in goats [97] have no or minimal apparent effects on milk production , potentially due to increased expression of other casein genes to compensate for the loss of β-casein or αS1-casein , respectively . Offspring receiving milk from α-casein-null mothers experience delayed growth and life-long body size reduction , but only transient effects on physical and behavioral development [96] . The most drastic change is noted in κ-casein null mice , which fail to lactate [94] . Similar to suppression of caseins , knockdown of individual tsetse MGPs had only minimal effects on tsetse fecundity; more drastic changes occurred upon silencing multiple transcripts . In addition , a reduction in tsetse MGPs accelerated separation of lipid emulsions . Caseins likely interact similarly with lipids in mammalian milk to promote lipid emulsifications . Indeed , in addition to their biological roles , caseins have also been industrialized as emulsifying agents [98] , [99] . This feature highlights the ability of these proteins to stabilize lipids present in the milk , as noted in tsetse . Proteomic studies examining mammalian milk fat globules have identified caseins , indicating that these proteins are associated with milk lipids [100]–[102] . Specifically , casein modification alters lipid composition and protein components of the milk fat globule in goats [103] . The analogous functions of MGP2–10 and caseins suggest roles for these proteins as a source of amino acids , as stabilizers of milk homogeneity , and as carriers of polyatomic ions ( i . e . phosphate groups ) . These roles must be fulfilled by a specific abundant protein or protein family to satisfy nutritional requirements of an immature organism during periods of lactation . This study provides the first complete examination of the mechanisms underlying tsetse fly lactation . In general , our results show that the majority of genes have lower expression during lactation with the exception of those directly involved in milk production . The combination of transcriptomic and proteomic analyses reveals there are 12 major milk gland proteins , which comprise ∼47% of the transcriptome of lactating flies , along with multiple minor protein constituents of tsetse milk . We have provided an overview of the combined results of this study ( Fig . 11 ) . Furthermore , we discovered a novel , tsetse-specific protein family , MGP2–10 , that is expressed highly during lactation . Interference with expression of these proteins reduces tsetse fly fecundity , suggesting that this family of MGP genes may provide a target for development of tsetse-specific abortifacients . This study has also revealed that many of the underlying functional aspects of tsetse fly lactation are analogous to those of other lactating organisms . This example of convergent evolution suggests that tsetse flies could be used as an invertebrate model system to investigate the complex molecular and physiological aspects associated with obligate lactation .
Colonies of G . morsitans morsitans were reared at Yale University and the Institute of Zoology at the Slovak Academy of Sciences ( SAS ) . The other two species ( G . pallidipes and G . fuscipes ) were reared at SAS . Flies were maintained on blood meals provided through an artificial feeding system at 48 h intervals [104] . Two groups of females were used for transcriptome analysis: the first group carried a third instar larva ( lactating ) while the second group was examined 48 h post parturition ( dry or non-lactating ) . Developing progeny were removed from each female to ensure transcript changes were representative of differences between the mothers . For sex specific transcript analysis , males and females were collected 16–18 d after adult emergence . Tissue samples were collected from pregnant females ( 16–18 d after adult emergence ) carrying third instar larvae 24 h after blood feeding . Samples for temporal expression analyses were collected according to progeny development status based on previous studies [9] , [47] . Flesh flies , S . crassipalpis , acquired from Ohio State University were reared according to standard procedures [105] . Total RNA was extracted from individual flies or dissected tissues using Trizol reagent ( Invitrogen , Carlsbad , CA , USA ) , following the recommended protocol . RNA was treated twice with the TURBO DNA Free kit ( Ambion , Austin , TX , USA ) to remove DNA , alcohol precipitated to remove residual salt , and further cleaned using the RNeasy kit ( Qiagen , Maryland , USA ) . Total RNA ( 2–3 µg ) was pooled from 10 flies extracted individually for each treatment . Sample quality and concentration was determined using a Bioanalyzer 2100 ( Agilent , Palo Alto , CA , USA ) . Library construction was performed using standard protocols for Illumina mRNA-Seq sequencing by the W . M . Keck Foundation Microarray Resource at the Yale School of Medicine . Each single-end library was sequenced on one lane of the Genome Analyzer II platform ( Illumina , San Diego , CA , USA ) . To determine Illumina read quality , FastQC analysis was performed on the transcriptomes generated from dry and lactating flies . Due to the prevalence of tsetse symbiont sequences in the reads , a specific quality control step was included to reduce bacterial sequence reads using the known whole genome sequence data from Wigglesworthia [106] , Wolbachia ( unpublished ) and Sodalis [107] determined from the same host species G . morsitans . Following symbiont specific sequence removal , remaining sequences were trimmed in CLC Genomics ( CLC Bio ) to remove ambiguous nucleotides . Contig libraries were constructed using Abyss [16] , [17] followed by a secondary assembly with Trinity [18] . Functional annotation was accomplished using the BLASTx algorithm through comparison with sequences included in the NCBI protein database [108] as well as the KOG [109] and GO databases [110] . Conserved protein domains were detected using rpsBLAST [111] searches against the CDD , Pfam and Smart databases [112] . Predicted protein translations were submitted to SignalP to identify potential secretion products by screening for secretion signal motifs [113] . Additionally , contigs were compared to several proteomes obtained from Flybase [114] ( D . melanogaster ) and Vectorbase [115] ( An . gambiae ) . Each read from each library was compared by BLASTn to the assembled coding sequences ( CDS ) using a word size of 25 , m8 output and low complexity filter turned off . CDS coverage and CDS number of read “hits” from each library were computed from the BLAST output file . A hit was only considered significant if it had 97% or better identity to its target and no more than one gap . The same read could be mapped up to three different CDS to the extent that their BLAST scores were identical . Expression levels were determined using CLC Genomics Workbench ( CLC bio , Cambridge , MA ) . Reads were mapped to our de novo assembly with an algorithm allowing only two mismatches and a maximum of 10 hits per read . RPKM was used as a measure of gene expression [116] . The proportion of read counts for each contig in relation to the total read counts in each sample was determined in order to calculate P-value differences in proportions by a Z-test following Bonferroni correction [117] . Fold change was determined as the ratio of RPKM of lactating flies vs . RPKM of dry flies . In addition to the analysis of the complete Illumina libraries , a secondary analysis was conducted featuring Illuminia libraries filtered to eliminate milk-specific contigs to reduce bias by these highly abundant proteins [116] , [117] . Data from this study are available in Sequence Read Archive ( SRA075330 ) . Pulled glass capillary tubes were used to collect milk samples by negative pressure from the guts of feeding third instar larvae , which were microscopically dissected from the uterus of pregnant females . Samples were stored in 1× protease inhibitor cocktail ( Sigma-Aldrich ) . Proteins were precipitated with 10% trichloroacetic acid ( Fisher Scientific ) at 4°C overnight , collected by centrifugation ( 11 , 000×g , 30 minutes , 4°C ) and washed two times with ice-cold acetone . Protein pellets were briefly dried and dissolved in 10 µl of protein pellet buffer ( 8M urea , 3M thiourea , and 1% dithiothreitol ) . Trypsin digestion was performed at 37°C for 12–16 h following dilution with distilled H2O to a final volume of 100 µl . Samples were stored at −80°C until analysis . Peptides were separated with a Waters nanoAcquity UPLC system ( 75 µm×150 mm BEH C18 eluted at 500 nl/min at 35°C ) with Buffer A ( 100% water , 0 . 1% formic acid ) and Buffer B ( 100% CH2CN , 0 . 075% formic acid ) . A linear gradient was established with 5% Buffer B , increasing to 50% Buffer B at 50 minutes and finally to 85% Buffer B at 51 minutes . MS/MS was acquired with an AB Sciex 5600 Triple Time-of-Flight mass spectrometer using 1 microscan followed by four MS/MS acquisitions . Neutral loss scans were obtained for 98 . 0 , 49 . 0 and 32 . 7 amu . Seven separate 1 µl injections at an estimated 0 . 351 µg/µl concentration for a total of 2 . 457 µg on the column were used for analysis . Mascot algorithm was used to analyze uninterrupted MS/MS spectra [118] . The Mascot Distiller program used MS/MS spectra to generate Mascot compatible files by combining sequential MS/MS scans from profile data that have the same precursor ion . Charge states of +2 and +3 were preferentially located with a signal-to-noise ratio of 1 . 2 or greater . A list of protein sequences was created and used in the BLASTx search against Trinity-assembled library from the pregnancy-specific analysis and positive matches were identified by tBLASTx against the NCBI and Swiss-Prot databases . Mascot scores were based on MOlecular Weight SEarch ( MOWSE ) relying on multiple matches of more than one peptide to the same predicted protein [119] , [120] . The MOWSE based ions score is equal to ( −10 ) * ( Log10P ) , where P is the absolute probability that a match is random . Matches were considered significant when the probability of a random match fell below 5% ( E value<0 . 05 ) . Therefore , Mascot scores greater than 68 were above the significance threshold when searching the newly assembled library . Proteins were considered to be successfully identified when two or more peptides matched the same predicted protein and the Mascot score exceeded the significance threshold . The exponentially modified protein abundance index ( empai ) was employed to estimate levels of protein species based on the number of species detected compared to the number of possible peptides for specific protein [121] , . Chromosomal organization of genes and full length mRNA sequences for mgp2–10 were obtained by mapping Illumina high-throughput reads against G . m . morsitans genomic scaffolds in the CLC Genomics software package . Nucleotide and predicted protein sequences were aligned using PROMALS3D [123] and Clustal [124] and formatted with BioEdit [24] . Flesh fly , Sarcophaga crassipalpis , sequences were obtained from a previous EST project [20] , [21] . Sequences of mgp2–10 from other tsetse species ( G . pallidipes and G . fuscipes ) were obtained from female cDNA by RT-PCR followed by cloning into T-vector plasmid ( Invitrogen ) and sequenced at the DNA Analysis Facility at Yale University ( New Haven , CT ) . Pairwise phylogenetic tree construction and bootstrap analysis ( 10000 replicates ) were performed using the MEGA4/5 sequence analysis suite [25] , [125] . dN/dS analyses were performed using the FEL ( Fixed Effects Likelihood [28] ) and PARRIS [27] algorithms available via DataMonkey [29] , [30] , which is a web-based implementation of the HyPhy phylogenetic analysis program [31] . Sequences were translated , aligned , reverse translated and the stop codons removed in accordance with the requirements for sequence input to DataMonkey . Under the FEL method , posterior probabilities cutoffs were set at 95 , which is equivalent to a p-value of 0 . 05 for the site-specific detection of codons under positive or negative selection . Analysis of specific regions of the MGP2–10 coding regions was conducted using MEGA5 according to previous milk protein studies [32] and individual regions were based upon protein coding regions with high or low levels of amino acid homology . For sex- and tissue-specific RT-PCR expression analyses , total RNA isolated from males and females and from dissected tissues was used as template for the Superscript III reverse transcriptase kit following the manufacturer's protocols ( Invitrogen ) . Fat body and milk gland were analyzed as a combined samples since complete separation is nearly impossible due to the intricate association of these organs . PCR was performed with gene-specific primer pairs ( Table S1 ) using the GoTaq DNA polymerase kit ( Promega ) . The PCR amplification conditions were as follows: 95°C for 3 min , 35 cycles of 30 sec at 95°C , 52–56°C for 1 min , and 1 min at 70°C using a Bio-Rad DNA Engine Peltier Thermocycler ( Hercules , CA ) . For pregnancy-specific transcript abundance determination , qPCR analyses were performed using a CFX PCR detection system ( Bio-Rad , Hercules ) . Data were analyzed with CFX manager software version 3 . 1 ( Bio-Rad ) . Primer sequences used were the same as used in RT-PCR analyses ( Table S1 ) . Comparative Ct values for genes of interest were standardized by Ct values for the control gene ( tubulin ) relative to the average value for the control treatment or newly emerged flies , yielding the delta Ct value . All experiments were analyzed in triplicate and subject to ANOVA followed by Bonferroni correction and Dunnett's test . Short interfering RNAs ( siRNA ) consisting of two Duplex sequences ( Table S1 ) were designed using Integrated DNA Technologies online software ( IDT ) . Control siRNAs were designed against Green Fluorescent Protein ( GFP; Table S1 ) . Each oligo , designed to target a single mgp gene , was also compared to the reference RNA library/G . morsitans genome [14] ) and the Trinity contigs library from this study to verify target specificity . The oligos for each strand of the siRNA were combined , and the concentration was determined spectrophotometrically followed by adjustment to 800–850 ng/µl per siRNA . Each female fly was injected with 2 µl siRNA solution into the thorax 8–10 d after adult emergence . Five days post-injection , gene expression levels were determined by qPCR and normalized to tubulin transcripts . For combined knockdown studies , siMGP constructs were mixed to yield a sample concentration of at least 600 ng/µl for each siRNA targeting a specific MGP transcript . Fecundity following MGP knockdown was assessed as previously described [9] . Finally , milk protein content was determined by Bradford assay ( Bio-Rad ) after extraction from the larval gut contents as described above . Emulsification assays were based on milk turbidity measurements . For this assay , milk was acquired from the guts of actively feeding larvae as before and diluted 10× prior to the assay . Samples were vortexed for 1 min at 10 , 000 rpm , and absorbance of the diluted emulsion was measured at 500 nm . Changes in absorbance were measured hourly for 10 h . Results were analyzed based on the slope of a regression , where ln ( ABSt/ABS0 ) is plotted versus time based on the exponential model ( ABSt = ABS0 e−kt ) . For this model , ABSt denotes absorbance at any time t , ABS0 is the initial absorbance , and k is the rate of absorbance decline in %/h . To generate structural models for MGP2–10 , four web-based de novo protein modeling programs were consulted . QUARK is a recently developed ab initio assembly program that will first break proteins into small sequences , following which full-length sequence models are assembled using Monte Carlo simulations [126] . The I-TASSER program first develops a three-dimensional model and subsequently predicts function based on structural similarity with functionally defined proteins [127] . Phyre2 is a widely used protein homology/analogy recognition engine that can rapidly predict the structure of 250 residue proteins [128] . Finally , SPARKS-X is a program that performs well in comparison to other programs [129] . Each program was run under the default configuration and the resultant predicted protein structures were visualized using Discovery Studio 3 . 1 ( Accelrys ) .
|
Tsetse flies are the sole vector for African trypanosomes , causative agents of sleeping sickness in humans and nagana in cattle . Transcriptome and proteome analyses were utilized to examine the underlying mechanisms of tsetse lactation that occur during each reproductive cycle . These analyses revealed a dramatic shift to the synthesis of milk proteins during lactation and a novel milk-specific protein family . All members of this family were co-localized , shared sequence similarity and were expressed at 40× basal levels during milk secretion . Suppression of gene from this lactation-associated family impaired progeny development by reducing milk protein content and altering milk homeostasis . These novel genes represent an excellent target for tsetse-specific reproductive-based control mechanisms . In addition , the characterization of tsetse milk production revealed multiple factors that are functionally analogous between tsetse and mammalian lactation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"genomics",
"functional",
"genomics",
"entomology",
"proteins",
"genome",
"analysis",
"transcriptome",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"proteomics",
"computational",
"biology",
"zoology",
"parasitology"
] |
2014
|
A Novel Highly Divergent Protein Family Identified from a Viviparous Insect by RNA-seq Analysis: A Potential Target for Tsetse Fly-Specific Abortifacients
|
Chromosome conformation capture ( 3C ) techniques have revealed many fascinating insights into the spatial organization of genomes . 3C methods typically provide information about chromosomal contacts in a large population of cells , which makes it difficult to draw conclusions about the three-dimensional organization of genomes in individual cells . Recently it became possible to study single cells with Hi-C , a genome-wide 3C variant , demonstrating a high cell-to-cell variability of genome organization . In principle , restraint-based modeling should allow us to infer the 3D structure of chromosomes from single-cell contact data , but suffers from the sparsity and low resolution of chromosomal contacts . To address these challenges , we adapt the Bayesian Inferential Structure Determination ( ISD ) framework , originally developed for NMR structure determination of proteins , to infer statistical ensembles of chromosome structures from single-cell data . Using ISD , we are able to compute structural error bars and estimate model parameters , thereby eliminating potential bias imposed by ad hoc parameter choices . We apply and compare different models for representing the chromatin fiber and for incorporating singe-cell contact information . Finally , we extend our approach to the analysis of diploid chromosome data .
The rapid development of chromosome conformation capture techniques such as 3C [1] , chromosome conformation capture-on-chip [2] , the closely related circular chromosome conformation capture [3] ( both named 4C ) and 5C [4] culminated in a genome-wide variant , Hi-C [5] , which provides all-against-all contact information . Hi-C experiments confirmed previously established hallmarks of genome organization including the existence of chromosome territories [5 , 6] and led to important new findings such as the partitioning of chromosomes into alternating active and passive , megabase-sized compartments [5 , 7] and in non-tissue-specific topologically associating domains ( TADs ) on the sub-megabase scale [8–10] . Chromosome conformation capture experiments typically analyze populations of millions of cells , thereby only providing a population-averaged view . Recently , however , Nagano et al . [11] pioneered Hi-C on single cells by executing all of the steps of the original Hi-C protocol within permeabilized cells and selecting individual cells for further analysis . Although the single-cell Hi-C approach provided only very sparse contact data , the structural information was sufficient to reveal unprecedented insights into genome organization including a high cell-to-cell variability of interdomain and trans-chromosomal contacts as well as the persistence of TADs across single cells . Many structural insights such as the existence of TADs or the scaling behavior of contact probabilities with genomic distance can be found by analyzing genome-wide contact matrices . Nevertheless , it seems attractive to obtain a more direct view of the 3D architecture of genomes by structural modeling based on the experimental contact information . To compute representative 3D structures of genomes , various approaches have been explored . There is a growing array of computational methods for calculating consensus structures from population Hi-C data . Typically , these methods first derive distances from the experimental contact frequencies by using different heuristics . In the early work by Duan et al . [12] , a model of the yeast genome was computed based on data from a 4C-related experiment . Bau et al . [13] mapped inverse log Z-scores from 5C data to distances and used the Integrated Modeling Platform ( IMP ) [14] to compute structural models . PASTIS [15] addresses chromosome structure determination by means of maximum likelihood , whereas ChromSDE [16] relies on semi-definite programming . Along with the structure , both PASTIS and ChromSDE optimize an additional free parameter , which is used to translate contact counts into distances . Trieu et al . [17] used an optimization-based approach , but modeled contact counts explicitly . Also Zhang et al . [18] avoided the conversion of contact counts to distances , but instead of a consenus structure , they obtain structure ensembles by simulating from an approximate energy landscape for chromosomes . Two major challenges complicate the adaptation of methods for chromsome structure inference from population Hi-C to single-cell data . First , single-cell Hi-C measures only the formation of a contact rather than contact frequencies . Second , only a small subset of all chromosomal contacts is measured such that the contact information is very sparse . Therefore , specialized methods for the analysis of single-cell Hi-C contacts need to be developed . Multidimensional scaling ( MDS ) is a popular method to obtain three-dimensional structures from incomplete and noisy distance information and was already used in the first publication on chromosome conformation capture [1] . A major limitation of MDS is that with dwindling number of data serious artifacts are introduced , which eventually leads to a complete break-down of the procedure . Shortest-path reconstruction in 3D ( ShRec3D ) [19] and an approach employing manifold-based optimization ( MBO ) [20] are two recent variants of MDS that aim to overcome these challenges for single-cell Hi-C data . Both methods define a contact distance between loci that show a contact in the Hi-C experiment and introduce a similar distance between neighboring loci along the chromatin fiber . The missing entries of the distance matrix are imputed by shortest-path distances , which are computed for a graph derived from the experimental contacts and the fiber connectivity . ShRec3D applies MDS directly to the completed distance matrix , whereas Paulsen et al . downweigh the shortest-path distances and utilize optimization techniques on matrix manifolds [21] . Nagano et al . [11] used restraint-based modeling to obtain structures of the X chromosome from their single-cell data . They derived distances from the contact data and combined the restraint energy with a simple polymer model . To find chromosome structures that fit the restraints , they used simulated annealing ( SA ) combined with molecular dynamics . However , the application of optimization approaches such as SA or MDS to chromosome structure determination suffers from the same conceptual problems described by Rieping et al . [22] in the context of protein structure calculations from NMR data . First , the scoring function typically involves model parameters that are unknown and set to ad hoc values . An example is provided by the weighting factors that define the “strictness” of the restraints [23] . Second , structure ensembles generated by minimization approaches lack a sound statistical foundation . These ensembles are computed by running multiple minimizations from randomly varying initial structures . Although this practice seems plausible , the variability of the ensemble is not a valid “structural error bar” , because it does not only reflect the quality and amount of the data , but also the power of the optimization procedure , to mention but one reason . Third , minimization approaches fail to clearly separate model parameters from algorithmic parameters , again blurring the meaning of the structure ensemble . To address these issues , Rieping et al . [22] introduced Inferential Structure Determination ( ISD ) as an unbiased and parameter-free alternative to minimization approaches to biomolecular structure determination . ISD is a Bayesian probabilistic framework that views biomolecular structure determination as an inference problem . At the core of the ISD approach is a probability distribution over conformational space representing and combining both noisy and possibly incomplete data as well as prior knowledge about the unknown structure . Several Bayesian approaches have been developed to model the structure of chromosomes based on ensemble Hi-C data [24 , 25] . These methods try to infer a consensus structure of the population data , which is only of limited use . Moreover these approaches have not fully benefited from the Bayesian approach to chromosome structure inference . Wang et al . [26] calculate structural ensembles using a Bayesian approach , but only optimize the posterior probability and do not apply the full Bayesian inference machinery , which allows for the quantification of parameter uncertainties and the comparison of alternative models . Here we report on the application of ISD to infer statistically well-defined ensembles of chromosome structures from single-cell Hi-C data . We show that Markov chain Monte Carlo ( MCMC ) sampling allows us to compute diverse ensembles of coarse-grained chromosome conformations that reflect the sparsity of single-cell Hi-C contacts . MCMC techniques and the flexibility of our Bayesian approach also allow us to compare different models of the chromatin fiber as well as alternative models for Hi-C contacts . We use the conformational ensembles to map epigenetic marks into three-dimensional space . Furthermore , we demonstrate that ISD outperforms alternative methods on simulated data . Finally , we show how to extend the approach to diploid chromosomes and infer the structures of two chromosome copies simultaneously .
We model the chromatin fiber with a beads-on-a-string representation . Owing to the sparsity of single-cell Hi-C contacts , we use a highly coarse-grained model in which every bead represents 500 kb of chromatin and has a radius of approximately 215 nm . Beads are connected such that they form a linear chain . The connectivity is enforced by a harmonic backbone potential , which penalizes distances between consecutive beads as soon as they exceed the bead diameter a . Beads are soft and allowed to overlap to some extent . We investigated two volume exclusion terms: a purely repulsive potential under which two beads that are closer than their diameter repel each other , and a Lennard-Jones potential with repulsive and attractive contributions . The sum of the backbone and nonbonded potential are part of the prior distribution ( see Materials and Methods for details ) . We first studied the properties of our model for the single-copy X chromosome of male mouse , which measures 166 Mb in length , and which we represent using 333 beads . We generated structures from the prior distribution and reconstructed the distribution of the radius of gyration Rg as a measure for the compactness of the fiber . Because there are no or only weak attractive interactions between the beads , the vast majority of structures generated from the prior showed an extended conformation . Fig 1A displays the density of states ( or equivalently the microcanonical entropy ) as a function of the radius of gyration . The density of states counts how many chromosome conformations map to a particular Rg value . There is a strong entropic force that pushes the fiber into an extended state characterized by a large radius of gyration . For the repulsive excluded volume term we found Rg = 11 . 2 ± 1 . 1 μm; for the Lennard-Jones potential we have Rg = 8 . 6 ± 1 . 3 μm . Due to the attractive contribution in the excluded volume term , the Lennard-Jones potential shows a higher preference for compact structures . Yet for both potentials , the fraction of compact structures with Rg smaller than 2 μm , say , is vanishingly small: 1 . 8 × 10−24 for the quartic repulsion potential and 7 . 8 × 10−21 for the Lennard-Jones term . The preference of the prior probability for extended structures with large average radii of gyration is incompatible with fact that chromosomes localize in chromosome territories , i . e . relatively compact subcompartments of the nucleus [6] . To good approximation , the size of the chromosome is a function of the radius of gyration . Empirically , we found that 1 . 21 × R g 1 . 11 gives a good estimate of the size of the X chromosome for relatively compact structures ( Fig 1B ) ; a simpler , yet precise enough relation is 2 . 58 × Rg . Using this approximation , the experimental chromosome size measurement of 3 . 7 ± 0 . 3 μm obtained with X-chromosome paint FISH [11] corresponds roughly to Rg ≈ 1 . 43 ± 0 . 12 μm . Therefore , the average radii of gyration reported above are an order of magnitude too large compared to the experimental finding . To incorporate the information from FISH into our probabilistic chromosome model and thereby inform the prior probability about the expected chromosome size , we assume a Gaussian error model for the chromosome size measurements . Based on our approximate relation between chromosome size and Rg , this term corresponds to a harmonic radius of gyration restraint with an experimental Rg value of 1 . 43 ± 0 . 12 μm . By using probability calculus , it is possible to combine single-cell Hi-C contacts with the FISH data and our chromosome models ( see Materials and Methods ) . Nagano et al . [11] analyzed Th1 cells of male mouse and selected ten cells that passed various quality criteria . We first focused on the most promising data set from cell 1 showing 616 X-chromosomal cis-contacts , which represents the highest number of contacts among all ten cells . We mapped these contacts onto the 500 kb beads; the removal of intra-bead contacts ( “self-contacts” ) resulted in 438 contacts , out of which 399 were unique ( some contacts mapped onto identical pairs of beads ) . We used a logistic model to quantify the probability that a cis-chromosomal contact is observed in a Hi-C experiment . The logistic function is a smooth version of a step function and has probability close to one if the contact is formed in the model structure of the fiber , and vanishes if the beads are too far apart . Consistent with other approaches [27 , 28] , we chose a distance cutoff dc = 1 . 5 × a ≈ 650 nm to decide whether two beads i and j are in contact . The smoothness of the logistic function was chosen such that a distance violation of 2 . 5% of the contact distance dc has a probability smaller than 10−6 . We generated ensembles of X-chromosome structures for both excluded volume potentials , both without and with the additional model for FISH data introduced in the previous section . To sample chromosome conformations we used Hamiltonian Monte Carlo ( HMC ) [29] , a stochastic variant of molecular dynamics . We started the HMC simulation from a fully extended X-chromosome structure . To ensure correct conformational sampling , we used replica exchange Monte Carlo [30 , 31] which runs multiple HMC simulations at different temperatures in parallel and exchanges conformations between the different simulations ( see Materials and Methods ) . Replica exchange simulations are among the most powerful Monte Carlo methods to simulate complex probability distributions but computationally demanding . It is also possible to generate X-chromosome conformations that satisfy the experimental contacts by only running HMC at a fixed temperature . The HMC sampler rapidly finds a compact structure that fits the contact data very well without producing significant violations . Nevertheless , the following results were obtained by running replica exchange simulations . As a first validation of our inference approach we studied whether all experimental contacts can be satisfied in the 3D model of the X chromosome . Fig 2A shows that this is indeed the case . For all four prior distributions the number of violations fluctuates about a small percentage of less than 3% ( see also S1 Fig ) . By increasing the steepness of the logistic function , we could decrease this number to exactly zero . As a further validation we analyzed the average pairwise distance matrix computed from the sampled X-chromosome structures . Fig 2B and 2C compare the experimental contacts with the average distance matrix . We observe that loci with a high number of cis-chromosomal contacts correspond to patches of small pairwise distances in the distance matrices . Using a diagonal permutation test , we found that the average distance matrices based on the different prior probabilities agree to a large extent , which underlines the fact that the prior does not have a strong influence on the average properties of the structure ensemble . The correlation of the distance matrices ranges between 93% and 95% ( see S1 Fig for details ) . Nevertheless , we can use Bayesian model comparison to study whether the data show any preference for one of the prior probabilities . To do so , we estimated the model evidence ( also known as marginal likelihood ) from the MCMC simulations . The model evidence quantifies how likely a probabilistic model is in the light of the data . Fig 2D shows that the contact data prefer the Lennard-Jones term over the quartic repulsion term . The incorporation of the information from FISH raises the model evidence further ( see S1 Table ) . Fig 2E shows the radii of gyration obtained without and with FISH data when using the Lennard-Jones potential . The structures generated with the cis-chromosomal contacts only , without additional FISH data are already quite compact with an average Rg of 1 . 32 μm . Due to the strong forces exerted by the logistic contact restraints , the ensemble is slightly more compact than suggested by the FISH data . When incorporating the FISH term , the average Rg shifts towards larger values with an average of 1 . 38 μm , corresponding to a chromosome size of ∼3 . 6 μm . The FISH data do not compromise the fit with the contact restraints: the number of violations does not change upon incorporation of the Rg model ( see S1 Fig ) . However , because the additional radius of gyration term helps to focus the conformational sampling on reasonably compact chromosome structures , the model evidence of the FISH based posterior is higher than without FISH data . By looking at the variance of the pairwise distances in the structure ensembles ( Fig 2F and S1 Fig ) , we find that the regions at the start of the X chromosome and around the centromere show the highest degree of conformational diversity . It is unclear , however , if this increased variability reflects true conformational fluctuations or simply the fact that these regions are unmappable . Nonetheless , we also observe an overall rise in the distance fluctuations towards the telomeric region , which might indicate that this part is indeed more dynamic . Many approaches to infer chromosome conformation from Hi-C data resort to modeling based on distance restraints . To this end , pairwise distances need to be derived from the experimental contact information . For example , the contact frequencies measured in ensemble Hi-C experiments were converted to distances by assuming a power law that relates the contact probability to the inter-bead distance , which is motivated by results from polymer physics ( see e . g . [15] ) . In case of single-cell data , this is more challenging because only single contacts are observed and not contact frequencies . To interpret the observation of a single-cell Hi-C contact as a distance measurement , we introduce an unknown distance δ between two loci that will be crosslinked . For the unknown chromosome conformation X it should therefore hold that δ ≈ dij ( X ) where dij ( X ) is the model distance between beads i and j representing both loci . Because δ is unknown , we estimate this model parameter simultaneously with the chromosome structure . In contrast , Nagano et al . [11] used a / n i j 2 as experimental distance where a denotes the bead diameter and nij counts how often beads i and j form a contact after mapping the high-resolution contacts onto the coarse-grained representation of the fiber . At 500 kb resolution nij ranges from 1 to 3 . In our approach , the repeated occurrence of a contact ( nij > 1 ) does not lead to a shortening of the contact distance , but rather to an enforcement of the distance restraint , which is duplicated nij times . Due to experimental errors and shortcomings of our model , we have to account for discrepancies between the unknown experimental distance and the model distances . This is achieved by introducing a probabilistic model for the distribution of the discrepancy between δ and dij ( X ) . We studied two error distributions: The first assumes a Gaussian shape with a flat plateau for distances between δ ± 0 . 2 × a in accordance with the approach by Nagano et al . [11] . The second model is a lognormal distribution [32] . Both error models depend on an additional unknown error parameter σ , which reflects how well the experimental distance agrees with the model distance . The inverse variance w = 1/σ2 can be interpreted as the weight of the distance restraint potential [23] . We set w to relatively large values to reflect our assumption that all observed contacts are correct ( which we also assumed in the logistic contact model ) . We used w = 100 for the Gaussian with flat plateau and w = 500 for the lognormal model , because it has a softer shape than the harmonic restraint resulting from the Gaussian model . To estimate the experimental distance using ISD , we rewrite δ = a/γ where γ > 0 is an unknown scaling parameter such that ideally a = γdij ( X ) . Fig 3A shows histograms of the estimated distance scale γ . The distance scale γ attains similar values for both error models . The average values are 0 . 724 ± 0 . 003 for the Gaussian with a flat plateau and γ = 0 . 744 ± 0 . 018 for the lognormal model . These values translate into an estimated inter-locus distance of ∼1 . 4 × a , which is comparable to the contact distance dc assumed in the logistic contact model . The distances involved in an experimental contact ( shown in Fig 2A for the contact model ) are less restrained in the distance-based models ( see S2 Fig ) . Nevertheless the agreement between the structure ensembles generated with both distance-based models is fairly high . The correlation between the average distance matrices is 95% ( Fig 3B ) , and both ensembles also agree well with the ensemble based on the logistic contact model ( with a correlation coefficient of 96% between the average distances matrices generated by each of the distance-based models and the contact model ) . Using Bayesian model comparison , we can also answer which of the two error models , Gaussian with a flat plateau or lognormal model , is preferred by the experimental data . Fig 3C shows the model evidence Pr ( D|I ) as a function of the replica temperature parameter λ . The values for λ = 1 indicate that the data tell us to prefer the Gaussian model with a flat plateau over the lognormal error model . By running calculations in which we varied the error parameter σ , we found that the exact value of the distance scale γ depends on the assumption that we make about the reliability of the Hi-C data to some extent ( see S3 Fig ) . However , for reasonably small σ/large w this dependence is less strong and the distance scale γ reaches a plateau . Bayesian inference also allows us to estimate both model parameters , γ and σ , simultaneously , but this results in rather broad ensembles that do not adopt a well-defined structure ( see S1 Appendix ) . The failure to generate well-defined structure ensembles , if w is allowed to vary is due to the sparsity of the data: The intrinsic tendency of the chromatin fiber to adopt a disordered state is stronger than the forces exert by the distance restraints with variable weight . Only with a large enough weight it is possible to counter-balance the entropic forces . Another remedy is to improve the model of the chromatin fiber [33] , but this is beyond the scope of this article . Our results show that it is possible , in principle , to model single-cell Hi-C contacts as distance measurements . However , we will use the logistic contact model in the remainder of this article , because single-cell Hi-C observes binary contacts rather than continuous distances . We now take a closer look at the ensembles generated with the ISD approach and compare them to the published ensemble by Nagano et al . [11] . The structure of the X chromosome adopts a bipartite conformation formed by two super-domains that approximately span the centromeric half ( ∼1–100 Mb ) and the telomeric half ( ∼100–166 Mb ) of the chromatin fiber . This large-scale domain structure is readily apparent from both the experimental contacts and the average distance matrix ( Fig 2C ) and becomes immediately visible in an explicit representation of the structure ensemble ( Fig 4A ) . Cluster analysis reveals that the ISD ensemble comprises multiple principal conformations about which the structures fluctuate . Closer inspection shows that the cluster centers are partial mirror images of each other . None of the likelihood and prior factors contributing to the posterior distribution distinguishes between a particular bead configuration and its mirror image , because all factors depend on distances only . Moreover , since there are only few contacts between the super-domains , each super-domain can show two conformations which are mirror images of each other . This results in at least four possible chromosome conformations which all achieve a similar goodness of fit of the Hi-C contacts . Our cluster analysis finds that the eight most dominating structural clusters produced by ISD cover ∼90% of all states sampled from the posterior ( see S4 Fig for further details ) . The four most highly populated clusters are shown in Fig 4A . These are approximate mirror images of each other . We applied the same type of cluster analysis also to an ensemble of 200 X-chromosome structures computed by Nagano et al . [11] ( see S5 Fig ) . We found similar structural clusters that are partial mirror images of each other and have corresponding structural clusters in our ensemble ( details given in Supplementary Information ) . However , overall the ISD ensemble seems to be more diverse , showing more clusters than the ensemble by Nagano et al . Visual inspection of the structural clusters suggests that the overall variability in the ISD ensemble is quite high and comparable to the fluctuations in the ensemble by Nagano et al . But each cluster of the ISD ensemble appears to be slightly better defined than the clusters in the ensemble by Nagano et al . This might be due to the more exhaustive sampling achieved by our Monte Carlo algorithm , the attractive contributions in the excluded volume term and the fact that we model Hi-C measurements as contacts with a steep sigmoidal contact probability rather than distance restraints . For each cluster , we studied the local variability of the beads by using standard techniques for the analysis of NMR structure ensembles . We estimated the local precision of the bead positions by the root mean square fluctuation ( RMSF ) after superposition of the cluster members onto the cluster center . Fig 4B shows RMSF curves for both ensembles . The RMSF curves from the four major clusters of the ISD ensemble correlate almost perfectly , the same is true for the clusters of the previously published X chromosome ensemble . We also find a high agreement between the RMSF fluctuations in the ISD clusters and the fluctuations within the clusters of the ensemble by Nagano et al . The average Pearson correlation coefficient of the RMSF profiles is 91% , showing that the ensembles obtained with both approaches agree strongly in their conformational heterogeneity . Together with the high correlation of 85% between the average distance matrices of the ISD ensemble and the ensemble by Nagano et al . ( see S2 Table ) , these findings indicate that both ensembles show many similar properties . A tube representation of the local variability of the four principal conformers is shown in panel 4A . Again , we find a higher conformational diversity towards the telemore , which was already apparent from the standard deviation of pairwise bead distances ( Fig 2F ) . We also ran ISD simulations on contact data from 5 additional Th1 cells . The average distance matrices indicate that the ensembles are significantly different , indicating the cell-to-cell variability of chromosome conformations found by Nagano et al . [11] ( see S6 and S7 Figs ) . A comparison of the ensembles obtained with data from cell 1 to cell 6 is reported in S2 Table . This comparison shows that there is an overall agreement between the X-chromosome conformations obtained with the ISD approach and the restraint-based modeling approach by Nagano et al . also for the data sets from other cells . By comparing the average distance matrices ( S6 Fig ) , we conclude that chromosome structures from different cells share some common features , such as the partitioning into more or less well-defined domains , which appear as blocks of small distances along the diagonal . The size and location of these domains can differ significantly from cell to cell . While in cells 1 , 4 , 5 and 6 , a well-defined telomeric domain is visible as a block between ∼130 − 166 Mb , it is much less pronounced in cells 2 and 3 . This difference is also evident in the structural models ( S7 Fig ) , in which the telomeric domain appears as a separated structural domain ( colored in red ) . In the models for cell 4 , the telomeric domain is on average ∼10 Mb shorter than in the models for cell 1 , 5 and 6 . The six different models also exhibit considerable differences in the spatial proximity of loci that are far apart in sequence . An example involves the centromeric region on the lower left of the average distance matrices . In the ensemble representing the X chromosome in cell 6 , this region is spatially close to many loci that are up to ∼100 Mb distant in sequence , while in cell 5 , this region shows significantly larger distances to most other loci . S7 Fig confirms this by showing that the blue and cyan parts of the X chromosome structure are much more exposed in cell 5 than in cell 6 . Taken together , a picture of highly variable , presumably stochastic chromosome organization emerges , in which conformations nevertheless share large-scale properties across different cells . A problem with current Hi-C based chromosome modeling is that it is difficult to validate the calculated structures . However , there are some independent sources of information , not used during modeling , that should be consistent with a meaningful structure ensemble . One is the information provided by population Hi-C . Although population Hi-C looks at a large pool of cells , the information about the absence of contacts should also hold for the chromosome structure based on single-cell data . As can be seen from S8 Fig by comparing the contact probabilities derived from the ISD ensemble with population Hi-C data , the ISD ensemble based on single-cell data indeed avoids contacts between loci that have a low contact probability in the population Hi-C map . Another validation is provided by the location of beads involved in trans-chromosomal contacts . Fig 4C and S9A Fig show that beads which are engaged in contacts with loci on other chromosomes tend to accumulate on the periphery of the structure ensemble , which is indirect evidence in support of our chromosome ensembles . Based on the inferred structure ensemble it becomes possible to generate three–dimensional maps of genomic and epigenetic features and to correlate the features spatially . Fig 4C shows a volume representation of chromosomal regions that are enriched in H3K4me3 and lamin-B1 associated domains in the first structural cluster . Loci that are enriched in these epigenetic marks tend to aggregate in three-dimensional space . The lamin associated domains as well as H3K4me3-enriched regions both show a tendency to locate in the periphery of the X chromosome , where they occupy distinct regions . In accord with previous findings by Nagano et al . , we also find that H3K4me3 is enriched in some parts of the interior of the X chromosome ( see S9C Fig ) . Due to the sparsity of the chromosomal contacts , we have used a very coarse-grained representation of the chromatin fiber . At this resolution , we can only study large-scale chromosomal organization . Higher resolution representations are typically needed to gain biologically relevant insights into 3D chromosome organization . We therefore also applied the ISD approach using a ten-fold higher resolved chromatin fiber . Each bead now represents 50 kb of chromatin , thereby matching the finest resolution used by Nagano et al . [11] . At this resolution the radius of a bead amounts to ∼100 nm . At 50 kb resolution , we represent the X chromosome with 3330 spherical beads . We generated structure ensembles based on the Lennard-Jones volume exclusion term and the additional FISH restraint . We modeled the intra-chromosomal contacts with the logistic model; in contrast to Nagano et al . no additional “anti-contact” restraints from the ensemble Hi-C matrix were introduced . To compare the overall properties of the structure ensembles generated at 500 kb and 50 kb resolution , we downsampled the distance matrices from the 50 kb models to match the resolution of the coarse-grained models . Downsampling was achieved by averaging 10 × 10 patches of the average distance matrix at 50 kb resolution . The downsampled average distance matrix and the average distance matrix of the low-resolution model show a correlation of 86% , indicating that the overall shape of the X chromosome at both levels of resolution is similar ( see also S10 Fig ) . Fig 5A shows a representative conformation from the structure ensemble generated at 50 kb resolution . As with the models obtained at 500 kb resolution , the X chromosome adopts a bipartite conformation where two major domains corresponding to the centromeric and telomeric regions adopt a slightly kinked conformation . Moreover , the centromeric superdomain appears to be more densely packed . Trans-chromosomal contacts and epigenetic features show a similar distribution in the high-resolution ensemble in comparison with the chromosome ensemble at 500 kb resolution ( see Fig 5B ) . Fig 5C shows the contact frequency between beads in the ISD ensemble as a function of the genomic distance s ( separation between beads along the fiber ) . In the region from 200 kb to 2000 kb genomic distance , the contact frequency shows random coil behavior . It has been argued on the basis of ensemble Hi-C data [5 , 34] that chromosomes adopt a fractal globule packing . The fractal globule shows s−1 dependence of the contact frequency , in contrast to the equilibrium globule which is expected to show a scaling behavior of s−3/2 until it reaches a constant value . Our 50 kb ensemble shows a mixed packing supporting a more complicated packing than the fractal globule also in single cells . Because Th1 cells of male mice are diploid , most cis-chromosomal contacts of the single-cell Hi-C data from Nagano et al . show contacts involving two-copy chromosomes . The difficulty with these data is that we do not know whether a particular contact is formed in the first or the second copy of the chromosome . Therefore , we have to disentangle the contacts to use them for structural modeling . A naive approach to achieve this kind of demixing is to fit the contacts with a single chromosome structure and consider the violated restraints as contacts specifying the structure of the second copy . This approach is problematic because the first structure will try to explain as many contacts as possible , which results in a strong strain . Because the model of the fiber is very generic and flexible , it will be possible to fit a large fraction of all contacts with a single structure . This is shown in Fig 6A where we used a single structure to compute models from the single-cell cis-contacts of chromosome 1 . Only a small fraction of less than 10% of all 1662 experimental contacts is strongly violated in the single-copy ensemble . It is hard to imagine that this small fraction of contacts is sufficient to determine a meaningful structure of the second copy . Furthermore , the number of restraints per copy would be highly unbalanced , for which there is no plausible physical reason . A reason for the failure of the naive approach is that it tries to assign contacts to either of the two copies similar to an exclusive-or operation . The more appropriate operation would be a logical or: A contact is either formed in both chromosome copies or in only one of the two copies . To implement this approach efficiently , we can again benefit from concepts developed in NMR structure calculation . Ambiguous NMR crosspeaks showing inter-proton contacts are routinely modeled using ambiguous distance restraints ( ADRs ) [35] . The reason is that due to the chemical shift degeneracy a peak can often be explained by multiple alternative contacts . Which of the alternatives is the correct one , is not known at the beginning of the structure calculation . An ADR combines all alternative distances into a single average distance that has to match the experimental distance . The trick of ADRs is to average the distances not arithmetically , but by summing the inverse sixth powers of the distances followed by taking the inverse sixth root of this sum ( r−6 averaging ) . In case of NMR structure calculation , this type of averaging can be motivated by the physical nature of the NMR signal . But we can also use it in different applications to implement some kind of or operation [36]: Due to the strong decay of the inverse sixth power large distances contribute only very weakly to the ADR . Therefore all configurations that could have led to the observation of the contact ( contact formed in only one of the two copies and contact formed in both copies ) , approximately result in the same average distance . To model two-copy chromosomes with ISD , we use two independent chromatin fibers and describe the observation of Hi-C contacts using a logistic model . However , in contrast to the single-copy approach , the logistic function is evaluated for the ADR computed by r−6 averaging of the corresponding distance in each copy . We only introduce intra-fiber distance restraints , although contacts could in principle arise also from inter-fiber contacts between the two copies . We exclude this possibility and assume that homologous chromosomes segregate into distinct territories that have no physical contact [37] . Fig 6B plots the restrained distances in the final structure of copy 1 , d i j ( 1 ) , against the distances in copy 2 , d i j ( 2 ) . Since there are no distance pairs that are both significantly larger than the contact distance dc , all contacts can be satisfied in only one of the two copies or in both . There is a cluster of contacts that are found in both structures ( d i j ( 1 ) < d c and d i j ( 2 ) < d c ) . Other contacts are only formed in one of the two copies ( copy 1: d i j ( 1 ) < d c and d i j ( 2 ) > d c , copy 2: d i j ( 1 ) > d c and d i j ( 2 ) < d c ) . A classification of the contacts into these three classes is shown in Fig 6C . Note that the fraction of contacts that is either assigned to copy 1 or 2 is quite balanced: 42% of the contacts are assigned to copy 1 , 41% are assigned to copy 2; the shared contacts are mostly found close to the diagonal . Fig 6D and 6E shows the average structure for both copies of chromosome 1 . Again , we observe a bipartite domain architecture where the first copy is less compact than the second copy . Another way of looking at the structural differences between the two copies are the average distance matrices shown in Fig 6F . Our model of the two-copy chromosome 1 suggests that homologous chromosomes exhibit a degree of structural variability within the same cell that is similar to the cell-to-cell variability across different cells . With a correlation coefficient of 51% , the difference between the distance matrices of both chromosomal copies is similar to the difference between structures of the X chromosome from different cells ( see S2 Table ) . This finding agrees with FISH images of chromosome territories showing that homologous chromosomes can adapt very different shapes ( see , for example , chromosome paint images of mouse lymphocytes [38] or human fibroblasts [39] ) .
In conclusion , our work shows that ISD provides a statistically sound and viable alternative to restraint minimization or embedding approaches for single-cell Hi-C data and produces less biased and statistically valid ensembles of chromosome conformations .
The Inferential Structure Determination ( ISD ) approach [22] views the determination of macromolecular structures as a problem of statistical inference and asks how likely a structure X is in the light of incomplete and noisy data D as well as prior information I . In the language of probability theory , the answer is provided by the probability distribution Pr ( X|D , I ) defined over the entire conformation space . ISD uses Bayes’ theorem to compute the posterior probability Pr ( X|D , I ) : Pr ( X | D , I ) = Pr ( D | X , I ) Pr ( X | I ) Pr ( D | I ) ( 1 ) where Pr ( D|X , I ) is the likelihood , Pr ( X|I ) the prior probability and Pr ( D|I ) the model evidence or marginal likelihood . The likelihood Pr ( D|X , I ) quantifies the agreement between the data and the structural model . To construct the likelihood , we typically choose a forward model for calculating mock data from a structure X and an error model that assesses the agreement between experimental and mock data . The forward model may involve unknown parameters such as scaling factors; the error model depends on unknown noise levels . Therefore , the 3D structure X is not the only unknown parameter in our inference problem . Additional model parameters θ need to be inferred from the data along with the conformational degrees of freedom . This is accomplished with the augmented posterior distribution Pr ( X , θ | D , I ) = Pr ( D | X , θ , I ) Pr ( X , θ | I ) Pr ( D | I ) . ( 2 ) After specifying the prior distributions reflecting our background knowledge ( e . g . a force field for the conformational parameters or the positivity of a scaling parameter ) , the problem of inferring the unknown structure and additional model parameters is formally solved . Because the posterior distribution is typically of a nonstandard form and exhibits multiple modes , we have to use powerful random sampling techniques such as Markov Chain Monte Carlo ( MCMC ) [42] to generate representative structures and model parameters ( see below ) . We represent a chromosome by N spherical beads of diameter a , which determines the length scale of the model ( Fig 7 ) . A typical value for the density of a chromosome is 12 Mb/ μm3 [43] . Our coarse-grained model uses beads representing 500 kb of chromatin such that the volume occupied by a single bead is 1/24 μm3 . Therefore the radius of a bead is approximately 215 nm . At a resolution of 50kb , the bead radius is approximately 100 nm . The distance di , i+1 between two consecutive beads i , i + 1 is restrained to an upper limit of a above which deviations are penalized quadratically with a force constant kbb = 250/a2 . Overlaps between two beads are penalized using an excluded volume term Enb with force constant knb . The conformational prior distribution defined over bead positions X = ( x1 , … , xN ) is given by the canonical ensemble: Pr ( X | I ) ∝ exp - k nb E nb ( X ) - k bb E bb ( X ) ( 3 ) with E bb ( x ) = ∑ i θ ( d i , i + 1 - a ) ( d i , i + 1 - a ) 2 ( 4 ) where I denotes all prior assumptions , i . e . specific values for a , kbb , etc . Here and in Eq 5 , θ ( x ) denotes the Heaviside step function; θ ( x ≥ 0 ) = 1 and θ ( x < 0 ) = 0 . Here we study two nonbonded force fields to account for volume exclusion effects . The first force field consists only of a repulsive term that is activated as soon as two beads come closer than their diameter a . A repellent force resulting from a quartic repulsion energy pushes the beads apart: E nb ( X ) = ∑ i < j θ ( a - d i j ) ( a - d i j ) 4 . ( 5 ) The strength of the force is determined by the force constant , which we set knb = 5/a4 . The second excluded volume term is a Lennard-Jones potential with linear asymptotes: E L J ( X ) = ∑ i < j 13 - 14 s 6 s 12 - 12 1 - s 6 s 13 d i j a ; d i j < s · a a d i j 12 - 2 a d i j 6 ; s · a < d i j ≤ l · a 1 - 2 l 6 l 12 d cut - d i j d cut - l · a ; l · a < d i j ≤ d cut 0 ; d i j > d cut ( 6 ) with s = 0 . 6 , l = 1 . 25 , dcut = 1 . 375 × a . The effective potentials between successive beads are shown in Fig 7B . FISH experiments measure the size of chromosome territories and show that the chromatin fiber adopts a compact conformation during interphase . To incorporate FISH data into our model , we first need to measure the size of a chromosome given the bead positions X . The squared radius of gyration is defined as: R g 2 = 1 N ∑ n = 1 N ∥ x n - x ¯ ∥ 2 ( 7 ) where x ¯ = 1 N ∑ n x n is the center of mass of all beads . For compact chromosome conformations we can define the size of a chromosome as the maximum extent of a bounding box circumscribing all beads . Empirically we found that the chromosome size is proportional to the radius of gyration where the proportionality factor is ∼2 . 58 . Let us now assume that FISH experiments result in a list of chromosome size measurements s1 , … , sM . We model each observation using a Gaussian error model such that the probability of observing the m-th size measurement is: p ( s m | X , I ) = 1 2 π σ FISH exp - 1 2 σ FISH 2 s m - 2 . 58 × R g ( X ) 2 ( 8 ) where σFISH is the error of the FISH measurements . The complete probability of all FISH measurements is: p ( s 1 , … , s M | X , I ) = ∏ m p ( s m | X , I ) = 1 ( 2 π σ FISH ) M exp - M 2 σ FISH 2 [ s ¯ - 2 . 58 × R g ( X ) ] 2 + Δ s 2 ( 9 ) where s ¯ = 1 M ∑ m s m is the average chromosome size and Δs is the standard deviation . For the X chromosome of Th1 cells we have s ¯ = 3 . 7 μm and Δs = 0 . 3 μm [11] . We studied three probabilistic models to measure the compatibility of a chromosome structure X with single-cell Hi-C data D . The data is a list of experimental contacts C between pairs of loci i and j . Two of the three models that we used to analyze the Hi-C contacts are based on distance restraints similar to the approach by Nagano et al . [11] . To account for inaccuracies in the data and the model , we use an error model that describes the likelihood of a mismatch between the experimental and back-calculated distances . The third model is based on sigmoidal contact probability and prefers conformations satisfying all observed contacts similar to Trieu et al . [17] . Due to coarse graining , we might observe multiple contacts between two beads; nij denotes the corresponding number of counts in the binned contact matrix . Distance / contact restraints between beads i , j with nij > 1 are duplicated nij times . Because the full posterior probability Pr ( X , θ|D , I ) ( or Pr ( X1 , X2 , θ|D , I ) in case of diploidy ) is of a non-standard form , we have to use Markov chain Monte Carlo ( MCMC ) sampling to draw representative chromosome conformations and model parameters that we can then use to compute average values and error bars . Gibbs sampling [44] is an iterative MCMC algorithm that draws from the posterior probability by cycling over successive steps that alternate between sampling of the conformational degrees of freedom and the model parameters: X ( t + 1 ) ∼ Pr ( X | θ ( t ) , D , I ) ∝ Pr ( D | X , θ ( t ) , I ) × Pr ( X | I ) θ ( t + 1 ) ∼ Pr ( θ | X ( t + 1 ) , D , I ) ∝ Pr ( D | X ( t + t ) , θ , I ) × Pr ( θ | I ) ( 16 ) This permits the use of adequate samplers for each model parameter . The normalizing constant Pr ( D|I ) in Eq 2 is the marginal likelihood or model evidence . It is generally hard to calculate Pr ( D|I ) , which is a high-dimensional integral over all unknown parameters X and θ . A great advantage of MCMC methods is that they can sample from unnormalized probability distributions . Therefore , the evidence Pr ( D|I ) does not need to be known when applying MCMC to sample from the ISD posterior . Although knowledge of Pr ( D|I ) is not necessary for parameter estimation ( i . e . sampling of X and θ ) , the model evidence is crucial , if we want to do model comparison , i . e . if we want to quantify , for example , if the chromatin beads should exhibit only repulsive forces or both repulsive and attractive forces . The model evidence Pr ( D|I ) reflects how well the chosen likelihood and prior distributions describe the Hi-C data D and allows us to rank different modeling assumptions in the light of the data . To calculate the evidence , we use histogram reweighting techniques [48–50] , which allow to use all samples from a RE simulation to estimate the model evidence in replica with high accuracy . We used similar techniques to study the properties of the chromosome model ( see Chromosome models ) . Because the prior alone prefers extended chromosome conformations , we imposed an additional radius of gyration term whose strength is varied smoothly via the inverse temperature λ of the replicas: p ( X | λ ) ∝ exp { - λ R g ( X ) - k bb E bb ( X ) - k nb E nb ( X ) } . ( 18 ) This RE simulation allowed us to explore all degrees of compactness of the chromatin fiber by varying λ between λ = 0 and λ = 500 using 64 replicas . The same histogram reweighting techniques that we used to calculate model evidences allowed us to obtain a precise estimate of the entropy as a function of Rg . The entropy measures the number of conformations with a certain radius of gyration and is shown in Fig 1A . Structure ensembles reconstructed from single-cell Hi-C contacts are typically less well defined than NMR structure ensembles . Therefore traditional measures to characterize and compare structure ensembles such as RMSDs are of limited use . To validate our structure calculation method , we ran several test calculations for a three-dimensional Hilbert curve comprised of 512 beads . We assumed a cutoff distance of 1 . 5 × bead radius , which resulted in a maximum of 3696 contacts ( ∼2 . 8% of the full distance matrix , which has 512 × 511/2 = 130816 elements in total ) . With the complete set of observable contacts , we obtained an ensemble that comprised two principal conformers which are mirror images of each other . The first cluster shows an average RMSD of 0 . 09 bead radii from the correct model and is populated by 48% of all sampled structures . The second cluster achieves the same RMSD to the mirror image of the ground truth and is populated by the remaining 52% of all structures . This shows that ISD is capable of drawing correct samples from the posterior distribution: Distance data alone cannot distinguish between a 3D structure and its mirror image . Therefore , we expect that the original structure of the Hilbert curve and its mirror image should be present in the ISD ensemble with identical probability . We also computed structures with sparsified versions of the full set of contacts by randomly selecting a smaller number of contacts . In the sparsified data sets , the number of contacts ranged from 3326 to 578 contacts corresponding to 2 . 8% to 0 . 44% of all distances ( the sparsity of the contact data in case of the ISD bead model of the X chromosome at 500 kb resolution is 0 . 68% ) . For all data sets , we obtained an ensemble of structures that comprised two major conformers: an approximate version of the input structure and its mirror image . The average structures of both clusters are shown in S11 Fig . Even with as few as 0 . 44% of all distances , ISD inferred structures that showed the correct domain architecture . The RMSD ranged from 0 . 13 to 1 . 46 bead radii for both the cluster approximating the ground truth and the cluster showing its mirror image ( see S12A and S12B Fig ) . Also the spread and population of the clusters were almost identical ( S12C and S12D Fig ) . These tests show that ( 1 ) ISD is capable of generating statistically valid ensembles , which is indicated by the fact that we obtained a bimodal posterior ensemble , in which the peaks corresponding to the original structure and its mirror image have identical spread and population ( see S12 Fig ) . ( 2 ) ISD generates structure ensembles that reflect the quality/completeness of the data , because the precision of the ensemble decreases with increasing sparsity of the contact data ( see S12C and S12D Fig ) . To better judge the performance of ISD , we also computed structural models with other methods for chromosome structure inference for the Hilbert curve data . Simulations were performed using an extended version of the ISD library [57] . Analyses were carried out with Python scripts that rely only on widely used additional packages and the CSB toolbox [58] . Code and scripts for analysis are available at https://github . com/michaelhabeck/isdhic .
|
Spatial interactions between distant genomic regions are of fundamental importance in gene regulation and other nuclear processes . Recent chromatin crosslinking ( “Hi-C” ) experiments probe the spatial organization of chromosomes on a genome-wide scale to an extent that was previously unattainable . These experiments report on contacting loci and thus provide information about the three-dimensional structure of the genome . Unfortunately , the data are noisy and do not determine the structure uniquely . There is also little quantitative prior knowledge about the large-scale organization of chromosomes . Here , we address these challenges by developing a Bayesian statistical approach that combines a minimalist polymer model with chromosome size measurements and conformation capture data . Our method generates statistical ensembles of chromosome structures from extremely sparse single-cell Hi-C data . We remove potential bias by learning modeling parameters from the experimental data and apply model comparison techniques to investigate which among a set of alternative models is most supported by the Hi-C data . Our method also allows for modeling with ambiguous contact data obtained on polyploid chromosomes , which is an important step towards three-dimensional modeling of whole genomes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"chromosome",
"structure",
"and",
"function",
"mathematics",
"statistics",
"(mathematics)",
"protein",
"structure",
"epigenetics",
"structural",
"genomics",
"chromatin",
"research",
"and",
"analysis",
"methods",
"sex",
"chromosomes",
"chromosome",
"biology",
"proteins",
"mathematical",
"and",
"statistical",
"techniques",
"gene",
"expression",
"x",
"chromosomes",
"monte",
"carlo",
"method",
"molecular",
"biology",
"genetic",
"loci",
"biochemistry",
"chromosomes",
"cell",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"genomics",
"statistical",
"methods",
"macromolecular",
"structure",
"analysis"
] |
2016
|
Inferential Structure Determination of Chromosomes from Single-Cell Hi-C Data
|
Post-translational modifications of histone proteins modulate the binding of transcription regulators to chromatin . Studies in Drosophila have shown that the phosphorylation of histone H3 at Ser10 ( H3S10ph ) by JIL-1 is required specifically during early transcription elongation . 14-3-3 proteins bind H3 only when phosphorylated , providing mechanistic insights into the role of H3S10ph in transcription . Findings presented here show that 14-3-3 functions downstream of H3S10ph during transcription elongation . 14-3-3 proteins localize to active genes in a JIL-1–dependent manner . In the absence of 14-3-3 , levels of actively elongating RNA polymerase II are severely diminished . 14-3-3 proteins interact with Elongator protein 3 ( Elp3 ) , an acetyltransferase that functions during transcription elongation . JIL-1 and 14-3-3 are required for Elp3 binding to chromatin , and in the absence of either protein , levels of H3K9 acetylation are significantly reduced . These results suggest that 14-3-3 proteins mediate cross-talk between histone phosphorylation and acetylation at a critical step in transcription elongation .
The ability of cells to modulate their transcriptional program in response to physiological stimuli is vital for the proper development of eukaryotic organisms . This is partly achieved by post-translational modification of histone proteins that associate with the DNA to form nucleosomes , the basic units of chromatin . Histone modifying enzymes respond to signaling cues by acetylating , methylating or phosphorylating specific amino acids of histone N-termini . Such modifications alter the affinity and accessibility of chromatin to transcription activators and silencers and consequently dictate the expression profiles of genes [1] . Several transcription regulators are known to have chromatin-binding motifs that interact with histones modified at specific amino acids . One example is the chromodomain , which binds methylated histones . It was identified in the heterochromatin protein HP1 and the developmental regulator Polycomb ( Pc ) , which bind histone H3 when methylated at lysine 9 ( H3K9me ) and lysine 27 ( H3K27me ) , respectively [2] . Other regulators such as the chromatin remodeling enzyme Brd4 possess a bromodomain that recognizes acetylated histones [3] . More recently , the basal transcription factor TFIID was shown to bind H3 when methylated at K4 ( H3K4me ) via the plant homeodomain ( PHD ) finger of the subunit TAF3 [4] . Cross-talk between histone modifications adds another layer of complexity . In yeast , histone H2B monoubiquitination is required for both H3K4 and H3K79 methylation during transcription activation [5] . In Drosophila , a decrease in H3S10 phosphorylation leads to the ectopic spread of H3K9 methylation on polytene chromosomes [6] . Furthermore , phosphorylation of H3S10 during mitosis in mammalian cells antagonizes the effect of H3K9 methylation and leads to the dissociation of HP1 from chromosomes [7] . In addition to defining distinct transcription and chromatin states , histone modifications can also mark different stages of transcription of active genes . During the transcription process , RNA polymerase II ( Pol II ) communicates with histone modifying enzymes via its C-terminal domain ( CTD ) , which consists of a heptad repeat with serine residues at positions 2 and 5 . These serine residues are differentially phosphorylated depending on the specific phase of the transcription cycle . Ser5 is phosphorylated during the very early steps of elongation , while Ser2 phosphorylation marks the release of Pol II from promoter-proximal pausing , a tightly regulated checkpoint that ensures proper capping of the mRNA . In yeast , the SET 1 enzyme responsible for H3K4 methylation interacts with the C-terminal domain of Pol II when it is phosphorylated at Ser5 . Consistent with this observation , H3K4 methylation is predominant at the 5′end of genes [8] . On the other hand , the SET 2 H3K36 methyltransferase associates with the Ser2-phosphorylated form of the CTD and H3K36 methylation is enriched at the 3′end of genes [9] . Phosphorylation of histone H3 at Ser10 ( H3S10ph ) has long been implicated in transcription activation in organisms ranging from yeast to humans [10]–[12] but it was only more recently that a mechanistic explanation for how H3S10ph could contribute to gene activation became apparent . Mahadevan and colleagues showed that several members of the 14-3-3 phospho-binding protein family interact with H3 only when phosphorylated and that human 14-3-3ζ is recruited to c-fos and c-jun when these genes are transcriptionally activated [13] . Furthermore , Winter et al . showed that 14-3-3 is recruited to the HDAC1 gene upon its activation in an H3S10ph-dependent manner and is required for the transcription of this gene [14] . More recently , Zippo et al . have shown that 14-3-3 plays a crucial role in the transcription of the mammalian FOSL1 gene by recruiting the histone acetyltransferase MOF . Acetylation of H4K16 by MOF helps recruit the double bromodomain BRD4 protein , which in turn recruits P-TEFb [15] . These authors suggest that H3K9ac is also involved in the process but evidence and details of the proteins involved are not available . The Drosophila genome contains two different 14-3-3 genes , 14-3-3ζ and 14-3-3ε , providing a simple model to study 14-3-3 function . The role and dynamics of H3S10ph during transcription activation in Drosophila are well characterized . Phosphorylation of H3 during interphase is carried out by the JIL-1 kinase and mutations in the JIL-1 gene lead to severe disruption of polytene chromosome morphology with a marked loss of the well-defined pattern of bands and interbands characteristic of wild-type chromosomes [16] . These structural perturbations do not affect the recruitment of transcription factors or Pol II to active genes . Instead , H3S10 phosphorylation is required during promoter-proximal pausing for recruitment of the P-TEFb kinase and phosphorylation of the CTD of Pol II at Ser2 , which is required for the release of Pol II and transcription elongation [17] . Here we explore the role of H3S10ph- mediated recruitment of 14-3-3 to chromatin during transcription activation . Results show that 14-3-3 proteins are recruited to active genes in a JIL-1-dependent manner and are required for phosphorylation of Pol II at Ser2 . Further analyses indicate that 14-3-3 interacts with elongator protein Elp3 , a histone acetyltransferase required during early elongation . The recruitment of Elp3 to chromatin and the subsequent acetylation of H3K9 are dependent on JIL-1 , suggesting that 14-3-3 proteins mediate crosstalk between H3 phosphorylation and acetylation during early transcription elongation .
Numerous studies in organisms ranging from yeast to vertebrates have established a role for H3S10 phosphorylation in transcription activation [10]–[12] . Specifically , studies in Drosophila indicate that this modification is involved in the release of Pol II from promoter-proximal pausing during early transcription elongation [17] . The JIL-1 kinase , which is the homologue of the vertebrate MSK1/2 kinases , is responsible for this modification in Drosophila [16] . Interestingly , mutations in JIL-1 not only result in a genome-wide decrease in transcription but also cause dramatic changes in the structure of polytene chromosomes [16] . Although the two effects are probably related , it has been recently questioned whether JIL-1 and H3S10 phosphorylation play a role in transcription and whether the observed recruitment of JIL-1 to heat-shock genes upon induction , and the ensuing H3S10 phosphorylation , are artifacts resulting from the fixation procedures utilized in the immunofluorescence microscopy analyses used to derive these conclusions [18] . To address these concerns , JIL-1 antibodies were used in standard chromatin immunoprecipitation ( ChIP ) experiments to examine whether JIL-1 is recruited to the hsp70 promoter when the gene is induced in Drosophila Kc cells . The results confirm our previous observations showing that JIL-1 binds to the promoter region of the hsp70 gene only after the cells are subjected to heat-shock ( Supporting Figure 1A in Text S1 ) . The same result was observed in ecdysone-induced genes after hormone stimulation ( data not shown ) . Total levels of H3K79 methylation and H3K36 methylation , which mark transcription elongation , were also examined and found to be reduced in JIL-1 mutants , providing additional and independent evidence for our previous finding of a role for JIL-1 in transcription elongation ( Supporting Figure 1B and 1C in Text S1 ) . Recruitment of JIL-1 to the hsp70 promoter explains the phosphorylation of H3S10 at heat-shock puffs , which has been observed consistently by various investigators using antibodies from different sources and varying fixation protocols [11] , [17] , [19]–[23]; also see Supporting Figure 4A in Text S1 . In agreement with this conclusion , polytene chromosomes from JIL-1 null mutants lack H3S10ph at the heat-shock puffs , while Pol II phosphorylated at Ser5 is still present [17] . Taken together , these findings strongly support a role for JIL-1 in transcription elongation in general and in that of hsp70 in particular . Results presented below as well as those of Zippo et al . , 2009 lend further support to this conclusion . In order to better understand the mechanisms underlying transcription regulation by JIL-1 mediated H3S10 phosphorylation and to elucidate downstream events , the role of 14-3-3 proteins in transcription was examined . Previous studies have shown that two human 14-3-3 proteins ( 14-3-3ζ and 14-3-3ε ) associate with active genes in an H3S10ph-dependent manner [13]–[15] . However , these studies were limited to the analysis of specific genes and it is yet unclear whether 14-3-3 plays a more general role in transcription . A pan 14-3-3 antibody was used to examine whether 14-3-3 proteins are broadly distributed on Drosophila polytene chromosomes . This antibody was raised against the peptide DKSELVQKAKLAEQAERY found in the N-terminus of human 14-3-3β , which is highly similar to an amino acid stretch in Drosophila 14-3-3ζ ( DKEELVQKAKLAEQSERY ) and partially so to Drosophila 14-3-3ε ( ERENNVYKAKLAEQAERY ) [24] . When protein extracts from salivary glands of third instar wild type larvae are analyzed on Western blots using this antibody , two bands of approximately 25 kDa can be seen ( Figure 1A ) . A hs-Gal4 strain that expresses Gal4 in the salivary glands in the absence of any heat-shock treatment [25] was then used to express UAS-RNAi against 14-3-3ζ , 14-3-3ε , or both simultaneously in order to determine the identity of the two bands ( Figure 1A ) . When RNAi against 14-3-3ζ is expressed , the lower band is severely diminished , with significant increase in the signal from the higher band . When RNAi against 14-3-3ε is expressed , a severe reduction of the higher band is observed , accompanied by partial reduction of the lower band . RNAi against both isotypes leads to a reduction in both bands , suggesting that the antibody recognizes both isotypes , and that the higher band corresponds to 14-3-3ε while the lower one corresponds to 14-3-3ζ ( Figure 1A ) . This antibody was then used in immunofluorescence microscopy to determine the localization of 14-3-3 proteins on polytene chromosomes derived from salivary glands of Drosophila third instar larvae . The results reveal a broad distribution of 14-3-3 proteins on chromatin as depicted in Supporting Figure 2A in Text S1 . This signal is dramatically decreased in larvae expressing RNAi against both isoforms of 14-3-3 , verifying the specificity of the antibody in immunofluoresence analyses ( Supporting Figure 2B in Text S1 ) ; antibodies against the insulator-binding protein Su ( Hw ) were used to control for signal intensity . Co-staining with anti-Pol IIoser5 ( Figure 2A ) shows that 14-3-3 proteins are found at interband regions and significantly co-localize with Pol II , as reflected by the abundance of yellow signal in the merge panel ( see close up in Figure 2A ) . Similar analyses using anti-H3S10ph ( Figure 2B ) and anti-JIL-1 ( Figure 2C ) antibodies show that 14-3-3 also co-localizes extensively with these two proteins , consistent with the hypothesis that they bind phosphorylated histones . The heat-shock paradigm was then used to better understand the dynamics of 14-3-3 distribution during transcription . When larvae are subjected to temperature elevation , the heat-shock ( hsp ) genes are turned on while all genes that were previously active are turned off . Immunofluorescence analyses of polytene chromosomes using antibodies against H3S10ph show that heat-shock treatment leads to disappearance of the modification from previously active genes and its redistribution to the induced hsp genes [11] . If 14-3-3 is recruited by phosphorylation of H3S10 , the distribution of 14-3-3 should exhibit similar behavior to that of H3S10ph upon heat-shock . Third instar larvae were incubated at 36 . 5°C for 20 min and their salivary glands were immediately dissected and fixed for immunofluorescence analysis using anti-14-3-3 in combination with anti-Pol IIoser5 ( Figure 2D ) or anti-H3S10ph ( Figure 2E ) antibodies . Results show that the pattern of 14-3-3 binding , like that of H3S10 phosphorylation , changes from a broad distribution throughout the genome to one that is restricted to the heat shock genes . This suggests that 14-3-3 proteins are recruited to actively transcribed genes and that their binding to chromatin correlates with H3S10 phosphorylation . The experiments described above were then repeated in JIL-1z2 null mutants to determine whether the recruitment of 14-3-3 is dependent on H3S10 phosphorylation . We have previously reported that despite the disrupted structure of the chromosomes of these mutants , Pol IIoser5 can still be detected on chromatin at wild-type levels [17] . Antibodies against Pol IIoser5 , as well as Su ( Hw ) , were therefore used to control for signal intensity levels . As can be seen in Figure 3B and 3C , and Supporting Figure 2C in Text S1 , no 14-3-3 protein can be detected on the chromosomes of JIL-1z2 mutants before or after heat-shock as compared to wild type ( Figure 3A , and Supporting Figure 2C in Text S1 ) whereas levels of Pol IIoser5 ( Figure 3B–3C ) and Su ( Hw ) ( Supporting Figure 2C in Text S1 ) are unaffected . This is not due to a decrease in total levels of 14-3-3 in the cell , as they remain unchanged in JIL-1z2 mutants ( Figure 1A ) . The binding of 14-3-3 to chromosomes was also examined in salivary glands that express a dominant negative form of the Brahma ( BRM ) chromatin remodeling ATPase , which acts very early during transcription initiation of many genes – excluding heat-shock genes – and is required for the recruitment of the Pol II machinery to chromosomes [25]; levels of phosphorylated H3S10 are significantly reduced in chromosomes from larvae carrying this dominant mutation [17] . Results indicate that the binding of 14-3-3 to polytene chromosomes is also disrupted in the dominant negative brm mutant , supporting the conclusion that H3S10 phosphorylation is required for the presence of 14-3-3 on chromatin ( Figure 3D ) . In contrast , similar analyses on chromosomes from heat-shocked brm mutants show that 14-3-3 is recruited to heat-shock puffs ( Supporting Figure 2D in Text S1 ) , consistent with the previous report that BRM is not required for activation of heat-shock genes [26] . Taken together , the findings described above establish a correlation between H3S10 phosphorylation , 14-3-3 recruitment and transcription . It remains possible , however , that 14-3-3 proteins bind to other components of the transcription machinery and that their association with H3S10 phosphorylation is strictly correlative and indirect . 14-3-3 recruitment to polytene chromosomes was therefore examined in larvae that are mutant for the kismet ( kis ) gene , which encodes a chromatin remodeling ATPase required during early elongation of transcription . In these mutants , like in JIL-1 mutants , transcription can initiate but productive elongation is impeded [27] . H3S10ph levels , however , remain comparable to wild-type [17] . As can be seen in Figure 3E , the binding of 14-3-3 to chromatin in kis mutants is similar to that observed in wild-type larvae , consistent with the hypothesis that 14-3-3 binding is specifically dependent on H3S10 phosphorylation and not simply on transcription elongation . Given the broad distribution of 14-3-3 proteins on polytene chromosomes as well as their close association with H3S10 phosphorylation , we hypothesized that 14-3-3 proteins may play a general role in transcription elongation . Western analysis was therefore used to determine levels of elongating Pol II in salivary gland cells expressing RNAi against each of the Drosophila 14-3-3 proteins separately or both simultaneously . As can be seen in Figure 1A , RNAi against 14-3-3ε alone or both 14-3-3 proteins leads to a significant decrease of total Pol IIser2 levels , while RNAi against 14-3-3ζ has no effect . Similarly , levels of H3K36 methylation , a marker of transcription elongation , were significantly reduced when both proteins were knocked down ( Supporting Figure 1C in Text S1 ) . Anti-lamin antibodies were used to control for loading in both experiments . 14-3-3 proteins function as versatile dimeric structures that can modulate various forms of protein-protein interaction in response to signaling cues . In many instances they serve as scaffolds , bridging proteins that cannot directly interact [28] . We therefore hypothesized that the recruitment of 14-3-3 to chromatin by phosphorylated H3S10 may serve to regulate the interaction between H3S10ph and other chromatin-binding proteins . Such proteins would have to ( 1 ) interact with 14-3-3 , ( 2 ) be chromatin-related , and ( 3 ) function during transcription elongation . Various reported biochemical screens , aimed at isolating 14-3-3-binding proteins , provide a vast database to search for candidates that fit these criteria . An exhaustive review of 14-3-3 interactors identified the histone acetyltransferase Elp3 ( Elongator protein 3 ) [29] , [30] , a subunit of the Elongator complex that co-purifies with Pol II [31] and is required for H3 acetylation specifically during transcription elongation in yeast [29] , [32] , [33] . In order to confirm an interaction between Elp3 and 14-3-3 in Drosophila , anti-14-3-3 antibodies were used to immunoprecipitate proteins from Kc cell extracts followed by Western analysis using anti-Elp3 antibodies . Figure 1B shows that both Elp3 and phosphorylated H3S10 co-precipitate when 14-3-3 anti-serum is used but not with beads alone , suggesting that these proteins directly or indirectly interact in vivo . The distribution of Elp3 on chromosomes and its co-localization with 14-3-3 was also examined . As expected from a histone acetyltransferase that is involved in active transcription , Elp3 is found at interband regions when analyzed by immunofluorescence microscopy ( Figure 4A , left panel ) . Co-staining with 14-3-3 revealed co-localization between the proteins at many sites ( see representative close up ) , suggesting that the interaction observed by co-immunoprecipitation may be relevant to their role in transcription . We then asked whether the role of Elp3 in transcription elongation is conserved in Drosophila . The Elp3 protein is present at heat-shock puffs after heat-shock ( Figure 5E ) , suggesting it is recruited to active genes upon induction . The Elp3EX1 allele was generated by imprecise excision of a P-element inserted 65 bp upstream of the transcription start site creating a deletion from the P-insertion site to the triplet encoding K277 [34] . This mutation results in pupal lethality and formation of melanotic tumors during larval stages . Since the mutant flies are unable to make any Elp3 protein , the late lethality is probably a consequence of the perdurance of maternal Elp3 protein into pupal stages of development . Therefore , larvae carrying the Elp3EX1 allele may still accumulate some Elp3 protein and they cannot be considered null [34] . This is especially obvious in the morphology of the polytene chromosomes , which present a range of chromatin alterations from complete absence of interbands to slightly distorted chromosomes still showing some band/interband pattern ( data not shown ) . We thus selected chromosomes with fully penetrant phenotypes such as those shown in Figure 4 and Figure 5 because they lack Elp3 protein ( see Figure 5B ) and are representative of a null effect . Chromosomes with some remnant of the band/interband pattern still contain some Elp3 protein and were not considered in our analysis . Levels of Pol II binding to chromatin were analyzed by immunofluorescence imaging of polytene chromosomes using antibodies to Su ( Hw ) as an internal control . Both anti-Pol IIoser5 and anti-Pol IIoser2 antibodies were used to determine whether flies carrying Elp3 mutations exhibit transcription defects . Consistent with a role in elongation , levels of Pol IIoser5 are maintained at wild type levels in the Elp3EX1 mutant ( compare Figure 4B and 4C ) while those of Pol IIoser2 were dramatically reduced ( compare Figure 4D and 4E ) . Western analysis of protein extracts from salivary glands of Elp3EX1 mutants also show that total levels of Pol IIoser2 are severely reduced when compared to wild type ( Figure 1A ) . Levels of methylated H3K36 , a marker of transcription elongation , are also reduced in these mutants , consistent with a role for Elp3 in transcription ( Supporting Figure 1C in Text S1 ) . If 14-3-3 recruits Elp3 in response to H3S10 phosphorylation , we would expect to see less Elp3 protein on polytene chromosomes in the absence of 14-3-3 or JIL-1 . To determine if this is the case , antibodies against Elp3 and phosphorylated Pol II ser5 were used to stain polytene chromosomes from wild type and JIL-1 mutant larvae or larvae expressing RNAi against both 14-3-3 mRNAs . Elp3 co-localizes with Pol IIser5 at interband regions in wild type chromosomes ( Figure 5A , representative close up ) . In contrast , no Elp3 protein can be detected on chromosomes of JIL-1 mutants or 14-3-3 RNAi knock-down when Pol IIser5 ( Figure 5C and 5D ) or Su ( Hw ) ( Supporting Figure 3C–3F in Text S1 ) were used as internal controls . This is most evident at heat-shock puffs , where levels of Elp3 proteins are significantly reduced when 14-3-3 protein levels are knocked down ( Figure 5F ) as compared to wild type ( Figure 5E ) . At the same time , JIL-1 protein can still bind to chromosomes in Elp3 mutants ( Supporting Figure 3B in Text S1 ) at wild type levels ( Supporting Figure 3A in Text S1 ) , suggesting that Elp3 functions downstream of JIL-1 during transcription elongation . In agreement with this conclusion , JIL-1 and H3S10ph are present at heat shock puffs after temperature elevation in polytene chromosomes from Elp3 mutant larvae ( data not shown ) . The above results point to a functional interaction between JIL-1 ( an H3 kinase ) and Elp3 ( an H3 acetyltransferase ) , suggesting crosstalk between histone H3 phosphorylation and acetylation during transcription elongation . H3 has been shown to be acetylated at K14 at induced heat-shock genes [11] , [21] . Immunofluorescence analysis of polytene chromosomes from heat-shocked wild type larvae detected a strong signal at heat-shock puffs when H3S10phK14ac antibodies were used ( Supporting Figure 4A in Text S1 ) , suggesting that the two modifications occur on the same histone tails . A similar analysis also detected H3K9 acetylation at the heat-shock genes ( Supporting Figure 4B in Text S1 ) . While we were unable to detect any signal using H3K9acS10ph antibodies by immunofluorescence analysis of polytene chromosomes , Western blot analysis shows that H3K9ac and H3S10ph also occur on the same histone tail in vivo ( Figure 6A ) . We first used these different antibodies to confirm that Elp3 acts as a histone acetyltransferase in Drosophila . Antibodies against H3K9ac and H3K14ac were used in Western blot analyses to compare H3 acetylation levels in Elp3 mutant and wild type larvae . As can be seen in Figure 6A , levels of H3K9ac are significantly diminished in Elp3 mutants , while those of H3K14ac are unaffected . In addition , immunofluorescence analysis with anti-H3K9ac shows that levels of this modification are diminished in polytene chromosomes of Elp3 mutants ( Supporting Figure 5D in Text S1 ) . To further confirm the specificity of Drosophila Elp3 as a histone H3K9 acetyltransferase , we isolated Elp3 by immunoprecipitation and used the protein in an in vitro acetylation assay using recombinant histones as a substrate . The results of these experiments show significant acetylation of H3K9 as compared to the no antibody control ( Figure 6B ) . Consistent with the data from Western analysis , acetylation of H3K14 could not be detected above background in this in vitro assay ( Figure 6B ) . Interestingly , levels of H3S10ph are slightly reduced in Elp3 mutants ( Figure 6B ) , despite the fact that JIL-1 binding to chromosomes is not affected ( Supporting Figure 3B in Text S1 ) . Antibodies to phospho-acetylated H3K9S10 and phospho-acetylated H3S10K14 were then used to determine the extent of the doubly modified histone H3 tails in Elp3 mutants . As can be seen in Figure 6A , mutations in Elp3 affect H3K9acS10ph to a larger extent than H3S10phK14ac as expected from the differential effects observed on the acetylation of the two Lys residues . Since JIL-1 is required for the recruitment of Elp3 , it follows that the H3K9 acetylation defect observed in Elp3 mutants would also be observed in the absence of JIL-1 and 14-3-3 proteins . Western analyses of protein extracts derived from salivary glands from JIL-1z2 mutant larvae or glands expressing RNAi against both 14-3-3 proteins ( Figure 6A ) and immunofluoresence analysis ( Supporting Figure 5B and 5C in Text S1 ) show that levels of acetylated H3K9 are markedly reduced in both cases . In addition , immunoprecipitation of 14-3-3 followed by the in vitro acetylation assay showed that 14-3-3 associates with a histone acetyltransferase activity that is specific to H3K9 and not H3K14 , similar to the effect observed with Elp3 ( Figure 6B ) . The results discussed above suggest that 14-3-3 proteins recruit the histone H3 acetyltransferase Elp3 , which in turn acetylates H3 in the Lys9 residue . This observation is interesting in the context of recent findings suggesting that 14-3-3 can recruit the histone acetyltransferase MOF to acetylate H4K16 in the mammalian FOSL1 gene [15] . To test whether this is also the case in Drosophila , we determined whether levels of H4K16 acetylation are affected by downregulation of 14-3-3 using RNAi against the two genes encoding this protein in Drosophila . Results show that the presence of H4K16ac in protein extracts from larval tissues ( Figure 6 ) or 3rd instar polytene chromosomes ( Supporting Figure 6 in Text S1 ) is not affected in flies lacking 14-3-3 proteins . To further confirm the absence of association between 14-3-3 and Drosophila MOF , we isolated 14-3-3 and associated proteins by immunoprecipitation with 14-3-3 antibodies and used the proteins in an in vitro acetylation assay using recombinant histones as a substrate . The results of these experiments show absence of H4K16 acetylation when compared to the no antibody control ( Figure 6B ) . These results suggest that recruitment of MOF may not require 14-3-3 or that the function of these proteins may be redundant .
Covalent modification of histone tails can influence the binding of transcription regulators and accordingly modulate gene expression . The phosphorylation of histone H3 at S10 was recently shown to be accompanied by the recruitment of members of the 14-3-3 protein family to specific genes upon induction [13]–[15] . Here we extend these observations and present evidence that 14-3-3 mediates a novel histone crosstalk that promotes transcription elongation in Drosophila . Drosophila 14-3-3 proteins display a broad distribution on polytene chromosomes , colocalizing with phosphorylated Pol II at many sites . This localization is dependent on the presence of H3S10P and , like H3S10 phosphorylation , redistributes to heat-shock genes upon temperature elevation . In addition , total levels of elongating Pol II are severely diminished in cells lacking both 14-3-3 isotypes . These findings establish a role for 14-3-3 proteins during transcription elongation of many Drosophila genes . The specific roles of the individual 14-3-3 isotypes , however , remain unclear . The two proteins have been shown to exhibit significant functional specificity with some aspects of redundancy across various cell processes [24] , [35]–[38] . It was recently shown that while 14-3-3ε is essential for hatching of Drosophila embryos , 14-3-3ε null mutants survive because one isotype of the 14-3-3ζ gene is upregulated at the time of hatching , compensating for the loss of 14-3-3ε [39] . This , together with the data presented here , suggests that the two isotypes may play redundant roles in transcription . Total levels of active Pol II remain unaffected when 14-3-3ζ is knocked down . This could imply that 14-3-3ε is the sole isotype required for transcription elongation . However , given that 14-3-3ε null mutants are viable [39] , it is more likely that the two proteins play redundant functions in transcription , and that the upregulation of 14-3-3ε in response to 14-3-3ζ knockdown is a compensatory mechanism . Further analyses are required to determine the exact function of each isotype . Antibodies with better specificity , for example , can be used to determine whether the two proteins colocalize at all sites on chromosomes or whether they exhibit unique binding patterns . Since these proteins can function as heterodimers , it is reasonable to speculate that while they may play a basic redundant function in transcription , specific combinations of 14-3-3 binding and interaction allow to fine tune the transcription process , thereby adding an extra layer of complexity to the histone code . From a mechanistic perspective , the data presented here suggest that 14-3-3 proteins mediate cross-talk between histone phosphorylation and acetylation during transcription elongation by creating a bridge between JIL-1 and Elp3 . In yeast , Elp3 is a subunit of the Elongator complex , which was first identified by its association with elongating Pol II [31] , and later shown to be involved in histone acetylation and transcription [29]; results here show that these functions are conserved in Drosophila . On the other hand , most Elp3 protein is cytoplasmic , and a few studies have implicated the complex in tRNA modification [40] , [41] . It is therefore possible that the mutant effects documented here are indirect and can be alternatively explained by translation defects . However , our results point to a direct role for Elp3 in transcription . Elp3 binds to transcriptionally active regions on polytene chromosomes and is recruited to heat-shock genes after heat-shock . It interacts with 14-3-3 and co-localizes with it at many sites on polytene chromosomes . More importantly , binding of Elp3 to chromosomes is dependent on JIL-1 and 14-3-3 . Consistent with these observations , decreased levels of H3K9 acetylation are detected in JIL-1 mutants and 14-3-3 knockdowns , suggesting that acetylation by Elp3 is dependent on phosphorylation by JIL-1 and the subsequent recruitment of 14-3-3 . Taken together , these results strongly support a direct role for histone acetylation by Elp3 during transcription elongation , downstream of JIL-1 and 14-3-3 . Interestingly , it was recently shown in mammals that 14-3-3 proteins bind phosphorylated H3S10 at the FOSL1 gene and serve to recruit the H4K16 acetyltransferase MOF , which then acetylates histone H4 at the Lys16 residue and is required for recruitment of BRD4 and P-TEFb . Although the specific aspects of the process have not been explored in detail , it appears that recruitment of BRD4 in mammals also requires acetylation of H3K9 [15] . Consistent with this observation , we have previously shown that phosphorylation of H3S10 is required for recruitment of P-TEFb to heat-shock genes in Drosophila [17] . Nevertheless , it appears that in Drosophila 14-3-3 does not play a major role in the recruitment of MOF; instead , 14-3-3 recruits Elp3 and this protein is then required for MOF recruitment , based on the observation that H4K16Ac is dramatically reduced in flies carrying a mutation in the Elp3 gene . Taken together , the data strongly support a role for crosstalk between histone phosphorylation and acetylation during the release of Pol II from promoter-proximal pausing . The relationship between histone H3 phosphorylation and acetylation has been the subject of some debate . These two modifications are known to occur in response to the same stimuli , in the same tissue and on the same histone tails . Two possible models have been consequently put forward to explain these observations [42] . The first proposes that the two modifications are synergistic and coupled such that one is dependent on the other . This is supported by the fact that , in vitro , HATs preferentially acetylate histone tails that are phosphorylated , suggesting that histone phosphorylation provides a stronger binding site for HATs than unphosphorylated ones [43] . The second model envisions the two modifications being performed by regulatory machineries that are recruited simultaneously yet independently to active genes [42] . Studies that support this model have utilized specific antibodies that recognize the modifications either individually or together . Two populations of histones were detected at active genes using these antibodies , a larger highly acetylated population that is not phosphorylated and a smaller phosphoacetylated population , suggesting that phosphorylation is not a prerequisite to acetylation [44] , [45] . Our results support the ‘synergistic and coupled’ model , but under the premise that 14-3-3 acts as a bridge , rather than one modification acting as a binding site for the next enzyme . At the same time , the data do not rule out the other scenario . In fact , while different groups have in the past advocated one model over the other , the two are by no means mutually exclusive . It is becoming increasingly evident that multiple layers of regulation come into play during transcription activation . Promoter-proximal pausing , rather than initiation , appears to be the rate limiting step in the activation of a significant number of genes in humans and Drosophila [46]–[48] . Along the same lines and more pertinent to this study , there are at least two reports of distinct pathways responsible for histone acetylation during gene activation in yeast: Gcn5 is known to acetylate histones during transcription initiation , while Elp3 acetylates histones during elongation [32] , [33] . It is therefore possible that the acetylation that occurs during initiation is not dependent on phosphorylation , while the acetylation associated with elongation is . This would account for the two different pools . Additional evidence of a functional link between Elp3 and JIL-1 was obtained from imaging polytene chromosomes of Elp3 mutants , which display defects in chromosome structure that are very similar to those caused by mutations in JIL-1 . These defects are characterized by the loss of organization of band and interband regions and a shortening of the chromosome arms . What these defects signify in terms of the roles of JIL-1 and Elp3 in transcription , however , remains unclear . Further analyses will be required to determine the exact cause of the chromosome morphological defects and precisely how they affect transcription . This will help explain the molecular mechanisms governing transcription regulation by histone phosphorylation and acetylation and further shed light into the complex relationship between chromosome structure and transcription .
Stocks were maintained in standard medium at 18°C or 25°C . Oregon R larvae were used for wild type ( wt ) controls in all experiments . The JIL-1z2 stock was a gift from Dr . K . Johansen ( Iowa State University ) . The brmK804R and kisk13416 stocks were a gift from Dr . J . Tamkun ( UC Santa Cruz ) . The Elp3EX1 mutant was a gift from Dr . J . Svejstrup ( Cancer Research UK London Research Institute ) . UAS-14-3-3ζ RNAi flies were obtained from VDRC ( Stock #48724 ) and UAS-14-3-3ε RNAi flies were obtained from NIG ( Stock #31196R-4 ) . To express siRNA in salivary glands these stocks were crossed to +/+; hsp70-Gal4/hsp70-Gal4 ( Bloomington , 1799 ) . To express RNAi against both 14-3-3 isotypes simultaneously , 48724/48724; 31196R-4/31196R-4 flies were crossed to +/+; hsp70-Gal4/hsp70-Gal4 . Approximately 100 pairs of salivary glands from third instar wild type or mutant larvae or glands subject to RNAi were homogenized in 100 µl RIPA buffer with EDTA free protease inhibitors ( Roche ) and phosphatase inhibitors ( Sigma #P2850 ) and left on ice for 15 min . Laemmli's buffer and beta-mercaptoethanol were added and lysates were incubated at 65°C for 20 min to solubilize proteins and then insoluble fractions were spun down . Samples were run on NuPage 4–12% gradient Bis-Tris gels and transferred to PVDF membranes for immunodetection . Membranes were incubated overnight at 4°C in antibody dilution buffer ( PBS/0 . 05% Tween/5% milk or BSA in case of Pol II antibodies ) containing primary antibodies at concentrations of 1∶1000 rabbit α-14-3-3 ( SCBT ) , 1∶1000 mouse α-Pol IIoser2 ( H5 , Covance ) , 1∶5000 mouse α-lamin C ( Developmental Studies Hybridoma Bank ) , 1∶1000 rabbit α-H3S10P ( Millipore ) , 1∶5000 rabbit α-H3K9Ac ( Millipore , 07-352 ) , 1∶5000 rabbit α-H3K14Ac ( Millipore , 07-353 ) , 1∶1000 rabbit α-H3S10PK9Ac ( Abcam , ab12181 ) , 1∶1000 rabbit α-H3S10PK14Ac ( Millipore , 07-081 ) , 1∶1000 rabbit α-H3K79me ( Abcam ) and 1∶10000 rabbit α-histone H3 ( Abcam ) . The membranes were washed twice with PBS/0 . 25% Tween , incubated for 1h at room temperature in the appropriate HRP secondary antibody ( Jackson ImmunoResearch Laboratories ) and washed twice with PBS/0 . 25% Tween . Antibody signal was visualized using chemi-luminescence detection methods ( SuperSignal West Pico kit , Pierce ) . Salivary gland polytene chromosome squashes were prepared from wandering third instar larvae maintained at 18°C . For heat-shock experiments , third-instar wild type and JIL-1z2 mutant larvae were subjected to heat-shock treatment as described previously ( Nowak et al . , 2003 ) . Salivary glands were dissected in 0 . 7% NaCl and fixed for 2 min in 45% acetic acid/1 . 85% formaldehyde . Fixed salivary glands were subsequently squashed in 45% acetic acid on subbed slides . The slides were frozen in liquid nitrogen and stored dry at −70°C . For immunostaining of 14-3-3 proteins , salivary glands were fixed for 1 min in 3 . 7% acetic acid , 2 min in 45% acetic acid/3 . 7% formaldehyde and 3 min in 45% acetic acid . Slides were incubated overnight at 4°C in antibody dilution buffer ( PBS/0 . 1% Triton X-100/1% BSA ) containing primary antibodies at concentrations of 1∶50 α-14-3-3 ( K19 , SCBT ) , 1∶100 rabbit α-JIL-1 , 1∶20 rabbit α-Elp3 [29] , 1∶30 mouse α-Pol IIoser2 ( H5 , Covance ) , 1∶30 mouse α-Pol IIoser5 ( H14 , Covance ) , 1∶150 rat α-Su ( Hw ) , 1∶50 rabbit α-H3S10phK14ac and 1∶50 rabbit α-H3K9ac . Following incubation , slides were washed three times in PBS/0 . 1% Triton X-100 and incubated for 1 h at 37°C in the appropriate secondary antibody ( Jackson ImmunoResearch Laboratories ) diluted 1∶200 in antibody dilution buffer . Slides were washed three times as described above and stained with 0 . 5 µg/ml of 4′ , 6-diamidino-2-phenylindole ( DAPI ) and mounted in Vectashield mounting medium ( Vector Laboratories ) for viewing . Goat anti 14-3-3ζ serum ( SCBT ) was covalently crosslinked to Protein G sepharose beads 4 Fast Flow ( GE Healthcare ) using DMP . Drosophila Kc cells grown to 80% density were spun down , washed twice in PBS and crosslinked for 10 min in 1% paraformaldehyde . Cells were washed twice in PBS to stop the reaction and then lysed in RIPA buffer containing phosphatase inhibitors and protease inhibitors for 20 min on ice . The insoluble fraction was spun down and the soluble fraction split in two , half on the beads with antibody and the other half with beads but no antibody as a control; a small fraction was reserved for the input lane . The samples were run on 4–12% NuPage Bis/Tris gel , transferred to nitrocellulose , blocked in 1% BSA and blotted against rabbit α-14-3-3ζ ( SCBT ) 1∶2000 , rabbit α-Elp3 1∶2000 [29] , and mouse α-H3S10P ( Millipore ) . Wild type third instar larvae were repeatedly washed in PBS and then ground and vortexed in RIPA buffer with protease inhibitors . The lysate was diluted 10-fold with 1% Triton X-100/150 mM NaCl/50 mM Tris and spun down to eliminate insoluble fractions . The soluble lysate was incubated with either no antibody , Elp3 polyclonal antibody , or 14-3-3 polyclonal antibody and pulled down with protein G beads . The beads were washed 5 times in 1% Triton X-100 buffer and 3 times in 50 mM Tris pH8/150 mM NaCl containing protease inhibitors . To 25 µl of the beads containing pull-down , 50 µl 50 mM Tris pH8 . 0/150 mM NaCl , 20 µl of 1 mg/ml histones and 5 µl of 5 . 69 mM AcetylCoA were added per 75 µl reaction and incubated at 30°C for 45 min with mixing . The histones were then collected for Western analysis . Drosophila Kc cells were grown at 25°C to 7×106 cells/ml in serum-free HyQ-CCM3 medium ( HyClone Laboratories , Inc . ) . Cells were subjected to heat shock by addition of an equivalent volume of medium preheated to 48°C to the growing cells . After holding the cells at 36 . 5°C for 15 min , the cells were immediately cooled down to 25°C with the addition of 1/3 total volume of 4°C medium immediately prior to cross-linking [49] . Cells were cross-linked with 1% formaldehyde for 10 min , quenched with 0 . 125 mM glycine and washed with PBS . Nuclear lysates were sonicated to generate 200–1000 bp DNA fragments . Immunoprecipitation was performed with 7 µl α-JIL-1 or with no antibody . Immunoprecipitated DNA was extracted and amplified with primers described in Boehm et al . , 2003: hsp70+4F , 5'-CAATTCAAACAAGCAAAGTGAACAC hsp70+112R , 5'-TGATTCACTTTAACTTGCACTTTA .
|
Activation of gene expression is thought to be regulated mainly at the level of transcription initiation . Nevertheless , many genes in Drosophila and vertebrates contain RNA polymerase that has started transcription but is paused 30–40 bp from the initiation site . Activation of these genes may thus be regulated by releasing the polymerase from the paused state rather than bringing this protein to the promoter . This release requires the recruitment of specialized proteins that modify the polymerase . It appears that the recruitment of these proteins takes place by modification of two chromatin proteins , histones H3 and H4 . Here we characterize a process composed of multiple steps required for release of RNA polymerase from the paused state . The process starts with the recruitment of an enzyme that adds a phosphate group to histone H3 . This phosphate serves as a signal to recruit a different protein , which in turn recruits a second enzyme capable of adding an acetyl group to the same histone molecule . The multiple steps involved may provide a variety of mechanisms to control the process .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"molecular",
"biology/transcription",
"elongation",
"molecular",
"biology/histone",
"modification",
"molecular",
"biology/chromatin",
"structure"
] |
2010
|
14-3-3 Mediates Histone Cross-Talk during Transcription Elongation in Drosophila
|
The interaction between phytohormones is an important mechanism which controls growth and developmental processes in plants . Deciphering these interactions is a crucial step in helping to develop crops with enhanced yield and resistance to environmental stresses . Controlling the expression level of OsAP2-39 which includes an APETALA 2 ( AP2 ) domain leads to phenotypic changes in rice . Overexpression of OsAP2-39 leads to a reduction in yield by decreasing the biomass and the number of seeds in the transgenic rice lines . Global transcriptome analysis of the OsAP2-39 overexpression transgenic rice revealed the upregulation of a key Abscisic Acid ( ABA ) biosynthetic gene OsNCED-I which codes for 9-cis-epoxycarotenoid dioxygenase and leads to an increase in the endogenous ABA level . In addition to OsNCED-1 , the gene expression analysis revealed the upregulation of a gene that codes for the Elongation of Upper most Internode ( EUI ) protein , an enzyme that catalyzes 16α , 17-epoxidation of non-13-hydroxylated GAs , which has been shown to deactivate gibberellins ( GAs ) in rice . The exogenous application of GA restores the wild-type phenotype in the transgenic line and ABA application induces the expression of EUI and suppresses the expression of OsAP2-39 in the wild-type line . These observations clarify the antagonistic relationship between ABA and GA and illustrate a mechanism that leads to homeostasis of these hormones . In vivo and in vitro analysis showed that the expression of both OsNCED-1 and EUI are directly controlled by OsAP2-39 . Together , these results reveal a novel mechanism for the control of the ABA/GA balance in rice which is regulated by OsAP2-39 that in turn regulates plant growth and seed production .
Plant hormones have synergistic or antagonistic effects on the physiological processes associated with growth and development . ABA and GA are hormone partners and act through a complicated network of antagonistic interactions . The coordination and interaction between phytohormone is essential to achieve normal growth and development . In Arabidopsis ( Arabidopsis thaliana ) , a high endogenous level of ABA causes a reduction in the endogenous level of GA [1] , and vice versa [2] . ABA generally regulates development by retarding plant growth , although there is recent evidence suggesting a growth promotion effect of ABA through reducing ethylene synthesis [3]–[5] . On the other hand , GA promotes growth and is involved in seed germination , leaf expansion , shoot and root elongation , and flowering and shoot fruit development [6] . These antagonistic hormones have a mutual biosynthesis , signalling and catabolism inhibition relationship [7] . Some components of the relationship between the GA and ABA synthesis and signalling pathways have been elucidated . For example , it has been shown that the upregulation of the ABA biosynthesis gene XERICO is controlled by the DELLA protein which is a negative regulator of GA response in Arabidopsis [7] . Further , GA suppression has been shown to occur through the ABA-inducible protein kinase ( PKABA1 ) present in the aleurone layer of barley [8] , [9] . In addition , the FUS3 transcription factor which specifies cotyledon identity in Arabidopsis has also been found to regulate the synthesis of ABA and GA during late embryogenesis [10] , [11] . However , a number of issues regarding this relationship are still unclear [12]–[14] . Transcription factors control a variety of physiological processes through altering the expression of genes involved in metabolic pathways including hormone biosynthesis and signalling in plants . One set of these is the large APETALA2 ( AP2 ) transcription factor family [15] . AP2 proteins are found only in plants and their unique feature is that they include the AP2 DNA-binding domain . For instant , there are 139 and 122 AP2 putative family genes in rice ( Oryza sativa L . subsp . japonica ) and Arabidopsis respectively [16] . The AP2 gene family plays a variety of functions throughout plant growth and development including the regulation of several developmental processes like floral organ and epidermal cell identity , and they are involved in the mechanisms used by plants to respond to various types of biotic and abiotic stresses ( e . g . ( Shukla et al . , 2006 [17]; Tang et al . , 2005 [18] ) . The AP2 domain specifically binds to the GCC box ( the consensus DNA binding motif is AGCCGCC ) which was originally identified as an ethylene response element ( ERF ) . The AP2 gene family has been found to control a wide range of physiological processes including through the regulation of genes involved in hormone metabolism and signalling . In Arabidopsis , the DWARF AND DELAYED-FLOWERING 1 ( DDF1 ) [19] and the LEAFY PETIOLE ( LEP ) are AP2 transcription factors involved in regulating GA metabolism and signalling [20] . In addition , the JERF1 is an AP2 transcription factor which modulates the expression of an ABA biosynthesis-related gene in tobacco [21] . Ectopic expression of AP2 related genes in plants causes a variety of phenotypic changes . While overexpression of JERF1 increases salt and cold tolerance in tobacco , overexpression of DDF1 caused late flowering and a dwarf phenotype in Arabidopsis . These phenotypic alterations were apparently due to an alteration in the endogenous hormonal balance in the plant . The OsAP2-39 gene which codes for a member of AP2 family in rice was initially identified as a strong nitrogen-responsive gene and was transformed into rice . The purpose of this work was to determine its role in controlling growth and development . Expression profiling data revealed the change in the expression of a large number of genes in the transgenic lines overexpressing OsAP2-39 including some key hormone biosynthetic and catabolic genes . We found that overexpression of OsAP2-39 increases the endogenous level of ABA in rice through a direct regulatory interaction between OsAP2-39 and a key ABA-biosynthesis gene ( OsNCED-1 ) . In addition , either a high level of ABA or the direct action of OsAP2-39 induces the expression of the ELONGATION OF UPPER MOST INTERNODE I ( EUI ) gene [22]–[24] . EUI reduces the level of the bioactive forms of GAs by epoxidizing the active GAs in rice . Therefore , overexpression of EUI causes a dwarf phenotype whereas mutation within this gene increases the internode length in rice . The alteration in the ABA/GA ratio due to the OsAP2-39 overexpression leads to a pleiotropic phenotype including short stems , decreased tiller and panicle number , late flowering and a low percentage of seed filling . Consequently , transgenic plants have a much lower seed yield than does the wild-type . Therefore , OsAP2-39 is a key participant in ABA synthesis and GA catabolism in rice and is involved in maintaining hormone homoeostasis which is crucial for normal plant growth and development .
The rice genome codes for 139 putative AP2 family proteins with a variety of functions and domain structures [16] . The OsAP2-39 ( Os04g0610400 ) cDNA was isolated by using a PCR strategy and cloned using standard protocols . The OsAP2-39 cDNA is composed of 666 bp and the genomic sequence contains no introns . The Protein Basic Local Alignment Search Tool ( BLASTP ) available at the National Center for Biotechnology Information ( NCBI ) website ( http://blast . ncbi . nlm . nih . gov/Blast . cgi ) showed that OsAP2-39 codes for a single AP2 domain present at the N-terminal part of the protein . This domain includes 11 putative DNA-binding sites ( Figure 1A ) implying a strong binding capacity . The OsAP2-39 protein is about 22 . 8 kDa with a predicted average pI of 9 . 62 . The hydrophobicity profile indicates that OsAP2-39 domains are mostly hydrophobic and is a folded protein ( Figure 1B ) . Sequence and phylogenetic analysis showed that OsAP2-39 is similar to other AP2 family members only within the AP2 conserved domain , but the rest of the protein sequence does not have a high level of similarity with other rice proteins ( Figure 1C ) or proteins from other plant species . Based on the BLASTP search results , a deduced amino acid sequence from maize ( Gene Bank accession number ACG28382 ) has the highest degree of homology with OsAP2-39 ( 65% identity and 69% similarity ) . OsAP2-39 was localized in the nucleus of the onion epidermal cells when it was fused with the C-terminal part of the GFP ( Figure 2 ) . Although OsAP2-39 does not code for a conventional nuclear localization signal , a prediction of subcellular localization using bioinformatics tools such as LOCtree of the University of Colombia available at ( http://cubic . bioc . columbia . edu/cgi-bin/var/nair/loctree/query ) showed that OsAP2-39 is a nuclear protein with a 95% chance of possibility . Together the sequence analysis and the subcellular localization results of OsAP2-39 suggest that this protein is a transcription factor . The OsAP2-39 cDNA was constitutively overexpressed in rice under the control of a corn ubiquitin promoter . Four independent transgenic lines were chosen for further studies using the phosphomannose isomerase ( PMI ) activity assay [25] as a selectable marker . These transgenic rice plants had pleiotropic phenotypes which led to overall biomass reduction ( Figure 3A–3C ) . These included less green leaves at the 1–2-weeks old stage , shorter inter-nodes including the upper most one , fewer leaves and tillers ( Figure 3B ) , reduction in seed yield ( Figure 3C ) , and delays in flowering by 1 to 2 week . The transgenic plant height was reduced by 55% , tillers by 75% , and the number of the leaves by 74% comparing with the wild-type plants . Consequently the yield of the transgenic plants was less than the wild-type by about 80% ( Figure 3B ) . The root system of the transgenic lines is also affected by OsAP2-39 overexpression . It has about 30% less total length , surface area , average diameter , and number of tips than the wild-type ( Figure 4 and Figure S1 ) . Wild-type plants transformed with an RNAi construct designed to block the production of OsAP2-39 were made . Of the initial lines produced , 5/29 initial transformed plants showed a decrease in the level of OsAP2-39 transcript of up to 5-fold ( Figure 5A ) . While overexpression of OsAP2-39 leads to an increase in gene expression level of both EUI and OsNCED-1 , the plants having a decreased expression of OsAP2-39 also have a decreased level of expression of both EUI and OsNCED-1 ( Figure 5A ) . This supports the direct OsAP2-39 regulatory effect on EUI , and OsNCED-I . Preliminary phenotypic showed that these RNAi lines were taller and a higher tiller number . Unfortunately , the decreased expression of the OsAP2-39 gene in any of the 4 lines was not inherited in the following generation which made it impossible to confirm this phenotypic analysis . Overexpression of OsAP2-39 affects several physiological processes . This includes low germination rate and shorter internodes . Thus , overexpression of the OsAP2-39 in rice shows a similar phenotype to that found in GA deficient mutants in plants like the rice lines harbouring mutations within various GA biosynthetic genes [26] , the gibberellin insensitive dwarf1-1 ( gid-1 ) [27] , the semi-dwarf-1 ( sd1 ) [28] , the slender rice-1 ( slr1 ) [29] , and the dwarf and gladius leaf 1 ( dgl1 ) [30] mutants . In addition , the Arabidopsis mutants gibberellin insensitive dwarf1 ( atgid1a and atgid1b ) [31] , sleepy1 ( sly1 ) [32] , [33] , and gibberellin-responsive dwarfs ga1-3 [34] have a similar impact on phenotype . Given these phenotypic similarities , GA deficiency was investigated as a reason for the phenotype caused by the overexpression of the OsAP2-39 and GA3 was exogenously applied to the transgenic lines . In addition , the GA biosynthesis inhibitor paclobutrazol ( PAC ) and ABA ( as a GA antagonist ) were used to confirm the effect of the GA on the phenotype . Seeds from both genotypes were germinated on a filter paper saturated with different hormone solutions and the number of germinated seeds was counted after 6 days . In a separate experiment , seeds were planted in magenta boxes containing solutions of different hormone treatments and incubated in the dark for one week . The results showed that hormonal treatment can modify the phenotype in the transgenic lines ( Figure 6A–6D ) . Treating the transgenic seeds with GA3 recovers the seedlings height ( Figure 6A ) and also the seed germination rate ( Figure 6D ) . In comparison with the wild-type , treating the rice seeds with 10 µM PAC decreases the germination rate ( Figure 6D ) and seedlings height ( Figure 6C ) . In addition , use of ABA delays seed germination in the transgenic lines and shows more drastic effect on the growth of transgenic lines than in the wild-type ( Figure 6B ) . The results also showed that treating rice plants at the 4-weeks old stage with 100 µM GA for 4 weeks with a dosage of 2 times/week rescued the normal height and flowering time in the transgenic plants , although the number of tillers did not recover with this treatment ( data not shown ) . This may reflect additional physiological processes associated with axillary bud initiation and development or could be due to inappropriate site and/or time of GA3 application . Thus , exogenous application of the GA3 recovers the endogenous level of this hormone in the transgenic lines . This leads to a restoration of most wild-type phenotypes in the transgenic lines . On the other hand , decreasing the endogenous level of GA by the application of PAC or reducing the effect of GA by addition of ABA magnified the OsAP2-39 effect on the transgenic line . This implies a shortage in bioactive GAs due to improper gene expression . In order to determine the molecular events associated with OsAP2-39 overexpression , global gene expression analysis on the transgenic rice was carried out using the Affymetrix gene chip microarrays . RNA samples were isolated from 4-week old leaves and processed for microarray analysis . Comparing with the wild-type , the gene expression analysis results showed an alteration in 409 genes in the transgenic rice lines ( Table S1 ) . The gene list includes 172 upregulated and 237 downregulated genes . Because microarray analysis may not detect every single gene whose expression is modulated in the transgenic line , the expression of additional genes involved in GA and ABA biosynthesis and signalling were tested using quantitative real time PCR ( qRT-PCR ) analysis . Interestingly , the results showed that the expression level of genes involved in ABA biosynthesis and GA catabolic and signalling pathways were changed due to the OsAP2-39 overexpression ( Figure 5B ) . This includes the upregulation of a putative OsNCED-1 ( Os03g0645900 ) , and OsNCED-3 ( Os07g0154100 ) genes coding for the 9-cis-epoxycarotenoid dioxygenase which are ABA-biosynthetic enzymes [35] . The OsNCED-1 codes for a protein with 83% identity and 90% similarity based on the Dayhoff matrix to the maize VIVIPAROUS14 ( VP14 ) protein ( Figure S2 ) which catalyzes the cleavage of 9-cis-epoxy-carotenoids to form C25 apo-aldehydes and xanthoxin , a precursor of ABA in higher plants . As a result , it is considered to be a key enzyme in the ABA synthesis pathway [35] . The VIVIPAROUS14 expression level is directly related with the ABA synthesis rate [35]–[37] . Consistent with this observation , the active endogenous ABA level of the OsAP2-39 transgenic rice lines was found to be 2-fold higher than the wild-type level ( Figure 7A ) . In addition , the ABA derivative compounds such as Dihydrophaseic acid ( DPA ) and Abscisic acid glucose ester ( ABAGE ) levels are also increased in the rice transgenic line ( Figure 7B ) . In addition to the OsNCED genes , the rice Zeaxanthin epoxiydase ( OsZEP-1 ) ( Os04g0448900 ) was downregulated in the transgenic line . These genes are involved in the ABA biosynthesis pathway . Free ABA is deactivated by oxidation to phaseic acid and by the formation of glucose conjugates . Induction of ABA oxidation may result from a feed back inhibition loop interaction due to the excessive level of the endogenous ABA in the transgenic line . Knockouts of the OsZEP-1 caused dwarf rice mutants [38] and the relative abundance of AB2 , which codes for a Zeaxanthin epoxiydase in tobacco ( Nicotiana plumbaginifolia ) is reduced due to the increase level of ABA [39] . These previous observations are consistent with the phenotype and the ABA level obtained in the OsAP2-39 overexpressed lines . The gene expression analysis also showed the upregulation of the ELONGATED UPPERMOST INTERNODE ( EUI ) ( Os05g0482400 ) gene which encodes for a cytochrome P450 monooxygenase , an enzyme which deactivates gibberellin through an epoxidation reaction [22]–[24] . GA deactivation can occur through other mechanisms . For example , in Arabidopsis GAs are deactivated through GA 2-Oxidase including AtGA2ox7 and AtGA2ox8 [40] and GA oxidase-6 ( AtGAox6 ) [41] . The microarray and qRT-PCR data showed that a putative gibberellin 2-beta-oxidase7 ( Os04g0522500 ) is upregulated in the OsAP2-39 transgenic line . In addition , the microarray and qRT-PCR data showed the upregulation of 3 gibberellin receptor proteins: OsGID1 ( Os07g0162700 ) and GID1L2 ( Os06g0214800 , Os07g0162900 ) ( Figure 5B ) . This result indicates a regulatory role of OsAP2-39 on GA activity in the transgenic line . Exogenous application of GA3 recovers the wild-type phenotype and application of the GA inhibitor PAC magnifies the effect of OsAP2-39 on the phenotype indicating a low endogenous content of active GA in the transgenic line . Analysis of the endogenous level of GAs revealed alterations in the hyroxylated GAs in the OsAP2-39 overexpression lines ( Table 1 ) . However , non-13-hydroxylated GAs was under the detectable limits . The non-13-hydroxylated GAs are supposed to be the EUI substrates in rice and similar results were previously obtained when the endogenous non-13-hydroxylated GA levels in the EUI overexpressed line were measured even after an exogenous treatment with GA3 [24] . Gene expression analysis revealed the upregulation of EUI in the OsAP2-39 transgenic rice . EUI encodes an enzyme that deactivates GA by catalyzing 16α , 17-epoxidation of non-13-hydroxylated GAs . At the same time , the transgenic lines have a higher endogenous ABA level than the wild-type . Since the overexpression of EUI is associated with a high level of ABA in the transgenic lines , the physiological relationship which links EUI with ABA was tested . Wild-type rice plants were sprayed with 10 µM ABA and the expression of EUI in leaves after 1 , 6 , 24 hours of ABA application was measured using qRT-PCR . The qRT-PCR results revealed that ABA induces EUI with a maximum level of expression after 6 hours of ABA treatment ( Figure 8 ) . Consistent with this result , sequence analysis of the EUI promoter using the Plant Cis-acting Regulatory DNA Elements ( PLACE , http://www . dna . affrc . go . jp/PLACE/signalscan . html ) showed the presence of one ABA Response Element ( ABR ) motif ( CACGTG ) in the EUI promoter at −2355 bp from the ATG start codon . These results show that a high endogenous ABA level is responsible at least in part for the EUI induction in the transgenic rice line and would explain how ABA is able to reduce bioactive GAs . After exogenous ABA treatment , OsAP2-39 is down regulated demonstrating a feed back mechanism leading to a reduction in the endogenous production of ABA ( Figure 8 ) . The gene expression analysis revealed that OsNCED-1 and EUI are upregulated in the OsAP2-39 transgenic lines . In order to investigate the mechanism of OsNCED-1 and EUI upregulation , the DNA sequences corresponding to the promoters of the both genes was analyzed using the PLACE software . The results showed that the OsNCED-1 promoter has 3 GCC sequence motifs located at 610 , 742 , and 1027 bp from the first ATG codon of the cDNA . Likewise , the sequence analysis showed that the EUI promoter has one GCC box located at 2488 bp from the first ATG codon of the cDNA . This motif is usually considered to be a binding box for AP2 transcription factors and therefore is a potential binding site for OsAP2-39 . To check the possibility that the OsAP2-39 protein binds to the GCC-box in vitro , recombinant OsAP2-39 protein was produced in Escherichia coli ( E . coli ) and used for Electrophoretic Mobility Shift Assays ( EMSA ) . The results demonstrate that OsAP2-39 strongly binds to the OsNCED-1 promoter sequence containing the GCC box motif . Substitution of the GCC box with poly adenine and thiamine sequence ( 5′-ATATAT-3′ ) inhibited the OsAP2-39-binding capacity to the DNA sequence ( Figure 9 ) . This result indicates an in vitro binding specificity of the OsAP2-39 to the GCC DNA motif . In order to investigate a direct relationship between the OsAP2-39 protein and OsNCED-1 and EUI gene expression , a transcription activation assay using a transient gene expression strategy was carried out using β-glucorinidase ( GUS ) as a reporter protein . A fusion construct was made either between the promoter region of OsNCED-1 or EUI and the GUS cDNA . In a separate construct , OsAP2-39 cDNA was cloned under the control of the 35S constitutive promoter and used as a transcription activator in the experiment . The pJD312 containing the firefly ( Photinus pyralis ) luciferase cDNA driven by the CaMV 35S promoter was used as the loading DNA control and the luciferase activity used to normalize the GUS activity in every sample ( Figure 10 A ) . DNA from the three vectors was co-transformed into the tobacco leaves using the particle bombardment method , with the empty vectors used as negative controls . Tobacco leaves were incubated 40 hours on Murashige and Skoog basal salt mixture ( MS ) solid media supplemented with ABA or GA at room temperature before protein from the leaves was isolated and used in the quantitative GUS and luciferase assays . The results demonstrated that OsAP2-39 slightly activates the expression of OsNCED-1 in tobacco epidermal cells when it is incubated on MS hormone-free medium . However , when the MS was supplemented with 100 µM GA , the OsAP2-39 was able to induce OsNCED-1 by almost 8 fold compared with the control experiment ( Figure 10B ) . Incubation of the bombarded tobacco leaves on MS media containing 10 µM ABA induces the expression of OsNCED-1 in the absence of the OsAP2-39 activator protein . However , in the presence of the OsAP2-39 , the expression of OsNCED-1 was reduced to 1/3 when compared with the control experiment . Interestingly , the results also showed that OsAP2-39 is able to highly activate the EUI promoter in the tobacco cells if incubated on hormone-free MS medium . In addition , the results demonstrated that EUI is induced by ABA and this induction was reduced in the presence of OsAP2-39 ( Figure 10B ) . Together these results indicate that OsAP2-39 directly regulates the expression of both OsNCED-1 and EUI and that this regulation is modulated by other factors induced by ABA and GA . OsAP2-39 was found to be more active in upregulating the OsNCED-1 gene in a high GA environment , which would lead to an increase in the ABA content . Rice lines transformed with the OsAP2-39 gene have fewer filled seeds in the spiklets ( 11A–B ) . Therefore it was of interest to determine whether the OsAP2-39 gene affects the pollination and fertilization processes in the flower . This was analyzed by investigating pollen grain morphology and viability . Compared to the wild-type , the results showed that OsAP2-39 overexpressing plants produced slightly smaller pollen grains and had a higher percentage with an irregular shape ( Figure 11C–11D ) . Similar observations were previously obtained when rice was treated with both cold and ABA [42] . In addition , a low active GA level due to EUI overexpression also leads to inhibited seed production in the transgenic lines [24] . This fact highlights the contribution of ABA and GA in this phenotype . Our findings that OsAP2-39 affects pollen grain morphology is consistent with the rice microarray data available through the public GENEVESTIGATOR database [43] which demonstrates that OsAP2-39 is highly expressed in rice anthers . Microarray and qRT-PCR data presented in this work showed that OsAP2-39 is expressed in the root at the early booting stage ( Figure S3 ) , when the ABA level is elevated in some grass plants such as barley [44] . In order to confirm the site of OsNUE39 expression in the plant tissue , the OsAP2-39 promoter was fused to the GUS reporter gene and transformed into Arabidopsis wild-type plants ( Figure S4 ) . Histochemical staining of GUS showed that OsNUE39 is predominantly expressed in the roots of the seedling ( Figure S4A ) , roots of adult plants ( Figure S4B ) ; and in the pollen grains ( Figure 4SD–4SF ) . This result is consistent with the microarray and RT-PCR data obtained from rice tissues . The plant hormone ABA regulates tolerance to environmental stresses such as drought and cold . In order to study the influence of high ABA on stress tolerance , the transgenic OsAP2-39 and wild-type plants were treated under cold and water stress conditions . While cold treatment did not show any specific effect on the transgenic lines , leaves of the transgenic lines are more susceptible to dehydration conditions than the wild-type . Drought tolerance experiments were carried out following the procedures described earlier by Yu et al . 2008 [45] . When plants of the two genotypes were grown in two different pots under water deprivation , wild-type plant dried faster than the OsAP2-39 overexpression line probably because of their large biomass which normally reflects a higher transpiration rate ( Figure S5 ) . When the two genotypes were grown in the same pot under water deprivation , the wild-type was able to grow for a longer time than the OsAP2-39 overexpression line likely due to their larger root system ( Figure 12A ) . Therefore it was difficult to reach any definitive conclusions from these two experiments . In order to better clarify this issue , an excised leaf water loss assay was done and the results showed that the transgenic OsAP2-39 lost water faster than does the wild-type ( Figure 12B ) indicating that OsAP2-39 has a lower leaf dehydration tolerance than the wild-type . It has been shown that dehydration induces ABA synthesis in plants . In order to determine the effect of dehydration on OsAP2-39 and OsNCED-1 expression , RNA was extracted from rice leaves dehydrated for 2 h and tested using qRT-PCR . The results show that OsAP2-39 and OsNCED-1 are highly induced by dehydration ( Figure 12C ) and this in turn would lead to an increase in ABA synthesis in that tissue . However , given that the overexpressing OsAP2-39 lines are not more resistant to dehydration implies that the increased production of ABA is not sufficient for drought tolerance . Despite the fact that a high ABA level is normally associated with stomatal closure and therefore drought tolerance , it is possible that the guard cells in the OsAP2-39 lines did not also have a higher ABA content .
Phytohormones regulate plant growth and development through a complex set of interactions . ABA and GA represent an example of a multidimensional and antagonistic relationship which has been studied over the last few decades . However , many important aspects of this relationship remain undiscovered . Here we demonstrate that a transcription factor containing the AP2 DNA–binding domain ( OsAP2-39 ) regulates ABA and GA crosstalk and homeostasis in rice . A hypothetical mechanism by which OsAP2-39 controls active ABA and GA levels is shown in Figure 13 . Overexpression of this transcription factor leads to an increase in the ABA content , which in turn reduces plant biomass and delays development . The mechanism by which OsAP2-39 controls the active ABA and GA is complicated and affected by the hormonal status in the tissue . OsAP2-39 slightly increases the expression of the OsNCED-1 , which is an ABA biosynthetic gene ( Figure 10B ) , in a hormone free environment . However , a high GA content leads to the upregulation of its expression by about 8 fold ( Figure 10B ) . As a result , high GA in turn activates the expression of OsNCED-1 which has been shown to be directly proportional to the ABA content in rice as observed in this study and also in other plant species where OsNCED-1 orthologues showed the same effect [46] , [47] . OsAP2-39 can induce EUI expression which in turn has previously been shown to reduce the bioactive forms of GAs [22]–[24] . In this study we found that this can occur through two different pathways . The first is an increased ABA content and the second is through the direct regulation of OsAP2-39 . As a result of these two pathways , the bioactive form of GA decreases and of ABA increases . In order to retain homeostasis , high ABA content inhibits OsAP2-39 gene expression . This set of mechanisms represents one of the pathways in which GA and ABA interact and communicate . In support of this hypothesis , the transactivation results in tobacco cells are consistent with the expression behaviour of the OsNCED-1 and EUI in rice leaves when they are subjected to exogenous application of ABA and GA . The increase in the active ABA level might be due to the inhibition of the ABA catabolism pathway . However the microarray data from the OsAP2-39 overexpression line did not show any expression alteration in the genes involved in this process . Overexpression of the OsAP2-39 leads to changes in the expression pattern of a large number of genes ( Table S1 ) . Some of these are due to the direct action of this transcription factor . For example , OsNCED-1 and EUI are both directly regulated by OsAP2-39 ( Figure 10 ) . This in turn would lead to an alteration in hormonal balance and change the level of expression of many other genes which together causes the pleiotropic phenotype present in these lines . This includes shorter internodes , a delay in flowering , smaller root mass , decreased tiller number and a lower seed yield . All of these phenotypes can be explained by the increase in ABA content and decrease in active GA except for the decreased tiller number . Information obtained from the global gene expression analysis provides a possible mechanism for this decrease in tiller number . These lines had a decreased level of the putative auxin-responsive proteins ( Os02g0769100 , Os10g0510500 and Os01g0753500 ) and also the upregulation of a gene ( Os02g0221900 ) which encodes a protein with similarity to MORE AXILLARY BRANCHES-1 ( MAX1 ) , a protein that regulates the rate of polar auxin transport in Arabidopsis [48]–[51] . Mutations within the MAX1 gene in Arabidopsis increase the number of axillary branches; and overexpression of this gene causes less axillary branches [49] . Therefore , one can hypothesize a similar situation for the OsAP2-39 overexpressing lines and a possible role for auxin transportation and signaling in the tillering phenotype akin to that seen in Arabidopsis for the auxiliary branching trait . Certainly , the idea that the axillary branching mechanisms in both rice and Arabidopsis are controlled by a common pathway is a reasonable one . The MONOCULM 1 ( MOC1 ) gene has been characterized and was found to control the rice axillary branches [52] . However , the expression level of this gene was not affected in the OsAP2-39 overexpressing lines indicating a different mechanism for the alteration in axillary branching in these transgenic lines . Given the notion that OsAP2-39 is not found in Arabidopsis , and so far only in the rice and maize sequence databases , it would not be surprising if this gene controls ABA and GA levels and axillary branching via a novel mechanism . A high ABA content is frequently associated with cold and drought tolerance in plants . When the transgenic and wild-type lines were grown together under water deprivation , the transgenic lines were more susceptible to water stress . These observations were confirmed when transgenic excised leaves lost water faster than did wild-type leaves . The constitutive overexpression of OsAP2-39 leads to changes in pollen grain morphology . This may explain the low seed yield in the transgenic lines . In fact , treatment of rice panicles with ABA induces pollen sterility and subsequently causes a significant reduction in grain number due to premature spikelet abortion [42] , [53] . Therefore , the fact that the transgenic lines overexpressing OsAP2-39 have a higher ABA content is consistent with the phenotype obtained where seed yield was reduced by about 80% . A low seed yield has been also observed when EUI gene was overexpressed in rice plants [24] . This result further supports the relationship between ABA synthesis and GA catabolism in rice . In conclusion , overexpression of the OsAP2-39 gene leads to a range of altered phenotypes that reduce biomass and seed yield including shorter internodes , delayed flowering , and lower tiller number . There are a large number of genes whose expression is altered in these lines . Some of these including key control genes regulating active ABA and GA levels are regulated directly by this transcription factor while many others no doubt are altered by changes in hormone levels . This work demonstrates for the first time a relationship between ABA biosynthesis and GA catabolic genes in rice . This relationship links the production of ABA and the inhibition of active GA and thus , provides a direct link in the antagonistic interactions between these hormones .
Rice ( Oryza sativa L . Kaybonnet ) was grown in a growth chamber with a 16 h light cycle , at 29°C during the day and 23°C during the night . Humidity was maintained at 70% . Plants were grown in pots containing 75% vermiculite and 25% peat moss and watered weekly with a nutrient solution [54] . Four-week old rice plants were treated with 100 µM GA or 10 µM ABA twice a week constitutively for four weeks . The construct for OsAP2-39 overexpressing was made using the maize ubiquitin promoter . OsAP2-39 transgenic rice lines were generated using Agrobacterium-mediated transformation method and positively transgenic plants were selected using Phosphomannose isomerase ( PMI ) . DNA sequence of low similarity to other rice genes located at the 3′ end ( 314–664 bp ) of OsAP2-39 was amplified by PCR using the following primers: AP2SiRNAF ( 5′- CACCTCGTCAGCCCGACCAGCAGCACG-3′ ) and AP2SiRNAR ( 5′- CTCCTCGATCGGCGGCGGCAG-3′ ) , cloned into TOPO pENTER vector ( Invitrogene ) , and the inverted DNA sequences separated by a GUS intron sequence were generated by site specific recombination method in the pANDA binary vector [55] down stream the maize ubiquitin promoter using the Gateway LR Clonase Enzyme Mix ( Invitrogene ) . Transgenic rice lines were obtained using Agrobacterium-mediated transformation and the positive lines were selected according to Miki et al . [56] . Roots of the wild type and transgenic plants were collected from three weeks old plants growing in turface supplemented with a full slow realize fertilizer ( 1 g/plant ) . Roots were scanned and analyzed using the WinRHIZO software ( v . 5 . 0 , Regent Instruments , Inc . , Quebec , QC , Canada ) . Five leaves from two OsAP2-39 transgenic rice lines and also from wild-type were pooled , freeze dried and the ABA contents were quantified at the hormone profile laboratory in the National Research Council , Plant Biotechnology Institute ( NRC-PBI ) Saskatoon , Canada and the method described by Chiwocha et al . [57] . Endogenous GA analysis using GC-MS was carried out using a MAT95XP mass spectrometer according to the previously published protocol [58] . The BLAST search program ( http://www . ncbi . nlm . nih . gov/BLAST/ ) was used to look for protein sequences homologous to OsAP2-39 and map the protein domains . Rice sequences with highest BLAST homology score were downloaded and used for the phylogenetic analysis using the Molecular Evolutionary Genetics Analysis ( MEGA4 ) software [59] . The neighbor-joining tree was generated with the Poisson correction method using the same software . Bootstrap replication ( 1000 replications ) was used for a statistical support for the nodes in the phylogenetic tree . The OSAP2-39 cDNA sequence was cloned in frame with the GFP protein under the control of the 35S promoter . The OSAP2-39 cDNA sequence was amplified by PCR using the following primer pair: APHindIII-ECoRIF: ( 5′-CCCAAGCTTATGGCTCCCAGGAACGC-3′ ) and APNdeI-ECoRR: ( 5′-CCGGAATTCCTACGCCTCCTCGATCG-3′ ) . After digestion with HindIII and EcoRI , the PCR products were cloned into pRLT2-GFP plasmid ( kindly provided from Dr . Robert Mullen , University of Guelph ) , amplified in E . coli , and transformed by particle bombardment into onion epidermal cells . A DNA sequence spanning the 2019 bp of the OsNUE39 promoter was amplified by PCR using the following primer pairs: promoterAp2F ( 5′-CGCGGATCCAATCTTGCTAAAATTTTGGCAAAG-3′ ) and promoterAp2R ( 5′-CATGCCATGGGTCCGTTCTTGTTCGGGTCG-3′ ) and cloned into the BamHI and NcoI sites upstream the GUS reporter gene of the pCAMBIA 3301 vector ( CAMBIA institute , Australia ) . Then the construct was stably transformed into the Wt Arabidopsis Col and various tissues of the transformed lines were assayed for GUS activity using the standard protocol . The recombinant full-length OsAP2-39 protein was expressed and purified using the Intein Mediated Purification with an Affinity Chitin-binding Tag system ( IMPACT ) ( New England Biolabs , ww . neb . com ) according to the manufacturers' instructions . The OsAP2-39 cDNA sequence was amplified by PCR using the following primer pair: APNdeI-ECoRF: ( 5′-GGAATTCCATATGGCTCCCAGGAACGCC-3′ ) and APNdeI-ECoRR: ( 5′-CCGGAATTCCTACGCCTCCTCGATCG-3′ ) . The PCR product and the pTYB12 plasmid ( New England Biolabs ) were digested with EcoRI and NdeI . After ligation , the construct was amplified in E . coli cells DH10B and transformed to the E . coli expression host strain ER2566 cells . Electrophoretic Mobility Shift Assay ( EMSA ) was carried out using the recombinant OsAP2-39 protein and the DNA products obtained using the PCR . The GGCGGC-box containing DNA sequence was amplified from the OsNCED-1 promoter using the following pair of primers: PMSA2bF: ( 5′-AATGTCTGCGGCGCCGGCGGC-3′ ) and PMSA2bR: ( 5′-AGTGTTCTGTTCCCCCGGGGAGATAAACCC-3′ ) . As a negative control , DNA in the EMSE reaction , the GCCGCC box motif sequence within the promoter was replaced by ( 5′-ATATAT-3′ ) using the site directed mutagenesis PCR and the following primer pair: PEMSA4F: ( 5′-AATGTCTGCGGCGCTATATACTGCGGTGTTTGTT-3′ ) and PEMSA4R: ( 5′-AGTGTTCTGTTCCCCCGGGGAGATAAACCC-3′ ) . The EMSA assay was carried out using the EMSA kit ( E33075 ) from Invitrogen ( Invitrogen , www . Invitrogen . com ) . After purifying the PCR product , a serial dilution of DNA ( 0 , 100 , 200 , 300 , and 400 ng ) were mixed and incubated with 30 ng of the purified OsAP2-39 recombinant protein according to the manufacturers' instructions . The DNA/protein complex samples were loaded into a Ready Gel TBE , gradient 4–20% polyacrelamide native gel ( Bio-rad Laboratories , www . bio-rad . com ) at 200 V for 45 minutes . The DNA in gel was stained using the SYBR Green provided in the same kit . As potential targets to OsAP2-39 transcription factor , DNA sequences corresponding to the OsNCED-1 and EUI promoters were cloned in an intron containing GUS reporter plasmid . The DNA sequence ( 1 kb upstream the ATG start codon of the cDNA ) of OsNCED-1 was amplified from the rice genomic DNA using the following promoter pair: OSNCEDF: ( 5′-CAATAACTGCAGGACGAGACCCTTTGCCG-3′ ) and OSNCED-1R: ( 5′-AGGGAATTCTCGATCGCACAACAATCTGAGC-3′ ) . The DNA sequence corresponding to the EUI promoter ( 2498 bp upstream the ATG start codon of the cDNA ) using the following primer pair: EUallF1: ( 5′-CTTTGCATTTGCCGCCGTGTT-3′ ) and EUallR1: ( 5′-GGCAGCCTACTCTCTCTTTCCCCG-3′ ) . After digestion with PstI and EcoRI , the PCR products were cloned into the pCAMBIA1391Z vector ( CAMBIA institute , Australia , www . cambia . au ) . OsAP2-39 induced by the 35S promoter in the pEGAD plasmid was used as an activator protein in the co-transformation transient expression analysis . To normalize the GUS activity values , firefly ( Photinus pyralis ) luciferase driven by 35S promoter in the pJD312 plasmid ( kindly donated from Dr . Virginia Walbot , Stanford University ) was used . Equal amounts of DNA from the different plasmids constructs was transformed by the particle bombardment to 4-weeks old tobacco ( Nicotiana plumbaginifolia ) leaves . After incubation for 40 hours at room temperature in the dark , the total protein was extracted from each sample and GUS and luciferase activities were measured . GUS activity was determined by measuring cleavage of β-glucuronidase substrate 4-methylumbelliferyl β-D-glucuronide ( MUG ) [60] . Luciferase activity was measured using the Luciferase Assay System kit ( Cat . E1500 ) ( Promega , www . promega . com ) following the manufacturers' instructions . Empty vectors were used as negative controls in this experiment . Double-stranded cDNAs was synthesized from 5 µg of total RNA from each sample . Labeled complementary RNA , synthesized from the cDNA was hybridized to Affymetrix rice whole genome array ( Cat . Number: 900601 ) . The hybridization signal of the arrays was obtained by the GeneChip scanner 3000 and quantified by MAS 5 . 0 ( Affymetrix ) . The probe set measurement was summarized as a value of weighted average of all probes in a set , subtracting bottom 5% of average intensity of the entire array using a custom algorithm . The overall intensity of all probe sets of each array was further scaled to a target intensity of 100 to enable direct comparison . Data was analyzed using GeneSpring software ( Agilent , CA , USA ) . The data was normalized with a default setting of the program , followed by gene filtering which required that each gene must have either a ‘P’ or ‘M’ flag in the three replicate samples . Genes with 2-fold change were identified first , and then ANOVA was used to identify significant genes ( Welch t-test p-value cutoff at 0 . 05 ) . For each genotype , leaf tissues from at least six plants were collected and pooled . The samples from 3 different pools were homogenized in liquid nitrogen prior to RNA isolation suing Tripure reagent ( Roche , http://www . roche-applied-science . com ) . cDNA was synthesized using the qScript cDNA Supermix ( Quanta Biosciences , http://www . quantabio . com/ ) . The qRT-PCR reactions were carried out using the SYBR Green PCR Master Mix ( Applied Biosystems , www3 . appliedbiosystems . com ) and the primers mentioned in Table S2 .
|
Hormones play an important role in controlling plant growth and development through a dynamic and complicated set of interactions . ABA and GA are well-known as antagonistic partners although the mechanism through which this occurs still needs further elucidation . In this project , we found that a transcription factor isolated from rice and coding for the AP2 domain ( OsAP2-39 ) directly controls a key ABA biosynthetic gene ( OsNCED-1 ) and also a gene that codes for a GA deactivation protein ( EUI ) . In addition , we show that ABA induces the expression of EUI which in turn would lead to GA deactivation . ABA also suppresses OsAP2-39 expression which would lead to a reduction in ABA synthesis . Therefore , OsAP2-39 links the ABA production and GA deactivation processes which results in ABA/GA balance and homeostasis .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"plant",
"biology/plant",
"growth",
"and",
"development",
"plant",
"biology/plant",
"genetics",
"and",
"gene",
"expression",
"plant",
"biology/plant",
"biochemistry",
"and",
"physiology"
] |
2010
|
The APETALA-2-Like Transcription Factor OsAP2-39 Controls Key Interactions between Abscisic Acid and Gibberellin in Rice
|
Evolutionary and ecosystem dynamics are often treated as different processes –operating at separate timescales– even if evidence reveals that rapid evolutionary changes can feed back into ecological interactions . A recent long-term field experiment has explicitly shown that communities of competing plant species can experience very fast phenotypic diversification , and that this gives rise to enhanced complementarity in resource exploitation and to enlarged ecosystem-level productivity . Here , we build on progress made in recent years in the integration of eco-evolutionary dynamics , and present a computational approach aimed at describing these empirical findings in detail . In particular we model a community of organisms of different but similar species evolving in time through mechanisms of birth , competition , sexual reproduction , descent with modification , and death . Based on simple rules , this model provides a rationalization for the emergence of rapid phenotypic diversification in species-rich communities . Furthermore , it also leads to non-trivial predictions about long-term phenotypic change and ecological interactions . Our results illustrate that the presence of highly specialized , non-competing species leads to very stable communities and reveals that phenotypically equivalent species occupying the same niche may emerge and coexist for very long times . Thus , the framework presented here provides a simple approach –complementing existing theories , but specifically devised to account for the specificities of the recent empirical findings for plant communities– to explain the collective emergence of diversification at a community level , and paves the way to further scrutinize the intimate entanglement of ecological and evolutionary processes , especially in species-rich communities .
Community ecology studies how the relationships among species and their environments affect biological diversity and its distribution , usually neglecting phenotypic , genetic and evolutionary changes [1–3] . In contrast , evolutionary biology focuses on genetic shifts , variation , differentiation , and selection , but –even if ecological interactions are well-recognized to profoundly affect evolution [4]– community processes are often neglected . Despite this apparent dichotomy , laboratory analyses of microbial communities and microcosms [5–14] as well as long-term field experiments with plant communities [15 , 16] and vertebrates [17 , 18] provide evidence that species can rapidly ( co ) evolve and that eco- and evolutionary processes can be deeply intertwined even over relatively short ( i . e . observable by individual researchers ) timescales [19] . Over the last two decades or so , the need to consider feedbacks between ecological and evolutionary processes has led many authors to develop a framework to merge together the two fields [20–41] . In particular , the development of quantitative trait models [42] and the theories of adaptive dynamics [43 , 44] and adaptive diversification [22–24 , 26–28 , 34] , reviewed in [40 , 42] , has largely contributed to the rationalization of eco-evolutionary dynamics , shedding light onto non-trivial phenomena such as sympatric speciation and evolutionary branching [40] . On the empirical side , the recent work by Zuppinger-Dingley et al . on long-term field experiments of vegetation dynamics appears to confirm many of the theoretical and observational predictions [45] . This study provided strong evidence for the emergence of rapid collective evolutionary changes , resulting from the selection for complementary character displacement and niche diversification , reducing the overall level of competition and significantly increasing the ecosystem productivity within a relatively short time . This result is not only important for understanding rapid collective evolution , but also for designing more efficient agricultural and preservation strategies . More specifically , in the experimental setup of Zuppinger-Dingley and colleagues , 12 plant species of different functional groups were grown for 8 years under field conditions either as monocultures or as part of biodiverse communities . Collecting plants ( seedlings and cutlings ) from these fields , propagating them in the laboratory , and assembling their offspring in new communities , it was possible to quantify the differences between laboratory mixtures consisting of plants with a history of isolation ( i . e . from monocultures ) and plants from biodiverse fields . While the former maintained essentially their original phenotypes , the latter turned out to experience significant complementary trait shifts –e . g . in plant height , leaf thickness , etc . – which are strongly suggestive of a selection for phenotypic and niche differentiation [41] ( see Fig 1 therein ) . Furthermore , there were strong net biodiversity effects [46] , meaning that the relative increase in total biomass production in laboratory mixtures with respect to laboratory monocultures was greater for plants from biodiverse plots than for plants coming from monocultures . These empirical results underscore the need for simple theoretical methodologies , in the spirit of the above-mentioned synthetic approaches [20 , 23 , 27 , 28 , 36 , 40 , 42 , 47 , 48] . These approaches should explain the community and evolutionary dynamics of complex and structured communities such as the ones analyzed in [45] . The phenotypic differentiation observed in the experiments of Zuppinger-Dingley et al . might be partially rationalized within the framework of relatively simple deterministic approaches to eco-evolution such as adaptive dynamics ( see e . g . [20 , 23 , 27 , 28 , 36 , 40 , 42 , 47] ) . In this context , diversification is the natural outcome of an adaptive/evolutionary process that increases fitness by decreasing competition through trait divergence . However , it is not obvious what would be the combined effects in this simplistic version of adaptive dynamics of introducing elements such as sexual reproduction , space , and multi-species interactions that could play an important role in shaping empirical observations . Moreover , questions such as whether phenotypic differentiation occurs both above and below the species level ( i . e . , within species or just between them ) , the possibility of long term coexistence of phenotypically equivalent species in the presence of strong competition ( i . e . , emergent neutrality ) , or the expected number of generations needed to observe significant evolutionary change remain unanswered and require a more detailed and specific modeling approach , within the framework of adaptive dynamics . Thus , our aim here is to contribute to the understanding of eco-evolutionary dynamics , emphasizing collective co-evolutionary aspects rather than focusing on individual species or pairs of them . For this purpose , we developed a simple computational framework –similar to existing approaches ( see Discussion ) – specifically devised at understanding the emerging phenomenology of the experiments of Zuppinger-Dingley et al . In particular , we propose an individual-based model , with spatial structure , stochasticity , sexual reproduction , mutation , multidimensional trait-dependent competition and , importantly , more than-two-species communities ( in particular , possibly owing to analytical difficulties , relatively limited work has been published about more than three-species communities , which is crucial to achieve a realistic integration of ecological and evolutionary dynamics for natural communities; see however [49–51] ) . Furthermore , our method is flexible enough as to be easily generalizable to other specific situations beyond plant communities and can rationalize the circumstances under which phenotypic diversification and niche specialization may emerge using simple , straightforward rules .
We construct a simple model which relies on both niche based approaches [52–54] and neutral theories [55–58] . The former prioritize trait differences and asymmetric competition , underscoring that coexisting species must differ in their eco-evolutionary trade-offs , i . e . , in the way they exploit diverse limiting resources , respond to environmental changes , etc . , with each trade-off or “niche” choice implying superiority under some conditions and inferiority under others [1 , 3 , 53 , 54] . Conversely , neutral theory ignores such asymmetric interactions by making the radical assumption of species equivalence , and focuses on the effects of demographic processes such as birth , death and migration . Here , we adopt the view shared by various authors [36 , 59–61] that niche-based and neutral theories are complementary extreme views . In what follows , we present a simple model that requires of both neutral and niche-based elements . In particular , our model incorporates trade-off-based features such as the existence of heritable phenotypic traits that characterize each single individual . However , the impact of these traits on individual fitness is controlled by a model parameter , that can be tuned to make the process more or less dependent on competition , in the limit even mimicking neutral ( or “symmetric” ) theories [55 , 56] . The traits of each single individual are determined by quantitative phenotypic values that can be regarded as the investment in specific functional organs . For instance , the traits could represent the proportion of biomass devoted to exploit soil nutrients ( roots ) , light ( leaves and stems ) , and to attract pollinators and capture pollen ( flowers; see Fig 1 ) . We then assume a hard limit –constant across generations– to the amount of resources that can be devoted to generate the phenotype , i . e . it is impossible to increase all phenotypic values simultaneously . Thus each individual is constrained to make specific trade-offs in the way it exploits resources . Because similar values in the trade-off space entail comparable exploitation of the same resource ( e . g . , water , light or pollinators ) similar individuals experience higher levels of competition , which translates into a lower fitness . This can be regarded as a frequency dependent selection mechanism providing an adaptive advantage to exceptional individuals , able to exploit available resources . Therefore , the ecological processes of competition , reproduction , and selection lead to evolutionary shifts in the distribution of phenotypic traits which feed back into community processes , giving rise to integrated eco-evolutionary dynamics . The basic components of the model are as follows ( further details are deferred to the Methods section ) . We consider a community of individuals of S different species , that are determined initially by mating barriers ( i . e . a species is defined as a set of individuals that can produce fertile offspring [62] ) . Each individual occupies a position in physical space ( represented as a saturated square lattice ) and is characterized by the label of the species to which it belongs and a set of intrinsic parameters ( i . e . trait values ) , specifying its coordinates in the “trade-off space” as sketched in Fig 1 ( see also [41 , 63] . All positions within the trade-off space are assumed to be equally favorable a priori . In what follows , we make a perfect identification between the trade-offs of a given individual and its phenotypic traits , which also determine the “niche” occupied by each individual . In principle , each individual , regardless of its species , can occupy any positition in the trade-off space . Positions near the center of the trade-off space ( Fig 1 ) correspond to phenotypes with similar use of the different resources ( i . e . , “generalists” ) , while individuals near the corners specialize in the exploitation of a given resource ( “specialists” ) . Individuals are subjected to the processes of birth , competition for resources , reproduction , descent with modification , and death . Individuals are assumed to undergo sexual reproduction , as in the experiments of [45] ( implementations with asexual reproduction are discussed later ) ; they are considered to be semelparous , so that after one simulation time step ( i . e , a reproductive cycle ) they all die and are replaced by a new generation . Importantly , demographic processes are strongly dependent on phenotypic values . In particular , the main niche-based hypothesis is that individual organisms with a better “performance” are more likely to reproduce than poorly performing ones . To quantify the notion of “performance” , we rely on classical concepts such as limiting similarity , competitive exclusion principle and niche overlap hypothesis [64 , 65] , which posit that in order to avoid competition , similar species must differ in their phenotypes . More specifically , our model assumes that the performance of a given individual increases with its trait “complementarity” to its spatial neighbors [65] , as quantified by its averaged distance to them in trade-off space ( see Methods ) ; i . e . the larger the phenotypic similarity among neighbors , the stronger the competition , and the worse their performance . Although the performance of a given individual depends on its complementarity with its neighbors , the model is symmetric among species and phenotypes; performance is blind to species labels and does not depend on the specific location in the trade-off space . The reproduction probability or performance of any given individual is mediated by a parameter β which characterizes the global level of competitive stress in the environment ( see Methods ) . In the limit of no competition , β = 0 , the dynamics become blind to phenotypic values and can be regarded as fully neutral , while in the opposite limit of extremely competitive environments , β → ∞ , niche effects are maximal and a relatively small enhancement of trait complementarity induces a huge competitive advantage . Finally , a mother selected as described in the competition process is assumed to be fertilized by a conspecific “father” in the population ( interspecies hybridization is not considered here ) which is also selected with the same reproduction probability function based on its performance . The offspring inherits its traits from both parents , with admixture and some degree of variation μ ( see Fig 1 and Methods ) . This process is iterated for all lattice sites and for an arbitrarily large number of reproductive cycles , resulting in a redistribution of species both in physical and in trade-off space . Species can possibly go extinct as a consequence of the dynamics . In this version of the model , speciation is not considered , though it could be easily implemented by establishing a dependence of mating on phenotypic similarity , making reproduction between sufficiently different individuals impossible [37] . Simulations are started with individuals of S different species ( e . g . S = 16 ) randomly distributed in space . In the initial conditions , the traits of all individuals are a sample from a common Gaussian distribution centered around the center of the simplex ( note that as shown in the S5 Appendix in S1 Text , results do not depend on the particular choice of initial conditions ) . Statistical patterns emerging from the eco-evolutionary dynamics described above are analyzed as a function of the number of generations and as a function of the number of species S , for different values of the two free parameters: the overall level of competition β and the variability of inherited traits μ . Results are illustrated in Fig 2 showing ( i ) phenotypic diagrams ( top row ) specifying the position of each single individual and its species in the trade-off space for different parameter values and evolutionary times ( ii ) ; values of complementarity for all individuals ( central row ) in the trade-off space , and ( iii ) the spatial distribution of individuals and species ( bottom row ) . Finally , several biodiversity indices are reported in Fig 3 .
In the present paper , we have developed a parsimonious modeling approach to integrate important ecological and evolutionary processes . In particular , we focused on understanding rapid phenotypic diversification observed in complex biological communities of plants such as those recently reported by Zuppinger-Dingley et al . in long-term field experiments [41 , 45] . Our model blends standard community processes , such as reproduction , competition or death , with evolutionary change ( e . g . , descent with modification ) ; i . e . community and evolutionary dynamics are coupled together , feeding back into each other . Over the last decades , attempts to integrate ecological and evolutionary dynamics have been the goal of many studies ( see e . g . [16 , 22–34 , 37 , 39 , 40 , 48 , 63] ) . In particular , a basic algorithm for modeling eco-evolutionary dynamics as a stochastic process of birth with mutation , interaction , and death was proposed in [22] and much work has been developed afterwards to incorporate elements such as spatial effects and different types of interspecies interactions [28] . Rather than providing a radically different framework , our model constitutes a blend of other modeling approaches in the literature of eco-evolutionary processes , and in fact it shares many ingredients with other precedent works , specially with the theory of adaptive dynamics [40 , 42] . For instance , Gravel et al . [36] also considered a spatially-explicit individual-based model with trait-dependent competition . However , our work has been specifically devised to shed light on the experimental findings of Zuppinger et al . [45] , and puts the emphasis on communities with arbitrarily large number of species , while usually the focus is on the ( co- ) evolution of pairs of species ( e . g . predator-prey , host-parasite , etc . ) or speciation/radiation of individual species . Finally , our modelling approach is sufficiently general as to be flexible to be adapted to other situations with slightly different ingredients . We explored some of these possible extensions in some Appendices ( S3 , S4 , S7 , S8 , S9 ) in S1 Text ( e . g long-distance dispersal , asexual reproduction , etc . ) , but other studies can be built upon the work laid here in a relatively simple way . The present model relies on a number of specific assumptions , two of which are essential in that they couple community and evolutionary dynamics: i . e . ( i ) demographic processes are controlled by competition for resources which is mediated by phenotypic traits and ( ii ) successful individuals are more likely to transmit their phenotypes to the next generation with some degree of variation . These two ingredients are critical for the emerging phenomenology . For instance , in the absence of competition ( i . e . β = 0 ) reproduction probabilities are identical for all individuals , implying that the model becomes neutral , and the evolutionary force leading to species differentiation vanishes ( see S4 Appendices in S1 Text ) . On the other hand , variation in inherited traits is necessary to allow for the emergence of slightly different new phenotypes and the emergence of drifts in trade-off-space . Although these constraints might be regarded as limiting , we deem them biologically realistic and do not think they hamper the predictive power of our model . Most of the remaining ingredients , such as the existence of a saturated landscape , semelparity ( i . e . non-overlapping generations ) , the specific form in which we implemented initial conditions , competition , dispersion , selection , inheritance linked to phenotypic characters rather than to a genotypic codification , etc . can be modified without substantially affecting the results . This flexibility could make the description of other type of communities possible with minimal model variations . Similarly , the model could be extended to incorporate phenotype-dependent reproductive barriers ( and thus speciation ) and the possibility of interspecies hybridization by making reproduction a function of phenotypic distance and relaxing its dependency on species labels . In addition to rapid phenotypic diversification , the experiments of Zuppinger et al . found an enhancement of the overall productivity in mixtures of diverse plants with respect to monocultures of the same plants [45] . Our model cannot be used to directly quantify such “biodiversity effects” [46] , as we assume a fully saturated landscape and there is no variable that accounts for total biomass production . However , in principle , under the hypothesis that larger trait complementarities correlate with greater resource capture and biomass production , the observed increase of relative complementarity in mixtures ( see Fig 3 ) could be used as a proxy for biodiversity effects . Observe , nonetheless , that the previous assumption might by wrong ( or incomplete ) as productivity can be profoundly affected by other factors such as , for instance , positive interactions between similar species , not modeled here , and more sophisticated approaches –see [73–77]– are necessary to validate this hypothesis . In the future we plan to modify our model to represent non-saturated landscapes and more detailed ecological dynamics , allowing for explicit analyses of biodiversity-productivity relationships . Beyond explaining most of the empirical observations in [45] , our model leads to some far-reaching predictions ( some of them already shared by existing theories ) ; one of the most remarkable ones is that optimal exploitation of resources comes about when the full community evolves into a reduced number of highly specialized species –the exact number depending on the dimensions of the trade-off space– that coexist in highly dispersed and intermixed populations . Such specialization might be unrealistic in the case in which all traits in trade-off space are essential for survival , and thus the convergence toward perfect specialization is capped . In any case , this result is congruent with the niche dimension hypothesis [78] , that postulates that a greater diversity of niches entails a greater diversity of species , i . e . a larger number of limiting factors ( and thus of possible trade-offs ) leads to richer communities [79] . However , this outcome might be affected by perturbations ( migration , environmental variability , etc ) which could be easily implemented in our model , and could prevent real communities from reaching the asymptotic steady state predicted here . It is also noteworthy that the resulting highly specialized species can be phenotypically equivalent , and a set of them can occupy almost identical locations in the trade-off space . Such species equivalence appears spontaneously , and supports the views expressed by other authors that “emergent neutrality” is a property of many ecosystems [80–82] . In future work we will explore the possibility of phase transitions separating an ecological regime based on the coexistence of multiple highly specialized species from an ecosystem dominated by generalists and the conditions under which each regime emerges . Beyond phenotype-dependent mating , upcoming studies will extend our approach to address communities where collective diversification phenomena based on both competition and cooperation are known to emerge ( see e . g . [13] ) , as well as investigate the evolution of communities with distinct types of interacting species such as plant-pollinator mutualistic networks . This research will hopefully complement the existing literature and help highlighting the universal and entangled nature of eco-evolutionary processes .
We implemented computer simulations in which each individual plant , i , is fully characterized by ( see also Fig 1 ) : ( i ) a label identifying its species , ( ii ) its coordinates in the physical space , and ( iii ) a set of real numbers specifying its phenotypic traits . In these simulations , time can be implemented either as discrete/synchronous updating or continuous/sequential updating without significantly altering the results . Species– , we consider a fixed number of species , labeled from 1 to S; while the emergence of new species is not considered here , some of them may become extinct along the course of evolution . Physical space– We consider a two-dimensional homogeneous physical space described by a L × L square lattice , assumed to be saturated at all times , in which the neighborhood of each individuals is determined by the closest K sites ( in our simulations , we took L = 64 and K = 24 ) . Phenotypic traits and trade-off space– As energy and resources are limited , each individual plant needs to make specific choices/trade-offs on how to allocate different functions . The way we implement the “trade-off space” is inspired in the field of multi-constraint ( non-parametric ) optimization that it is called Pareto optimal front/surface [83]; it includes the set of possible solutions such that none of the functions can be improved without degrading some other . Thus , the phenotype of any individual can be represented as a trade-off equilibrium , a point in this space and encapsulated in a set of real numbers T = ( T1 , T2 , … , Tn ) ( all of them in the interval [0 , 1] ) , such that ∑ k = 1 n T k = 1 where n is the number of trade-offs ( see Fig 1 and [41] ) . All positions within the trade-off space are equivalent a priori , although this requirement can be relaxed . Competition for resources– The trait “complementarity” between two individuals i and j is quantified as their distance in the trade-off space: c i j = ∑ k = 1 n | T k ( i ) - T k ( j ) | / n , which does not depend on species labels . The averaged complementarity , ( or simply “complementarity” ) over all the neighbors j of individual i is Ci = ∑j ∈ n . n . ( i ) cij/K . Complementarity-based dynamics– Each timestep , every individual is removed from the population; the resulting vacant site i is replaced by an offspring of a potential mother plant j which is selected from the list of K local neighbors of the vacant site with a given probability Pmother ( j ) . This probability controls the dynamical process; we assume it to increase as the mother’s trait complementarity Cj increases ( i . e . as its effective competitive stress diminishes ) : P mother ( j ) = e β C j / ∑ j ′ ∈ n . n . ( i ) e β C j ′ , where the sum runs over the set of K neighbors of i; eβCj is the “performance” of individual j and β is a tunable “competition parameter” controlling the overall level of competitive stress in the community . Once the mother has been selected , the father is randomly chosen from all its conspecific individuals l in the community , with a probability proportional to their performance , eβCl . In other words , individuals with lower competition pressure are more likely to sire descendants both as females and as males . Inheritance , admixture and variation of phenotypes– The traits of each single offspring are a stochastic interpolation of those of both parents with the possibility of variation: T new k = η T mother k + ( 1 - η ) T father k + ξ k , for k = 1 , … , n , where η is a random variable ( uniformly distributed in [0 , 1] ) allowing for different levels of admixture for each offspring , and ξk are ( Gaussian ) zero-mean random variables with standard deviation μ , a key parameter that characterizes the variability of inherited traits . To preserve the overall constraints Tk ∈ [0 , 1] and ∑k Tk = 1 , mutations are generated as ξk = ( rk − rk+1 ) , where {r1 = rn+1 , … , rn} are independent Gaussian random variables with zero-mean and standard deviation μ / 2; in the rare case that T new k < 0 ( resp . >1 ) , we set it to 0 ( resp . to 1 ) and added the truncated difference to another random trait . The centroid of species s is B ( s ) = {B1 ( s ) , … , Bn ( s ) } , with Bk ( s ) = ∑i Tk ( i ) /ns for each trait k , where i runs over the ns individuals of species s . Interspecies distance: is the distance between the centroids of two different species s and s′ in the trade-off space ds , s′ = ∑k|Bk ( s ) − Bk ( s′ ) |/n , averaged over all surviving species . Intraspecific distance is the average distance in trade-off space between all pairs of individuals of a given species s , ds = ∑i , j ∈ s cij/ns ( ns − 1 ) averaged over all surviving species . Local complementarity is the mean complementarity of individuals to their spatial neighbors , LC = ∑i ( ∑j ∈ n . n . ( i ) cij/K ) /N where N is the total number of individuals and K is the number of local neighbors . Global complementarity is the complementarity averaged over all pairs of individuals regardless of their relative positions in physical space , GC = ∑i , j ≠ i cij/ ( N ( N − 1 ) ) . Similarly GCinter is the averaged complementarity between individuals of different species and GCintra is the averaged complementarity between conspecific individuals . In the case of monocultures , GCinter ( S = 1 ) is measured from two different/independent realizations . Relative complementarity , RC = GCinter − GCintra , is a measure of the averaged difference in the level of competition between randomly sampled conspecific and non-conspecific individuals . Moran’s index: is a measure of spatial correlations between neighbors; it is negative when neighbors tend to belong to different species ( see S2 Appendix in S1 Text ) .
|
Population ecology and evolutionary biology have been traditionally studied as separate disciplines , even if feedbacks between community and evolutionary processes are known to exist , having been empirically characterized in recent years in different types of communities ( from microbes to plants and vertebrates ) , and theoretically analyzed with novel and powerful mathematical tools . Recent long-term field experiments with plants have proven that rapid co-evolution and diversification of species traits results in an overall enhancement of the ecosystem productivity , with important consequences for agriculture and conservation . Here , we propose a relatively simple computational eco-evolutionary model specifically devised to describe rapid phenotypic diversification in this type of species-rich communities . Our model captures the main phenomenology observed experimentally , and it also makes non-trivial predictions for long term phenotypic change and ecological interactions , such as the stable coexistence of highly specialized species or the possible emergence of phenotypically equivalent species occupying the same niche . Finally , the model is easily generalizable to analyze different eco-evolutionary problems within a relatively simple and unified computational framework .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"community",
"ecology",
"ecological",
"metrics",
"ecology",
"and",
"environmental",
"sciences",
"ecological",
"niches",
"species",
"diversity",
"ecology",
"species",
"delimitation",
"phenotypes",
"speciation",
"genetics",
"biology",
"and",
"life",
"sciences",
"theoretical",
"ecology",
"evolutionary",
"biology",
"biodiversity",
"evolutionary",
"processes"
] |
2016
|
Eco-evolutionary Model of Rapid Phenotypic Diversification in Species-Rich Communities
|
Transposable elements ( TEs ) make up the majority of many plant genomes . Their transcription and transposition is controlled through siRNAs and epigenetic marks including DNA methylation . To dissect the interplay of siRNA–mediated regulation and TE evolution , and to examine how TE differences affect nearby gene expression , we investigated genome-wide differences in TEs , siRNAs , and gene expression among three Arabidopsis thaliana accessions . Both TE sequence polymorphisms and presence of linked TEs are positively correlated with intraspecific variation in gene expression . The expression of genes within 2 kb of conserved TEs is more stable than that of genes next to variant TEs harboring sequence polymorphisms . Polymorphism levels of TEs and closely linked adjacent genes are positively correlated as well . We also investigated the distribution of 24-nt-long siRNAs , which mediate TE repression . TEs targeted by uniquely mapping siRNAs are on average farther from coding genes , apparently because they more strongly suppress expression of adjacent genes . Furthermore , siRNAs , and especially uniquely mapping siRNAs , are enriched in TE regions missing in other accessions . Thus , targeting by uniquely mapping siRNAs appears to promote sequence deletions in TEs . Overall , our work indicates that siRNA–targeting of TEs may influence removal of sequences from the genome and hence evolution of gene expression in plants .
While transposable elements ( TEs ) constitute a large fraction of plant , animal and human genomes [1]–[3] , their contribution to genome size can change rapidly during evolutionary time . In some taxa , TEs have been responsible for two-fold differences in genome size that arose over a few million years or less . These rapid fluctuations , which may be due to TEs being either more active or more efficiently deleted in certain species , indicate that control of TEs can differ greatly between closely related plant species [4]–[7] . The balance between TE transpositions and selection against TEs is influenced by factors ranging from mating system to silencing by short interfering RNAs ( siRNAs ) and chromatin modification . Therefore the control of TE activity and the removal of transposed copies can be considered key factors in the evolution of genomes . TEs are often regarded as genomic parasites due to the potentially detrimental effects of insertional inactivation of genes and ectopic recombination of DNA [8] . Twenty-four nt long siRNAs are associated with most TEs as part of a ‘double-lock’ mechanism of siRNA-mediated DNA methylation that controls transposition via transcriptional repression , with a reinforcement loop between DNA methylation , histone methylation and siRNAs [reviewed in 9] . siRNAs are a robust proxy for DNA methylation at TEs , with unmethylated TEs generally lacking matching 24 nt siRNAs [10]–[13] . Most plant TEs have cytosine methylation at CG , CHG and CHH sites , but a quarter is unmethylated and a further 15% have atypical methylation patterns . In the TE-dense heterochromatin , DNA methylation can spread about 500 bp into neighboring unmethylated TEs [13] . In the euchromatin , methylation spreads from TEs to approximately 200 bp beyond the siRNA target sites [13] , consistent with the effect of siRNAs on expression of proximal genes dissipating by 400 bp [14] . siRNA-targeted , methylated TEs are , on average , located farther away from expressed genes than TEs that are not strongly methylated or associated with siRNAs [13] , [15] . As expected from this correlation , siRNA-targeted TEs have more effects on nearby gene expression than those without [14] , [15] . Most poorly methylated TEs are short and have few CG dinucleotides [13] . This indicates a progression over evolutionary time from TEs that are active and targeted by siRNA-mediated DNA methylation , to inactive , degenerate relics that have changed through deletions and nucleotide substitutions initiated by deamination of methylated cytosines . These inactive TEs are then no longer targeted by siRNA-mediated DNA methylation . Presumably because of interference with cis-regulatory elements , Arabidopsis TEs reduce the average expression levels of adjacent genes , although the distance over which these effects are noticeable varies between A . thaliana and A . lyrata [14] . Differences in TEs next to genes contribute to the divergence of gene expression levels between orthologs in these closely related species [14] , and gene expression is negatively correlated with the number of nearby siRNA-targeted , methylated TEs [15] . In the selfing species A . thaliana , TEs account for only a fifth of the genome [7] , [13] , [16] , making it relatively depauperate of TEs . Given that the A . thaliana genome is small relative to other members of the family and that its close relative A . lyrata , an outcrosser , contains approximately three times as many TEs [14] , deletion of TEs in A . thaliana is likely an ongoing , active process . In accordance with this hypothesis , intraspecific polymorphisms and deletions in A . thaliana are disproportionately located within TEs and , to a lesser extent , intergenic regions [17]–[19] . A reference-guided assembly approach has been applied to accurately characterize complex sequence variation in several A . thaliana accessions [19] . Here , we exploit this information to examine TE variants and their effect on the expression of nearby genes in three divergent accessions . We report that TEs are more likely to be located in polymorphic regions of the genome . Where TEs are present in less polymorphic regions , they also tend to be less polymorphic themselves . Although polymorphic TE variants are less abundantly targeted by siRNAs , uniquely mapping siRNAs targeting polymorphic TE variants are strongly correlated with the TE regions that vary between accessions . These findings suggest a link between the ability to tolerate TE insertions , siRNA-mediated silencing and purging of TEs by deletion .
We annotated the sets of genes and TEs in three A . thaliana accessions: Col-0 , Bur-0 and C24 [19] , [20] . For reference accession Col-0 , we used the TAIR9 annotation of TEs and protein-coding genes . Excluding centromeric sequences , 21 , 913 full-length and degenerate TEs and 26 , 541 genes were considered further . We built genome templates of Bur-0 and C24 from re-sequencing data using the SHORE pipeline [21] . The reference coordinates of TEs and genes were projected onto these genome templates , and variation in TEs and genes was determined based on single nucleotide polymorphisms ( SNPs ) , 1 to 3 bp insertions/deletions ( indels ) and larger deletions of 4 to 11 , 464 bp ( median 30 bp , mean 113 bp ) . Larger insertions were not included because of the high false-negative rate [17] . Comparison of polymorphism densities confirmed that coding regions were relatively depauperate of SNPs , indels and large deletions compared to intergenic regions and TEs ( binomial test , p[Coding Regions/Intergenic Region] = 0 and p[Coding Regions/TE] = 0 for SNPs , indels or large deletions ) . Large deletions were significantly over-represented in TEs compared to intergenic regions , while SNPs and indels were not ( Figure S1a; binomial test , p[TE/Intergenic Region] = 0 for large deletions ) . Over 6% of reference TEs differed by at least 10% of total length in each of the two accessions , Bur-0 and C24 , compared to Col-0 ( Figure 1a and Figure S2 ) . Almost all of this variation , 93% , was due to large deletions ( Figure S1b; for distribution of large deletion sizes see Figure S1c ) . We defined TEs with at least 10% variation by length ( SNPs , indels and larger deletions combined ) , but not completely missing in Bur-0 or C24 , as TE variants or VarTEs ( please also see Figure S3 for abbreviation definitions ) . Close to 40% of VarTEs were shared between Bur-0 and C24 ( Figure S4a ) . TE density is highest in and next to the centromeres , where there are few genes . The fraction of VarTEs and the average level of TE variation were higher in the pericentromeric regions than on the gene-dense chromosome arms ( Figure 1b; Mann-Whitney U [MWU] test , p<2×10−16 for Col-0 versus Bur-0/C24 , Table S1 and Figures S5 and S6 ) . To examine whether gene proximity biases TE variation across the chromosomes , we calculated the distance between TEs and protein-coding genes for Col-0 . TEs were separated into two subsets: TEs within 2 kb of any gene , subsequently called proximal TEs , and TEs at least 2 kb away from the closest gene , called distal TEs . Distal TEs were on average more variable than proximal TEs ( Figure 1c; Figures S7 and S8; MWU p[Col-0/Bur-0] = 0 . 001 , p[Col-0/C24]<6×10−5 ) . Proximity to protein-coding genes may therefore influence TE variation , consistent with TEs closer to genes likely being under stronger selective constraint [15] , [22] . The correlation between TE variation and proximity to genes was compared among TE superfamilies [23] , [24] . For non-centromeric TEs , LTR retrotransposons were more distal from genes , while no significant difference in distance to genes was observed for other TE superfamilies ( Table S2 ) . However , for proximal TEs there were differences among TE superfamilies in distance to genes and , as expected , TE superfamilies that are closer to genes ( e . g . CACTA , MITE ) were less variable than superfamilies located farther away from genes , e . g . non-LTR retrotransposons ( Table S2 ) . To investigate the link between TE and proximal gene variation , we examined whether TE variation and location correlated with the polymorphism level of neighboring genes . We used the small-scale mutations to calculate the polymorphism level of non-centromeric genes . For each accession , genes were separated into two subsets; TE+ genes included genes within 2 kb of a TE and genes with TEs anywhere within the transcribed region , while TE- genes were at least 2 kb from the closest TE ( Table S3 ) . To be conservative , any TEs in Bur-0 or C24 with predicted deletions of at least 10% of the reference length were annotated as deleted . TE+ genes were on average more polymorphic than TE− genes in each accession ( Figure 2a; MWU p<2×10−16 for Col-0 , Bur-0 and C24 ) . The same analysis was repeated for 80 resequenced A . thaliana accessions [17]; we could confirm the correlations observed with Bur-0 and C24 in these accessions . Since polymorphism levels vary enormously among gene families , we further investigated whether there is a correlation of TE proximity with gene family using small-scale mutations from the 80 A . thaliana accessions ( 20 , 61 ) , and Col-0 , C24 and Bur-0 . Genes from highly polymorphic families such as those encoding NBS-LRR , F-box and Cytochrome P450s proteins were , on average , closer to TEs in all accessions ( Figure S9; distance is negatively correlated with gene polymorphism , Spearman's ρ ( Col-0 ) = −0 . 11 , ρ ( Bur-0 ) = −0 . 11 , ρ ( C24 ) = −0 . 10; p<2×10−16 ) , including a higher proportion of genes having proximal TEs ( Figure S10 ) . TEs are therefore either more likely to insert into or near polymorphic genes , or are less efficiently purged from such regions . To further examine the effects of TE variants on proximal genes , we divided TE+ genes into two subsets: genes where flanking TEs were <10% variant ( Invariant TEs: InvTE ) among the three accessions ( InvTE+ genes ) , and genes where at least one flanking TE showed ≥10% sequence ( VarTE ) variation between accessions ( VarTE+ genes; Table S3 ) . Three quarters of VarTE+ genes were shared in comparisons between Col-0 and Bur-0 or Col-0 and C24 ( Figure S4b ) . The VarTE+ genes were on average more polymorphic than InvTE+ genes ( Figure 2b; MWU p = 0 . 005 ) , also in the 80 accessions dataset [17] . We conclude that TEs close to genes are less polymorphic , while genes close to polymorphic TEs are themselves more polymorphic . A correlation between polymorphism levels of TEs and nearby genes is insufficient to address whether this is a direct link as opposed to high directional selection pressure on the genomic region in general . To address this question , we therefore compared the polymorphism level of TEs , the flanking regions and nearby genes . TEs in highly polymorphic regions are themselves more polymorphic than TEs in regions of low divergence ( Figure S11a; binomial test , p = 0 ) , with the exception that TEs in highly polymorphic regions with nearby lowly polymorphic genes show a similar level of divergence as TEs in regions of low polymorphism with no coding genes . Moreover , TEs in gene-free regions show significantly higher divergence than TEs within 4 kb of a gene , especially if those genes are less polymorphic . TEs are generally more polymorphic than their flanking sequences ( binomial test , p = 0 ) , with the exception of TEs in highly polymorphic regions with lowly polymorphic gene . The results for large deletions ( Figure S11b ) are consistent with our observation from Figure S1 that large deletions are over-represented in TEs compared to intergenic regions . Notably , there is no significant difference in the level of small-scale mutations between TEs and flanking regions ( Figure S11c ) . Taken together , TE variation through large deletions shows a positive correlation with flanking region polymorphism level , but is also strongly influenced by the conservation and presence/absence of nearby genes . The frequency of large deletions is however generally higher in TEs than in the flanking regions , indicating positive selection for large deletions within TEs . Genes that are close to TEs ( TE+ genes ) tend to have a lower expression average than TE− genes in the Col-0 reference accession [15] . We set out to determine whether this was true for the accessions studied here as well . Gene expression was measured using Affymetrix tiling arrays and RNA extracted from floral tissue of each accession . We considered presence/absence of TEs in the flanking regions of genes , taking into account the number of linked TE insertions and the distance from each gene to the closest TE . We confirmed the reported pattern for Col-0 [15] , and found that it applies to Bur-0 and C24 as well . In all three accessions , genes with proximal TEs ( TE+ genes ) were on average expressed at lower levels than those without proximal TEs ( TE− genes; Figure 3a; MWU p<2×10−16 for Col-0 , Bur-0 and C24 ) . This effect was even stronger if TEs were located simultaneously within , upstream and downstream of the gene ( Figure 3a; MWU p≤2×10−14 for Col-0 , Bur-0 and C24 ) . Moreover , the average expression level of neighboring genes was positively correlated with the distance to the nearest TE ( Figure 3b; Spearman's ρ ( Col-0 ) = 0 . 15 , ρ ( Bur-0 ) = 0 . 13 , ρ ( C24 ) = 0 . 13; p<2×10−16 ) , and negatively correlated with the number of proximal TEs ( Figure 3c; df = 55 , chi-square sums 915 , 588 and 553 for Col-0 , Bur-0 and C24 , respectively , p<2×10−16 ) . Thus , gene expression is suppressed by proximal TEs , especially if they are close to the gene and numerous . Since TE superfamilies may have different effects on proximal genes , we examined gene expression according to the TE superfamily of the closest proximal TE . TE+ genes are expressed differentially depending on the TE superfamily of the proximal TE . TE+ genes with DNA transposons are on average expressed at a higher level compared to TE+ genes surrounded by retrotransposons ( Figure S12; MWU , p = 0 . 02 for Col-0 , Bur-0 and C24 ) . However , this is solely due to the higher expression level of genes proximal to CACTA elements . Indeed , we did not find evidence for CACTA TEs having any effect on gene expression ( Figure S12 , MWU , p ( CACTA TE+ genes/TE− genes ) = 0 . 7 , 0 . 6 and 0 . 8 for Col-0 , Bur-0 and C24 , respectively ) , which may explain why they are on average closer to genes than TEs from other families . Within the retrotransposons , LTR retrotransposons are younger on average than non-LTR retrotransposons and have a greater suppressive effect on proximal genes ( Table S2; [25] ) . Therefore TE superfamilies can differ considerably in their effects on proximal genes . TEs suppress the expression of neighboring genes at least partially through DNA methylation , which in turn is linked to 24-nt long siRNAs [12] , [15] , [22] , [26] , [27] . To investigate the influence of siRNAs on TE silencing , we sequenced siRNAs from mixed inflorescence tissue ( shoot meristem plus flowers , stages 1–14 ) of each accession and mapped the reads to all possible positions of the respective genomes without any mismatches . As expected from previous work , the density of siRNAs over TEs was about four times higher than the genome average ( Table S4; Figure S13 ) . We have reported before that siRNA-targeted TEs are more effective in suppressing expression of neighboring genes than are non-siRNA-targeted TEs , and that they are farther from genes [15] . We determined whether this held true in the current , more comprehensive dataset . If at least one 24-nt siRNA mapped to a TE it was labeled as siRNA+ ( Table S5 ) . siRNA+ and siRNA− TEs were overall similar in number , but retrotransposons were targeted by siRNAs more frequently than DNA transposons ( Figure S14; binomial test , p = 0 for Col-0 , Bur-0 and C24 ) . siRNA+ TEs were farther from genes ( Figure 4a; Figure S15a; MWU p<2 . 2×10−16 for Col-0 , Bur-0 and C24 ) , and this bias was consistent among TE superfamilies ( Figure S16 ) . To examine the effects of siRNA-targeting on the expression of flanking genes , we classified genes by whether the nearest TE was siRNA+ or siRNA− ( Table S5 ) . In each accession , genes flanked by siRNA+ TEs had lower average expression levels than genes with adjacent siRNA− TEs ( Figure 4b; Figure S15b; MWU p[Col-0] = 0 . 0001 , p[Bur-0] = 0 . 002 , p[C24] = 2×10−6 ) . The effect of suppression was stronger if the closest siRNA+ TE was within 2 kb of the gene ( Figure 4b; Figure S15b; MWU p<2×10−16 for Col-0 , Bur-0 and C24 ) . Therefore , as found previously for Col-0 , siRNA-targeting of TEs represses nearby genes and TEs that are close to genes are less likely to be targeted by siRNAs , either due to stronger selection for deletion of siRNA-targeted TEs close to genes or selection against siRNA-targeting of these TEs . Because siRNAs that map to unique positions in the genome ( usiRNAs ) correlate more closely with DNA methylation than siRNAs that map to multiple positions ( msiRNAs; [12] ) , we investigated whether usiRNAs and msiRNAs target TEs differentially , and how usiRNA− and msiRNA-targeted TEs might affect the expression of nearby genes . All TEs with at least one usiRNA were labeled as usiRNA+ ( Table S5 ) . In both Bur-0 and C24 , over 83% of siRNA+ TEs were usiRNA+ , similar to what has been reported for Col-0 [14] . usiRNA+ TEs were farther away from genes than msiRNA+ TEs ( Figure 4a; Figure S15a; MWU p[Col-0]<2×10−16 , p[Bur-0] = 6×10−13 and p[C24] = 2×10−6 ) . We also observed that the average expression level of genes within 2 kb of usiRNA+ TEs was lower than the expression of genes within 2 kb of msiRNA+ TEs ( Figure 4b; Figure S15b; MWU p[Col-0] = 3×10−6 , p[Bur-0] = 5×10−5 , p[C24] = 0 . 01 ) . Therefore , even though TEs targeted by usiRNAs and msiRNAs are on average farther from genes , they more strongly reduce expression of proximal genes compared to TEs targeted by only msiRNAs . Overall , we confirmed that siRNA+ TEs , especially usiRNA+ TEs , suppress neighboring gene expression , consistent with a trade-off between reduced TE mobility and deleterious effects on neighboring gene expression [14] , [15] . If TEs suppress the expression of adjacent genes , presence of gene-proximal TEs in the different accessions should be associated with differences in expression levels of proximal genes . We found that expression of TE− genes varied less between accessions than TE+ genes , and further that expression varied less between genes proximal to invariant TEs ( InvTE+ genes ) than genes proximal to variant TEs ( VarTE+ genes; Figure 5a; MWU p[TE−/TE+]<2×10−16 , p[InvTE+/VarTE+] = 2×10−5 ) . However , because TEs , and especially VarTEs , are found more often next to polymorphic genes , these conclusions could be confounded by correlated differences in genic polymorphisms . We therefore classified genes based on the extent of sequence variation ( Table S6 ) . Regardless of degree of genic polymorphism , VarTE+ genes were the ones that varied most in expression between accessions ( Figure 5b ) , indicating that TE variation increases variance in gene expression . We next determined whether differential siRNA-targeting influences gene expression . To remove the potentially confounding effects of variation in TEs themselves , we focused on InvTE+ genes and grouped these based on whether siRNAs for the adjacent TE could be detected in either all or none of the three accessions , or whether accessions differed in siRNA-targeting of the adjacent TE . We found that while variation in siRNA-targeting increased expression differences between accessions , this increase was not statistically significant ( Figure 5a ) . It should be noted that in our analysis we could not distinguish between the effects of differential siRNA-targeting and any perturbations of cis-regulatory sequences . Since each TE that differs in presence/absence or each siRNA-targeting variant between accessions represents a natural mutagenesis experiment , this offers an opportunity to study the effects on individual genes , to confirm the inferences drawn from averaging over all genes . We selected siRNA+ TE+ genes in Col-0 that are siRNA− TE+ or TE− in Bur-0 or C24 and tested for differential expression between Bur-0 or C24 and Col-0 . To remove the potential confounding effect of genic polymorphism , we excluded genes with a polymorphism level greater than 2% . Overall 706 genes were retained for this analysis . The effect of siRNA-targeting on gene expression was further verified by comparing expression profiles among wild-type , rdr2-1 and a ddc ( drm1drm2cmt3 ) DNA methyltransferase triple mutant [28] . Fifteen genes out of 706 showed significant up-regulation ( top 5% ranking ) in Bur-0 or C24 and in at least one of the RNA silencing mutants ( Table S7 ) . Although not statistically significant , this observation is consistent with siRNA-targeting and TE presence affecting gene expression . Moreover , it is likely an underestimate of TE effects on gene expression , given our stringent selection criteria . Because siRNA+ TEs suppress neighboring gene expression particularly efficiently , we asked whether targeting of different regions of TEs was reflected in the expression of adjacent genes . We first investigated whether invariant and variant TEs ( InvTEs and VarTEs ) differed in siRNA-targeting , normalized by TE length , and whether there were differences between invariable and variable regions of VarTEs ( Figure 6a; Table S8 ) . Fewer siRNAs mapped to siRNA+ VarTEs than to siRNA+ InvTEs ( Figure 6a; MWU p<2×10−16 for Col-0 versus Bur-0/C24 ) , but there were more siRNAs in variable regions than invariable regions of siRNA+ VarTEs in Col-0 ( Figure 6a; MWU p[Col-0/Bur-0] = 1×10−5 , p[Col-0/C24]<2×10−16 ) . Furthermore , usiRNAs were overrepresented in variable regions ( binomial test , p[Col-0/Bur-0] = 7×10−18 , p[Col-0/C24] = 0 ) , while msiRNAs were biased towards invariable regions ( p[Col-0/Bur-0] = 1×10−6 , p[Col-0/C24] = 0 ) . Therefore , usiRNAs strongly correlate with variability of TE sequences and are over-represented in the variable regions of variant TEs . This finding raised the question whether TE regions that varied between accessions and were targeted by siRNAs had a particularly large effect on expression of adjacent genes . We therefore separated Col-0 genes within 2 kb of variable TEs into three subsets: genes next to siRNA− VarTEs ( siRNA− VarTE+ genes ) ; genes next to VarTEs with an siRNA-targeting bias towards invariable TE regions ( InvsiRNA+ VarTE+ genes ) ; and genes next to VarTEs with an siRNAs targeting bias towards variable TE regions ( VarsiRNA+ VarTE+ genes; Table S8 ) . As expected , siRNA− VarTE+ genes had a higher average expression level compared to InvsiRNA+ VarTE+ genes ( Figure 6b; MWU p[Col-0/C24] = 0 . 01 , p[Col-0/Bur-0] = 0 . 01 ) or VarsiRNA+ VarTE+ genes ( MWU p[Col-0/C24] = 9×10−5 , p[Col-0/Bur-0] = 0 . 003 ) . The InvsiRNA+ VarTE+ genes , however , were expressed on average more highly than the VarsiRNA+ VarTE+ set ( MWU p[Col-0/C24] = 0 . 01 , p[Col-0/Bur-0] = 0 . 04 ) . This indicates that gene suppression by neighboring TEs may not only be influenced by siRNA presence or absence at the TEs , but may also depend on which TE regions are targeted by siRNAs . We speculate that siRNA-targeting of particular TE regions suppresses the expression of nearby genes to such an extent that there is significantly higher selection pressure for these regions to be excised or mutated . Alternatively , due to the skew of usiRNA mapping towards variable regions , and the greater correlation between usiRNAs and TE methylation , the lower expression level of VarsiRNA+ VarTE+ genes may reflect a higher degree of epigenetic silencing of these elements compared to InvsiRNA+ VarTE+ genes .
TEs may be prevented from reaching fixation within a population through negative selection , especially for gene-proximal , methylated TEs [13] , [15] , [34] . Therefore , it is perhaps unsurprising that TEs are over-represented in analyses of structural variants among accessions and between species [17] , [18] , [35] , [36] , and that a recent comparison of 80 A . thaliana genomes reported evidence of structural variation in 80% of TEs [17] . Similarly , Hollister and Gaut [15] found that 44% of over 600 TE insertions were polymorphic among 48 accessions . Since most TEs in A . thaliana are relatively old [7] , the simplest way to explain these patterns is ongoing deletion of TEs , which is also consistent with TEs in A . thaliana being on average farther from genes than in the closely related but outcrossing A . lyrata [7] . This may , however , be too simplistic an explanation as non-LTR retrotransposons are skewed towards an older insertion distribution than LTR retrotransposons [25] , even though they are not significantly more variable ( Table S2 ) . While TE presence/absence polymorphisms in different accessions have been previously characterized [17] , we have shown that there is substantial sequence variation in about 6% of TEs when comparing accessions ( Figure 1a ) . These TE variants are equally distributed throughout the genome ( Figure 1b ) . TEs can affect the expression of proximal genes via mechanisms including disruption of promoter sequences , reduction of transcription through the spread of epigenetic silencing [13] , or read-though antisense transcription [37] . Often TEs suppress the expression of proximal coding genes [15] , [22] , [38] however , TEs can also introduce new promoter sequences , leading to up-regulation of proximal genes [37] . In both plants and animals , TE-derived sequences have been recruited to form regulatory sequences and have contributed to coding regions [8] , [39]–[42] . Methylated TEs suppress expression of proximal genes in A . thaliana , regardless of insertion upstream or downstream of the coding region . Purifying selection is therefore greatest for methylated TEs proximal to genes [15] . Notably , the effects of siRNAs on expression of proximal genes can only be detected up to 400 bp [14] , while measurable TE effects extend to 2 kb [14] . This supports the assertion that TEs either directly affect gene expression by disruption of positive regulatory sequences , or otherwise act through DNA structure and epigenetic marks to affect genes over longer distances . We found that TEs that with variable siRNA-targeting do not affect proximal genes more strongly than TEs that are targeted in all three accessions ( Figure 5 ) . It is possible that siRNA-targeting varies independently of TE sequence variation , as observed recently for DNA methylation [43] , and that such TEs mask more subtle differences between the TE classes examined . However , the region of the TE targeted by siRNAs does seem to matter , with siRNA-targeting of TE sequences within an accession that are variant/absent in other accessions showing a greater suppression of proximal genes ( Figure 6 ) . This agrees with the observation that genes close to usiRNA-targeted TEs have a lower expression average than those close to msiRNA-targeted TEs , and that usiRNAs are over-represented in the variable regions of transposons . A recent study of hybrids between parents of different ploidy found that a reduction in 24 nt siRNAs is associated with up-regulation of more TE-associated genes than when there is no significant change in siRNA levels [44] . This result supports the hypothesis that siRNAs , or linked epigenetic changes , can affect the expression of nearby genes , with deletion of the siRNA-targeted regions alleviating repression of adjacent genes . While TEs in the euchromatin are often found close to genes , methylated TEs are underrepresented upstream of genes , likely because changes in the promoter more easily affect gene expression than variation in the 3′ region [13] . In agreement , methylated TEs have a skewed distribution , with older elements farther from genes , but unmethylated TEs do not show such a bias [15] . In a comparison of humans and chimpanzees , TE insertion site preference appears to be the main cause for TEs being found more often in the vicinity of genes with increased interspecific expression variation [45] . This is reminiscent of what we have observed , with additive effects of polymorphism , TE presence and TE variance on the variability of orthologous gene expression ( Figure 2 and Figure 4 ) . In a comparison of two rice subspecies , TE presence/absence polymorphisms were also found to be underrepresented in SNP deserts [35] . There are several possible explanations for these observations: some genomic regions may suffer from generally elevated mutation rates TEs near highly conserved genes are more efficiently purged; or TE integration into more mutable genomic regions is favored . In the latter case , new mutations may destabilize DNA packing and facilitate TE insertions , similar to the TE insertion preference for transcribed genomic regions [42] . With our observation of TE deletions correlating with siRNA-targeting , we can expand the current model for TE evolution [15] . Our model starts with the duplication of a TE that is already present and targeted by siRNAs within the genome ( Figure 7a and 7b ) , leading to all siRNAs produced by and targeting the original TE now being multiply-mapping siRNAs ( msiRNAs ) . As the two copies of the duplicated TE gain mutations ( enhanced by deamination of methylated cytosines ) , uniquely-mapping siRNAs ( usiRNAs ) are produced in addition to msiRNAs ( Figure 7c ) . Hollister and colleagues [14] noted that usiRNA-targeting increases with TE age , while msiRNA-targeting decreases , and that TEs are expressed at lower levels when also targeted by usiRNAs . Furthermore , usiRNAs are more closely correlated with DNA methylation than are msiRNAs [12] and they are expressed at higher levels than msiRNAs [14] . With usiRNAs , the duplicated TEs will therefore be more effectively silenced , probably with a concurrent increase in methylation , a further reduction in the expression level of proximal genes , and thus increased selection against the TEs . usiRNA-targeting may then facilitate TE inactivation through preferential deletion of usiRNA-targeted regions ( Figure 6 and Figure 7d ) . This may be actively promoted by the usiRNAs and attendant epigenetic marks , in a mechanism analogous to the siRNA-guided removal of “internal eliminated sequences” including TEs in Tetrahymena [46] , [47] . In favor of such a scenario , small deletions within TEs have been shown to occur more frequently than ectopic recombination events at the LTRs [31] , [48] . Ectopic recombination appears to be less important for TE elimination in A . thaliana , as TE density and recombination rate are not correlated in this species [48] , and because ectopic recombination is lower in homozygotes [49] . No matter what the mechanism , deletions within TEs would reduce selection pressure by removing usiRNA target sites , inactivating TEs so they are no longer transposition-competent , and relieving proximal gene repression . In apparent contrast to the majority of TEs , some are under positive selection [50] , [51] , and TEs can also contribute to new regulatory networks [52] . Our model is only appropriate for TEs under neutral or negative selection . Modeling of TE dynamics suggests that transposition events occur in a cyclical manner [53] , [54] , with some activation events creating new favorable genetic variants . One such example is provided by transposition of a TE that is induced upon heat stress in genetic backgrounds impaired in siRNA biogenesis confers heat-responsiveness to proximal genes [55] . We have exploited high-quality genome information from multiple accessions of a single species to study the effects of TE variation on proximal gene expression . We discovered a link between siRNA-targeting and TE variation that illuminates how epigenetic mechanisms may help to shape genomes , but several questions remain: Do usiRNAs directly facilitate TE deletions or do they act indirectly through differences in selection for deletions ? Are TE deletions in other species also associated with regions of increased usiRNA-targeting ? And do species differ in the rate of TE deletion via this mechanism ? Because of the rarity of TE deletions , this is a challenging process to dissect . Genomes with a large fraction of TEs , such as those of many crop plants , might therefore prove more tractable systems for studying mechanism of TE removal than the TE poor A . thaliana genome .
We extracted positions of genes and TEs from the A . thaliana Col-0 genome sequence TAIR version 9 from http://www . arabidopsis . org . We excluded genes and TEs within the centromeric regions [56] . To define gene and TE sets in Bur-0 and C24 , we built genome templates using published Illumina paired-end reads of Bur-0 and C24 [19] . We used the SHORE pipeline [21] to align the reads to the Col-0 reference genome and extracted the consensus sequences as genome templates by calling bases with quality>24 , support>6 , concordance>0 . 7 and average hits = 1 . We then applied a naïve projection of the coordinates of genes and TEs from Col-0 onto the genome templates to define the gene and TE sets of Bur-0 and C24 . SHORE was also used to detect genomic variations by calling SNPs , small ( 1–3 bp ) insertions/deletions and larger deletions from the genome templates of Bur-0 and C24 compared to the Col-0 genome using the same parameters for quality control . The distance between TEs and genes in Bur-0 and C24 was estimated from Col-0 using the annotated TE and gene coordinates , and adjusted to account for insertions and deletions between TEs and genes . For each polymorphism type ( i . e . , SNPs , small indels , and large deletions ) , we compared the densities pairwise across coding regions , intergenic regions and TEs . To test whether a higher density was significant in a particular genomic region ( e . g . TE ) compared to others ( e . g . coding region ) , a cumulative binomial probability distribution was applied:p is the polymorphism density in coding regions , and k and n are the total number of polymorphic sites in TEs and the total length of TEs , respectively . We calculated gene polymorphism levels as the fraction of genic region containing small-scale variations in at least C24 or Bur-0 , or one of the 80 A . thaliana accessions [17] . Genes with more than 20% zero sequencing coverage or no base calls among 80 accessions were excluded from the analysis . 4 kb 5′ and 3′ flanking regions for each TE were extracted . For each flanking region ( FR ) , or genic regions ( GR ) within the FR , small-scale mutations and large deletion polymorphisms between Col-0 and Bur-0/C24 were calculated . Using all mutations , the polymorphism levels of TEs , FRs and GRs were ranked . A threshold of 50% was used to split FRs and GRs into high or low polymorphism datasets and thereby classify the TEs by genomic environment . The polymorphism levels of the FRs were calculated in 200 bp bins for each group of TEs , with binomial tests to compare polymorphism levels between TEs and FRs , and between different TE groups . Inflorescences ( meristem and flowers up to stage 14 ) were pooled from five plants of each accession grown at 23°C . Triplicate samples were collected between 7 and 8 hours into a 16 hour light cycle . RNA was extracted using the Qiagen ( Hilden , Germany ) Plant RNeasy Mini kit . Each biological replicate was analyzed with Affymetrix ( Santa Clara , CA , USA ) tiling 1 . 0R arrays and the data were processed according to published methods [57] , [58] . Tiling array probes that were polymorphic for C24 or Bur-0 were removed from the dataset for the affected accession ( s ) . For gene expression estimates , ≥70% and at least 3 probes had to be present; all other genes were not considered . Tilling array data from Arabidopsis Col-0 and the RNA silencing mutants rdr2-1 and ddc ( drm1-1;drm2-2;cmt3-11 ) mutants were downloaded from GEO ( GSE12549; [28] ) and processed according to published methods [57] , [58] . Expression level changes for each dataset were estimated by fold-change differences between Bur-0/C24 and Col-0 , and between the RNA silencing mutants and wild type Col-0 . Background distributions of fold-change were calculated and genes , with a fold-change exceeding a one-sided 95% quantile in each dataset were considered as significantly up-regulated in Bur-0/C24 or the mutants . The siRNA datasets have been published [19] ( GEO accession number GSE24569 ) . We mapped the 24-nt siRNA reads onto both strands of the genome templates ( see below ) and the TEs of Col-0 , Bur-0 and C24 , respectively , using the Vmatch package ( http://www . vmatch . de ) . Only reads with perfect matches were considered . The statistical significance of over-representation of usiRNAs or msiRNAs within the variable regions of siRNA+ VarTEs in comparison to all siRNAs was tested using the cumulative binomial probability distribution given above . p , expected frequency , is the ratio between the number of siRNAs mapped to the variable regions the total number of siRNAs mapped to any region of siRNA+ VarTEs , and n and k are the total number of usiRNAs/msiRNAs mapped to any region and the number of usiRNAs/msiRNAs mapped to the variable regions , respectively . We defined an siRNA+ VarTE as either InvsiRNA+ or VarsiRNA+ if siRNAs are overrepresented in the invariable regions and variable regions , respectively . For siRNA+ VarTEs that contain siRNAs in both variable and invariable regions , we employed the cumulative binomial probability distribution described above to test whether siRNA-targeting shows statistically significant bias towards variable or invariable regions . For each siRNA+ VarTE , p in the formula above is the abundance of siRNA-targeting at the TE . To test the bias towards variable regions , n and k represent the genomic length of variable regions and the number of siRNAs targeting variable regions , respectively . Similarly , to test the bias towards invariable regions , n and k represent the genomic length of invariable regions and the number of siRNAs targeting invariable regions , respectively . P-values were adjusted for multiple hypothesis testing with the Benjamini-Hochberg method to control for a false discovery rate of 5% [59] . The siRNA and microarray data reported in this paper have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus ( NCBI GEO ) ( http://www . ncbi . nlm . nih . gov/geo/ ) under accession numbers GSE24569 and GSE24669 . The genome assemblies are available from http://1001genomes . org/projects/MPIWang2012/ while the transposable element annotations for Bur-0 and C24 are available from Dryad under doi 10 . 5061/dryad . 8674d .
|
Transposable elements ( TEs ) are selfish DNA sequences . Together with their immobilized derivatives , they account for a large fraction of eukaryotic genomes . TEs can affect nearby gene activity , either directly by disrupting regulatory sequences or indirectly through the host mechanisms used to prevent TE proliferation . A comparison of Arabidopsis thaliana genomes reveals rapid TE degeneration . We asked what drives TE degeneration and how often TE variation affects nearby gene expression . To answer these questions , we studied the interplay between TEs , DNA sequence variation , and short interfering RNAs ( siRNAs ) in three A . thaliana strains . We find sequence variation in genes and adjacent TEs to be correlated , from which we conclude either that TEs insert more often near polymorphic genes or that TEs next to polymorphic genes are less efficiently purged from the genome . We also noticed that processes that cause deletions within TEs and ones that silence TEs appear to be linked , because siRNA targeting is a predictor of sequence loss in accessions . Our work provides insight into the contribution of TEs to gene expression plasticity , and it links TE silencing mechanisms to the evolution of TE variation between genomes , thereby linking TE silencing mechanisms to expression plasticity .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"transposons",
"genomics",
"molecular",
"cell",
"biology",
"plant",
"science",
"genome",
"evolution",
"plant",
"evolution",
"plant",
"biology",
"plant",
"genomics",
"biology",
"computational",
"biology",
"epigenomics",
"molecular",
"biology",
"genetics",
"and",
"genomics"
] |
2013
|
Transposon Variants and Their Effects on Gene Expression in Arabidopsis
|
Formation of centromeric heterochromatin in fission yeast requires the combined action of chromatin modifying enzymes and small RNAs derived from centromeric transcripts . Positive feedback mechanisms that link the RNAi pathway and the Clr4/Suv39h1 histone H3K9 methyltransferase complex ( Clr-C ) result in requirements for H3K9 methylation for full siRNA production and for siRNA production to achieve full histone methylation . Nonetheless , it has been proposed that the Argonaute protein , Ago1 , is the key initial trigger for heterochromatin assembly via its association with Dicer-independent “priRNAs . ” The RITS complex physically links Ago1 and the H3-K9me binding protein Chp1 . Here we exploit an assay for heterochromatin assembly in which loss of silencing by deletion of RNAi or Clr-C components can be reversed by re-introduction of the deleted gene . We showed previously that a mutant version of the RITS complex ( Tas3WG ) that biochemically separates Ago1 from Chp1 and Tas3 proteins permits maintenance of heterochromatin , but prevents its formation when Clr4 is removed and re-introduced . Here we show that the block occurs with mutants in Clr-C , but not mutants in the RNAi pathway . Thus , Clr-C components , but not RNAi factors , play a more critical role in assembly when the integrity of RITS is disrupted . Consistent with previous reports , cells lacking Clr-C components completely lack H3K9me2 on centromeric DNA repeats , whereas RNAi pathway mutants accumulate low levels of H3K9me2 . Further supporting the existence of RNAi–independent mechanisms for establishment of centromeric heterochromatin , overexpression of clr4+ in clr4Δago1Δ cells results in some de novo H3K9me2 accumulation at centromeres . These findings and our observation that ago1Δ and dcr1Δ mutants display indistinguishable low levels of H3K9me2 ( in contrast to a previous report ) challenge the model that priRNAs trigger heterochromatin formation . Instead , our results indicate that RNAi cooperates with RNAi–independent factors in the assembly of heterochromatin .
Eukaryotic genomes are characterized by domains of transcriptionally permissive euchromatin and relatively transcriptionally inert heterochromatin . In addition to its important role in transcriptional regulation , heterochromatin plays a critical role in the regulation of genomic stability . In fission yeast constitutive heterochromatin assembles at the centromeres , telomeres and the mating type locus . This heterochromatin is required for high fidelity chromosome transmission , protecting chromosome ends from fusion to form dicentric chromosomes , and preventing co-expression of both sets of mating type information which could lead to haploid meiosis and cell death . A major hallmark of heterochromatin in most eukaryotes is the presence of methyl groups on lysine 9 of histone H3 . In fission yeast ( Schizosaccharomyces pombe ) , methylation of H3 K9 is carried out by a single enzyme , Clr4 ( the homolog of Suvar39 enzymes in higher eukaryotes ) , which is responsible for mono , di and tri-methylation of H3K9 [1] . This mark is in turn bound by proteins bearing a chromodomain , including the HP1 homologs Swi6 and Chp2 , and importantly , Clr4 itself , leading to models for perpetuation and spreading of heterochromatin [2]–[5] . A fourth chromodomain protein , Chp1 , has high affinity for binding the methyl mark [6] . Chp1 is a component of the RITS complex ( RNA-induced initiation of transcriptional silencing complex ) , which is critical for the accumulation of heterochromatin at centromeres [7] , [8] . Heterochromatin assembly in several organisms also depends upon the cellular RNA interference ( RNAi ) pathway [9] . RNAi is triggered by double-stranded RNA ( dsRNA ) , which is processed by the RNAseIII-like activity of Dicer into short interfering ( si ) RNAs . siRNAs are loaded into RNA –induced silencing complexes ( RISC ) , where they associate with Argonaute proteins , and base-pair with and promote the sequence-dependent destruction of RNA via cleavage mediated by Argonaute . In fission yeast , the RNAi effector complex is called RITS , and consists of the sole argonaute protein , Ago1 , in complex with Tas3 which physically links Ago1 to Chp1 [8] , [10] . Association of the RITS complex with centromeres is co-dependent on the RNA dependent RNA polymerase complex , RDRC [11] . RITS and RDRC physically associate with the Clr4-containing Clr-C complex [5] , [12] , and trigger an RNAi-mediated positive feedback loop to enhance H3K9me2 accumulation and heterochromatin assembly [13] . Accumulation of H3K9me2 allows recruitment of heterochromatin- binding proteins such as Swi6 ( the fission yeast homolog of HP1 ) and cohesin to centromeric repeats , and is required for efficient chromosome segregation ( reviewed in [14] ) . Clearly , the mechanism by which RITS and RDRC are initially recruited to centromeres to promote RNAi-dependent accumulation of H3K9me2 is critical to our understanding of heterochromatin assembly . Somewhat paradoxically , the outer repeats of the centromere are transcribed by RNA polymerase II , and this transcription correlates with heterochromatin assembly [15]–[18] . Recently , two models have been proposed for how centromeric transcripts may initiate recruitment of RITS/RDRC . The first proposes that single stranded centromeric transcripts fold into hairpin structures to provide dsRNA template for the activity of dicer ( Dcr1 ) to generate centromeric siRNAs to target RITS to homologous centromeric sequences [19] . Alternatively , RNA degradation products ( priRNAs ) associate with Ago1 , and if derived from centromeric transcripts , target Ago1 to centromeres [20] . Ago1 slices transcripts that are homologous to priRNAs , recruiting RDRC which promotes dsRNA synthesis . dsRNAs are cleaved by Dcr1 to form centromeric siRNAs that recruit RITS to centromeres [8] . In both models , centromeric RITS/RDRC then promotes Clr-C association , and H3K9 methylation facilitating binding of Chp1 to chromatin . These models infer that small RNAs and the RNAi pathway act as the priming signal for heterochromatin assembly , with Dcr1 or Ago1 playing the initiating role respectively . However , RITS possesses two potential centromeric targeting motifs: Ago1 which binds siRNAs and can target centromeric transcripts and Chp1 which has high affinity for binding H3K9me2 [6] , [7] . We questioned whether indeed RNAi is the upstream event for heterochromatin assembly , or whether Clr4 functions upstream of RNAi to generate H3K9me2 to recruit RITS . Dissection of the requirements for the initial assembly of centromeric heterochromatin is greatly hampered by the inter-relatedness of the RNAi and chromatin modifying pathways and the positive feedback loops involved in full heterochromatin assembly . Deletion of genes required for assembly of heterochromatin ablates heterochromatin , complicating analysis of whether the gene contributes to the establishment or maintenance of heterochromatin . To define the contribution of siRNAs and H3K9me to targeting RITS to centromeres , we have generated mutants that separate Ago1 from the RITS complex [10] , that remove Chp1 from the RITS complex [21] , and mutations within the chromodomain of Chp1 [6] , [22] that weaken the high affinity of Chp1's chromodomain for binding H3K9me2/3 . Data accumulated from analysis of these mutants strongly supports that centromeric targeting of RITS critically depends on Chp1 and in particular , Chp1 chromodomain's high affinity for binding H3K9me2/3 . The tas3WG mutant separates Ago1 from Tas3-Chp1 [10] . This mutant bears a two amino acid alanine substitution of residues W265 and G266 within the conserved WG/GW Ago “hook” ( or interaction domain ) of Tas3 [10] , [23] , that renders the Chp1-Tas3 subcomplex incapable of associating with Ago1 . Surprisingly , tas3WG cells can maintain preassembled heterochromatin , most likely via retention of the subcomplexes of RITS at centromeres because of Chp1-Tas3 association with H3K9me2 and Ago1's association with centromeric siRNAs [10] . Interestingly , following removal of all H3K9me2 and loss of heterochromatin –dependent siRNAs ( in clr4Δ backgrounds ) , tas3WG cells fail to generate de-novo heterochromatin on reintegration of clr4+ [10] . We reasoned that genes that are particularly important for the establishment of heterochromatin would likewise be defective for heterochromatin assembly if transiently depleted in the tas3WG background . In contrast , genes with a less critical role for heterochromatin assembly might be expected to assemble heterochromatin efficiently if transiently depleted in tas3WG cells . Here , we directly test the contribution of proteins involved in the RNAi pathway or in methylation of H3K9 to the initiation of centromeric heterochromatin . We find that transient depletion of any Clr-C component perturbs the establishment of heterochromatin in tas3WG cells . In contrast , transient depletion of genes involved in the RNAi pathway does not block heterochromatin assembly in tas3WG cells . Thus there appears to be a continuous requirement for the Clr-C complex , but not RNAi , during heterochromatin assembly in cells bearing a disrupted RITS complex . Consistent with this , RNAi-defective cells retain low levels of the heterochromatin mark , H3K9me2 , whereas this mark is completely absent from Clr-C mutant cells . Because RNAi-defective cells maintain residual H3K9me2 , it is not heterochromatin initiation which is being monitored following reintroduction of RNAi components , but more downstream aspects of heterochromatin assembly . To determine if RNAi is required for the initial step in heterochromatin assembly , we additionally removed H3K9me2 from RNAi defective cells by making compound mutants with clr4Δ , and interrogated whether on re-expression of clr4+ , Clr-C could target centromeric sequences . We found that Clr4 can promote de novo methylation of centromeric repeats when overexpressed in cells otherwise lacking Clr4 and either of the RNAi components Ago1 or Dcr1 . Thus , Clr-C can target H3K9me2 to centromeric repeats independently of the RNAi pathway . This data , plus our finding that ago1-deficient cells retain significant levels of H3K9me2 on centromeric repeats shows that Ago1 and Ago1 bound priRNAs are not necessary for the initiation of assembly of centromeric heterochromatin . Instead , our data strongly indicates that RNAi-independent mechanisms function together with RNAi in the cooperative assembly of centromeric heterochromatin .
To precisely define the point of action of regulators of heterochromatin assembly , we have employed the tas3WG allele which disrupts the RITS complex . Transient gene depletion experiments in tas3WG cells previously showed that the H3K9 methyltransferase , Clr4 , is required to establish centromeric heterochromatin when Ago1 is separated from Chp1-Tas3 [10] . Clr4 is a component of a large complex of proteins called Clr-C . Cells lacking Clr-C components lose centromeric heterochromatin [24]–[29] , but the role of individual Clr-C components in heterochromatin assembly is poorly understood . We therefore assessed the point of action of Clr-C components in heterochromatin establishment , using transient gene depletion experiments in the tas3WG background . Rik1 was the first protein identified in complex with Clr4 [30] , and it resembles the UV DNA damage binding protein , UV-DDB1 , including homology to the CPSF-A factor involved in RNA processing [31] . Rik1 is thought to act upstream of Clr4 , and to help recruit Clr-C to chromatin , since Rik1 remains localized at centromeres in mutants that mislocalize Clr4 [5] , [28] . We tested whether transient depletion of rik1+ in tas3WG cells would prevent assembly of heterochromatin . We introduced the tas3WG-TAP and tas3-TAP alleles into a rik1Δ strain that carries a ura4+ transgene within the outer repeats of centromere 1 ( cen::ura4+ ) . Wild type cells efficiently assemble heterochromatin on cen::ura4+ , silencing its expression , and allowing growth on media containing the drug 5-FOA , which is toxic to cells that express ura4+ . Cells lacking rik1+ fail to silence the centromeric transgene . The re-establishment of centromeric heterochromatin was monitored following reintroduction of rik1+ into its normal genomic locus . Addition of rik1+ to rik1Δ tas3-TAP cells allowed efficient establishment of heterochromatin and silencing of the cen::ura4+ transgene ( Figure 1B ) . In contrast , on reintroduction of rik1+ into tas3WG-TAP cells , heterochromatin did not reassemble to silence the cen::ura4+ reporter . Transcription of endogenous dg and dh centromeric repeats was measured by real time PCR in cDNA prepared from these strains . In wild type cells , centromeric transcripts are processed by siRNA-dependent Ago1-mediated processing and by RNAi-independent turnover [32]–[34] . In addition , heterochromatin that assembles on repeat sequences can reduce access of RNA polymerase , thus preventing transcript accumulation [35] . Centromeric transcripts from dh ( Figure 1C ) and dg ( Figure 1D ) accumulate in cells lacking rik1+ , similar to cells lacking clr4+ . On reintegration of rik1+ into tas3-TAP cells , centromeric transcripts become normally processed , resulting in no net gain in transcript levels in rik1Δ to rik1+ tas3-TAP cells relative to tas3-TAP cells . Strikingly , both dg and dh transcript levels remain high in tas3WG cells following reintegration of rik1+ , consistent with the observed silencing defect of the cen::ura4+ reporter in these strains . This accumulation of transcripts is at least in part due to defective processing of centromeric transcripts into siRNAs by the RNAi machinery since siRNAs were not detectable by Northern blotting in tas3WG-TAP rik1Δ to rik1+ cells whereas rik1+ reconstituted tas3-TAP cells synthesized centromeric siRNAs as efficiently as tas3-TAP cells ( Figure 1E ) . Raf1 ( Cmc1 , Dos1 , Clr8 ) and Raf2 ( Cmc2 , Dos2 , Clr7 ) have also been identified as components of Clr-C [27] , [29] . They are required for localization of Swi6 [25] , and are important for silencing the mating type locus [26] . raf1+ encodes a WD repeat protein which can bind Rik1 [25] , and raf2+ encodes a putative Zn finger protein which binds to Pcu4 [26] . raf1Δ and raf2Δ were crossed into tas3-TAP and tas3WG -TAP backgrounds , and wild type genomic copies of raf1+ or raf2+ were reintegrated into the corresponding deletion mutants and assessed for heterochromatin assembly . As seen for transient depletion experiments with rik1 , tas3-TAP cells efficiently re-assembled centromeric heterochromatin on reintroduction of raf1+ or raf2+ , whereas silencing of the cen::ura4+ reporter was not apparent in tas3WG-TAP backgrounds ( Figure 2A and 2B ) . Centromeric transcripts accumulate to high levels in raf1 and raf2 deleted cells , and although transcript levels drop following reintegration of raf1+ into raf1Δ tas3-TAP cells or of raf2+ into raf2Δ tas3-TAP cells , high levels of dg and dh transcripts are maintained in both raf1+ and raf2+ reconstituted tas3WG-TAP cells ( Figure 2C and 2E , Figure S1A and S1B ) . Consistent with this failure to suppress high levels of centromeric transcription in tas3WG-TAP cells transiently depleted for raf1+ or raf2+ , these cells fail to engage the RNAi pathway to promote destruction of centromeric transcripts into siRNAs ( Figure 2D and 2F ) . Pcu4 is the fission yeast cullin4 , and it has been identified in complex with the UV-DDB1 E3 ubiquitin ligase [35] , [36] , and with the related Rik1 protein in the Clr-C complex [27]–[29] . To define the role of Pcu4 in heterochromatin establishment , we monitored heterochromatin assembly following reintroduction of the pcu4+ gene into pcu4Δ tas3-TAP and tas3WG-TAP strains . Clearly centromeric transcripts accumulate in pcu4Δ cells , and processing of transcripts is efficiently resumed following re-introduction of the wild type gene into tas3-TAP cells ( Figure 3A and 3B ) . However , both dh and dg transcript levels are maintained at high levels following pcu4+ reintroduction into pcu4Δ tas3WG -TAP cells . This failure to silence centromeric transcripts was reflected in the failure to produce abundant centromeric siRNAs in these tas3WG reconstituted cells ( Figure 3C ) . In summary , all components of Clr-C are defective for silencing of endogenous centromeric or centromeric reporter transcripts following their transient depletion in tas3WG cells , suggesting that their continuous presence is required for the initiation of heterochromatin in RITS-defective cells . Next we analyzed H3K9 methylation on centromeric sequences following transient depletion of components of the Clr-C complex . In wild type cells , H3K9me accumulates to high levels on centromeric repeats through both RNAi-dependent and RNAi-independent mechanisms [15] , [30] . In cells lacking pcu4 , H3K9me2 levels on dh sequences are not above the background seen for clr4Δ cells which lack H3K9me2 ( Figure 3D , upper panel ) . Following pcu4+ reintroduction into pcu4Δ tas3-TAP cells , H3K9me2 levels rise to that seen in tas3-TAP cells , whereas no significant accumulation of H3K9me2 is observed in pcu4+ reconstituted tas3WG-TAP cells . Thus Clr4 mediated H3K9 methylation is abrogated in tas3WG-TAP cells transiently depleted for pcu4 . This methylation defect is not likely due to defective reassembly of the Clr-C complex following transient depletion of pcu4 , since H3K9 methylation resumes effectively in pcu4+ reconstituted tas3-TAP cells . Chp1 binds H3K9me2/3 and Chp1 recruitment to centromeres is a hallmark of heterochromatin . ChIP experiments performed with anti-Chp1 antibodies demonstrated that pcu4Δ cells are also defective for Chp1 association with centromeres , and pcu4+ reconstitution of pcu4Δ tas3-TAP but not of pcu4Δ tas3WG-TAP cells promotes Chp1 association with centromeres ( Figure 3D , lower panel ) . We also assessed H3K9me2 levels at centromeres in other strain backgrounds . In all Clr-C mutants ( raf2 , rik1 , raf1 ) , centromeric H3K9me2 levels were no higher than seen in clr4Δ cells ( Figure 4A and 4C ) . Following reconstitution with raf2+ , clr4+ , rik1+ , or raf1+ , tas3-TAP cells accumulated “wild type” levels of H3K9me2 , but tas3WG-TAP cells failed to accumulate H3K9me2 at centromeres . Very similar results were obtained for Chp1 association with centromeres ( Figure 4B and 4D ) , consistent with tas3WG-TAP cells being dependent on constitutive expression of all components of the Clr-C complex to provide H3K9me2 at centromeric sites for recruitment of Chp1 . In sum , these experiments demonstrate that in cells where the association of Ago1 with Chp1-Tas3 has been abrogated , that reintroduction of Clr-C components is not sufficient to direct H3K9me2 accumulation on centromeric repeats . Clr-C defective cells should still express heterochromatin independent siRNAs , and Ago1 in these cells would be expected to maintain association with primal RNAs . Thus targeting of Ago1 by priRNAs to centromeric repeats is not sufficient to drive Clr-C recruitment to centromeres when Ago1 is physically separated from Tas3-Chp1 . Next we examined whether RNAi components contribute to the initial steps in heterochromatin assembly . RNAi defective cells , such as dcr1Δ , are expected to retain priRNAs but lose most of their siRNAs . In contrast to Clr-C defective cells , dcr1Δ cells maintain low levels of H3K9me2 at centromeres ( Figure S2D ) . Following overexpression of dcr1+ , both dcr1Δ tas3-TAP and dcr1Δ tas3WG -TAP cells efficiently assembled heterochromatin [10] . This suggested that H3K9me2 , and not siRNA , acts at an early stage of heterochromatin initiation . However , in these experiments it was unclear whether overexpression of dcr1+ suppressed an establishment defect in tas3WG dcr1+ reconstituted cells [10] , [14] . We directly tested whether integration of dcr1+ into the genomic dcr1Δ locus of tas3-TAP and tas3WG-TAP cells could support reassembly of centromeric heterochromatin . Cells lacking dcr1+ fail to silence the cen::ura4+ centromeric transgene . Following reintegration of dcr1+ , silencing of the cen::ura4+ reporter resumed in both dcr1Δ to dcr1+ tas3-TAP and dcr1Δ to dcr1+ tas3WG-TAP cells ( Figure S2A ) . Cells lacking dcr1+ accumulate high levels of centromeric transcripts , but following dcr1+ reintegration , centromeric transcript levels were reduced in both the dcr1+ reconstituted tas3-TAP and tas3WG-TAP cells , confirming that reintegration of dcr1+ promoted efficient assembly of centromeric heterochromatin ( Figure S2B , S2C ) . In addition , dcr1Δ cells cannot generate siRNAs from centromeric transcripts , but on reintegration of dcr1+ , siRNA production resumed efficiently in both tas3-TAP and tas3WG-TAP backgrounds ( Figure S2D ) . Together , these results confirmed and extended our data obtained with overexpressed dcr1+ [10] . dcr1+ and siRNAs are not critical for Clr-C activity at centromeres , but are important for amplification of the H3K9me2 signal during later stages of heterochromatin assembly . We next asked whether transient depletion of genes that act upstream of Dcr1 in the RNAi pathway would cause defective heterochromatin establishment . RDRC acts upstream of Dcr1 , generating dsRNA for siRNA production . RDRC consists of the RNA-dependent RNA polymerase ( Rdp1 ) , the RNA helicase Hrr1 , and a non-canonical poly ( A ) polymerase , Cid12 [11] . Cells lacking any component of RDRC show reduced RITS association and reduced H3K9me2 at centromeres , and have reduced siRNA production [11] , [19] , [20] . We introduced the tas3-TAP and tas3WG-TAP alleles into deletion mutants of all components of RDRC , and then tested whether silenced chromatin assembled on the cen::ura4+ reporter following reintegration of genomic clones encoding these genes . For cells lacking cid12+ , hrr1+ , or rdp1+ , reintegration of these genes into knockout tas3-TAP cells allowed efficient assembly of heterochromatin . Interestingly , as seen for Dcr1 , reintroduction of the genes into the mutant tas3WG-TAP strains also supported silencing of the cen::ura4+ reporter ( Figure 5A and 5B , Figure 6A ) . Transcript analyses performed on the RDRC reconstituted strains revealed that cells lacking RDRC components accumulate both centromeric dg and dh transcripts , but that on reconstitution with the wild type gene , dg and dh transcript levels dropped to levels close to those normally found in tas3-TAP or tas3WG-TAP cells , which is considerably less than seen in RDRC mutant cells ( Figure 5C and 5E , Figure 6C , and Figures S3A , S3B , and S4A ) . Thus processing of centromeric transcripts is efficiently resumed following reintroduction of RDRC components into RDRC deficient tas3WG cells , and this conclusion is further supported by detection of siRNAs in RDRC reconstituted cells ( Figure 5D , 5F and Figure 6B ) . In summary , in contrast to cells transiently depleted for Clr-C components , centromeric heterochromatin assembly can occur efficiently following the transient depletion of RDRC components or of dcr1+ in tas3WG cells . Cells lacking dcr1+ accumulate H3K9me2 on centromeric sequences , whereas Clr-C deficient cells completely lack H3K9me2 . We therefore asked whether cells lacking RDRC components accumulate centromeric H3K9me2 , and whether H3K9me2 levels could signal the difference in outcome , promoting heterochromatin assembly in tas3WG cells following transient depletion of RDRC , but not Clr-C components . We assessed centromeric H3K9me2 levels in RDRC deficient cells and following reintegration of RDRC components . In these experiments , we note that in all RDRC mutants , the level of H3K9me2 at centromeres is considerably higher than seen in cells lacking clr4+ , although at least 2 fold reduced compared with wild type cells . Reintegration of RDRC components into the corresponding RDRC null cells supported centromeric accumulation of H3K9me2 of both tas3-TAP and tas3WG-TAP cells to levels found normally ( Figure 6D and 6E ) . In addition , although Chp1 association with centromeres is diminished in hrr1Δ cells , reintroduction of hrr1+ into tas3WG-TAP cells promoted Chp1 association ( Figure S4B ) . Together then this data shows that heterochromatin assembly occurs efficiently following transient depletion of genes required for siRNA synthesis , including RDRC components that act upstream of Dcr1 . In addition , the ability of heterochromatin to reform efficiently , following transient depletion of RNAi components in tas3WG cells , correlates with the persistence of low levels of H3K9me2 on centromeric repeats in the mutant backgrounds . Very recently it has been proposed that Ago1 is the most upstream factor in heterochromatin assembly . It is thought to act as an acceptor for RNA degradation products , termed pri-RNAs , which , based on frequency of occurrence , preferentially target antisense transcripts resulting from bidirectional transcription of DNA repeats . Cleavage of nascent centromeric transcripts by priRNA-directed activity of Ago1 is proposed to recruit the RDRC complex and eventually promote siRNA-dependent recruitment of RITS , and subsequent robust assembly of heterochromatin via recruitment of Clr-C [20] . This model therefore places Ago1 as an initiator , upstream of RDRC and Dcr1 and of the Clr-C complex and H3 K9 methylation . This model is supported by the detection of small RNAs ( priRNAs ) in dcr1Δ strains , and of siRNAs in cells lacking clr4+ , or in which heterochromatin assembly is blocked because of mutation of H3K9 , supporting that heterochromatin is not essential for small RNA generation [19] , [20] . Finally , the model would suggest that siRNAs and priRNAs act upstream of heterochromatin assembly , and that Ago1 is the most upstream component of the RNAi pathway . Indeed , Halic and Moazed argue that Ago1 activity is required for the initial deposition of H3K9me , since in their publication strains lacking Ago1 exhibit lower levels of centromeric H3K9me accumulation than strains deficient in other components of the RNAi pathway [20] . To test the role of Ago1 in heterochromatin assembly , we first performed transient depletion experiments for Ago1 in the tas3WG-TAP background ( Figure 7 ) . We integrated a genomic clone of ago1+ into ago1 null tas3-TAP and tas3WG-TAP cells , and monitored heterochromatin assembly . In contrast to ago1 null cells , where centromeric transcripts are highly elevated ( above the levels seen in clr4Δ cells ) , reintroduction of ago1+ into either tas3-TAP or tas3WG-TAP cells reduced centromeric dh and dg transcripts to levels seen normally in tas3-TAP and tas3WG-TAP cells ( Figure 7A and Figure S5A ) . Consistent with this suppression , we found that unlike ago1 null cells , where siRNA production is severely reduced , centromeric siRNAs are synthesized at normal levels following reintroduction of ago1+ ( Figure 7B ) . Next , we performed ChIP experiments to monitor H3K9me2 levels in ago1Δ cells . ago1 deletion reduces H3K9me2 accumulation at centromeres below that of wild type cells , but above that of clr4Δ cells . On reintroduction of ago1+ , centromeric H3K9me2 levels accumulate to normal levels ( Figure 7C ) , similar to the results seen on reintegration of other RNAi components into tas3WG cells . These data suggest that cells lacking ago1 behave similarly to other RNAi defective strains , and functionally that there is sufficient H3K9me2 in ago1Δ tas3WG cells to drive heterochromatin assembly following reintroduction of ago1+ . We further analyzed centromeric H3K9me2 levels in RNAi defective strains . At the 2 sites tested within centromeric dg and dh repeats , H3K9me2 levels were significantly elevated in ago1Δ cells above the background levels in clr4Δ or ago1Δ clr4Δ cells , and H3K9me2 accumulation at centromeres was similar in all RNAi deficient backgrounds tested . This data would suggest that Ago1 , like other RNAi components , is not acting upstream of Clr-C for heterochromatin assembly . Our demonstration that heterochromatin can assemble following transient depletion of RNAi components in tas3WG cells is suggestive that the RNAi pathway is acting downstream of Clr-C . However , given that low levels of centromeric H3K9me2 are maintained in RNAi-defective cells , it is difficult to assess whether RNAi is required for the initial step in heterochromatin initiation . To address this question , we removed residual H3K9me2 from RNAi defective cells by introduction of the clr4Δ allele . We then tested whether H3K9me2 could be deposited at centromeres following expression of Clr4 in these cells that lack both Clr4 and Ago1 or Clr4 and Dcr1 ( Figure 8A ) . Following overexpression of clr4+ in ago1Δclr4Δ cells , H3K9me2 could be detected on centromeric repeats above the background observed in clr4 null cells , and similar to levels found normally in ago1Δ cells . Similar results were obtained following overexpression of clr4+ in dcr1Δclr4Δ cells ( Figure 8B ) . Thus , when overexpressed , Clr4 can target centromeric repeats to initiate H3K9me2 deposition in the absence of a functional RNAi pathway . We note , however , that reintroduction of clr4+ into its normal locus in these cells is not sufficient , in the absence of the RNAi pathway , for accumulation of detectable centromeric H3K9me2 ( data not shown ) . Together , these experiments strongly indicate that Clr-C can initiate H3K9me deposition at centromeres via RNAi-independent mechanisms , but that cooperation between RNAi-dependent and RNAi-independent factors normally results in full heterochromatin assembly ( summarized in Figure 9 ) .
We have utilized our novel mutant , tas3WG , to identify genetic requirements for heterochromatin initiation as opposed to those required for the maintenance of pre-existing heterochromatin . We used an approach in which genes required for heterochromatin formation are deleted and reintroduced and then heterochromatin assembly is examined . In wild-type cells , heterochromatin can be established regardless of the factor removed , indicating that establishment mechanisms are robust to perturbations of the system . However , in cells harboring a disrupted RITS complex , we found that the establishment of silencing becomes sensitive to the prior presence of particular silencing factors . Our data demonstrate that the ability to assemble heterochromatin in such gene removal-restoration experiments in tas3WG cells correlates with the prior presence of H3K9me2 on centromeric repeats , but does not require the prior presence of small RNA species . These data strongly suggest that RNAi-independent mechanisms of recruitment of Clr-C play a key role in the assembly of centromeric heterochromatin . We also tested whether RNAi is required in an obligate manner to initiate de novo heterochromatin assembly in cells that lack any prior H3K9me . To accomplish this , we generated clr4Δ dcr1Δ cells and found that some deposition of H3K9me2 at centromeric sequences occurred upon overexpression of clr4+ ( Figure 8C ) . These experiments clearly demonstrate that Clr-C can function to initiate de novo centromeric heterochromatin assembly independently of the RNAi pathway . Recently , Ago1 and its associated priRNAs have been proposed to trigger heterochromatin formation . This notion is partly based on an observation that an ago1Δ strain had little or no H3K9me2 at centromeres , suggesting an upstream role for Ago1 [20] . If this hypothesis were correct , we anticipated that in our tas3WG system , that transient depletion of Ago1 might block heterochromatin assembly similar to what we observed for Clr4 . In contrast , we found that tas3WG cells formed robust centromeric heterochromatin on reintroduction of ago1+ into ago1Δ cells ( Figure 7A–7C ) . Consequently , we re-examined the reported critical Dicer-independent role for Ago1 in driving H3K9Me . In contrast to a recent study [20] , we found that H3K9me2 levels in ago1Δ cells were no lower than in other RNAi-defective backgrounds ( Figure 7D and 7E ) . To further probe for a potential role of priRNAs in heterochromatin initiation , we generated cells that lack both Ago1 ( which binds priRNAs ) and Clr4 , and tested whether reintroduction of Clr4 could promote de novo centromeric H3K9me2 in the absence of Ago1-priRNA targeting activities . This experiment revealed that indeed Clr4 , when overexpressed , can initiate H3K9me2 deposition at centromeres independent of Ago1 ( Figure 8C ) . Taken together , our data demonstrate ( 1 ) that Clr-C components can act independently of members of the RNAi pathway to initiate heterochromatin assembly and ( 2 ) the ability to promote heterochromatin assembly in tas3WG cells correlates with the prior levels of centromeric H3K9me2 in mutant backgrounds , and not the initial small RNA abundance and ( 3 ) , that Ago1-bound priRNAs are unlikely to be the key initiator of heterochromatin assembly . The functional data that we present therefore counters the model for heterochromatin initiation proposed recently [20] , and supports that RNAi-independent factors , together with the RNAi pathway , are necessary for full heterochromatin assembly . The conclusions that derived from our observations contrast with the widely held belief that RNAi initiates heterochromatin assembly at fission yeast centromeres . Although it has been shown by several labs that small RNAs derived from exogenous hairpin RNAs can induce silencing of genomic loci [37] , [38] , these effects tend to be very weak and very locus specific . In these experiments , silencing efficiency correlates with proximity to sites of heterochromatin , or is enhanced by overexpression of heterochromatin proteins . The production of the majority of centromeric small RNAs depends on the presence of heterochromatin . However , low levels of small RNAs are found in Clr-C deletion backgrounds [12] , [20] or in histone H3K9R mutant cells [19] . Interestingly , Clr-C mutants that completely lack H3K9me are deficient for heterochromatin establishment in our reintegration assay , in spite of the presence of centromeric siRNAs ( which are below the level of detection of our Northern assay ) . In contrast , RNAi-defective strains that are devoid of , or express even lower levels of centromeric small RNAs than Clr-C mutants [12] , [19] , [20] , can assemble heterochromatin effectively following reintegration of the wild type gene into the tas3WG background . Although not easy to detect , priRNAs , which have been postulated to prime heterochromatin establishment , are expected to be present in all of the genetic backgrounds that we tested for heterochromatin initiation . This class of small RNAs therefore does not appear to contribute to the differential ability of mutants to initiate heterochromatin assembly in the tas3WG background . Finally , our data showing that transient depletion of ago1+ does not impair heterochromatin establishment in tas3WG cells , and that ago1Δ cells retain H3K9me2 , would argue that ago1+ and priRNAs are not the initiating trigger for heterochromatin assembly . These results beg the question of how low levels of siRNA synthesis occur in the absence of heterochromatin , since early models suggested that localization of the RITS and RDRC complexes to centromeres was a prerequisite for centromeric siRNA generation , and that RITS and RDRC complex localization was dependent on Clr4 [8] , [11] . One study suggested that single-stranded transcripts from centromeric sequences can adopt secondary structures to yield dsRNA that can be targeted by Dcr1 to form siRNAs [19] . Such models for the heterochromatin-independent synthesis of centromeric siRNAs may now help to explain our previously puzzling result that Chp1 chromodomain mutants that are defective for heterochromatin establishment following transient depletion of clr4+ express abundant siRNAs [6] . How low levels of H3K9me2 are initially placed at centromeres remains an open question , but our data supports that it is not absolutely dependent on small RNAs and the RNAi pathway . We suggest that H3K9me2 deposition is linked to the transcription of the centromere . Mutants in three separate components of the RNA pol II complex show defects in heterochromatin assembly , including a mutant that truncates the C terminal repetitive tail of the largest subunit of the polymerase , Rpb1 [16]–[18] . This raises the intriguing possibility that , similar to other histone modifying enzymes such as the Set1 and Set2 methyltransferases , that Clr-C may associate with , and be brought to chromatin , via RNA polymerase II [39] . Another mutation in RNA pol II that causes defective heterochromatin assembly resides in the Rpb7 subunit [16] , which together with its partner , Rpb4 , is thought to be an accessory and non-obligate component of yeast RNA pol II . One interesting possibility would be if Clr-C recruitment to regions of chromatin that are destined to become heterochromatic was controlled by modulation of RNA pol II by the Rpb4/7 subcomplex . In contrast , in plants , two pol II-related RNA polymerase activities , Pol IV and Pol V , have evolved to mediate heterochromatin assembly on repetitive sequences ( reviewed in [40] ) .
Integration plasmids for genomic clones were constructed by PCR using Phusion polymerase ( NEB ) and standard cloning or Gateway ( Invitrogen ) techniques . Oligonucleotide sequences are listed in Table S2 . Full details of plasmid construction are listed in Text S1 . Strains used in this study are listed in Table S1 and details of their construction and verification are in Text S1 . Cells were cultured overnight at 25°C in rich YES medium to a density of approximately 5×106 cells/ml . Cells were washed extensively in PMG media , counted , and five-fold serial dilutions made , such that plating of 4 ul of cells yielded 1 . 2×104 cells within the most concentrated spot . Plating was performed on PMG complete media , PMG media lacking uracil , and PMG complete media supplemented with 2g FOA per liter as described previously [21] , and incubated for 5 days at 25°C . Transcript and siRNA analyses were performed as previously described [10] , [21] . Oligos for real time PCR analysis: ( dh ) JPO-769 and JPO-770 , ( dg ) JPO-986 , JPO-987 , adh1 , JPO-793 and JPO-794 [10] . RNA was prepared from duplicate cultures for every experiment , and for analyses following gene reintegration , multiple independent re-integrants were assessed . Real-time PCR was performed on an Eppendorf Mastercycler ep Realplex machine using Quantifast Sybr green ( Qiagen ) . Data was analyzed using the ΔCt method , ensuring that all samples gave Ct values within the experimentally determined linear range . Chromatin immunoprecipitation was performed as previously described [10] , [21] , using antibodies that recognize H3K9me2 ( Abcam ) and Chp1 ( Abcam ) . Further details are in Text S1 .
|
Centromeres are the chromosomal regions that promote chromosome movement during cell division . They consist of repetitive DNA sequences that are packaged into heterochromatin . Disruption of centromeric heterochromatin leads to chromosome loss that can result in miscarriages and genetic disorders . We have sought to define the precise steps leading to heterochromatin assembly using fission yeast as the model system . To accomplish this we employed our novel Tas3WG mutant strain that can propagate preassembled heterochromatin but cannot support its de novo establishment . Current models suggest that small RNAs initiate heterochromatin assembly by targeting the RNAi machinery and subsequently the Clr-C chromatin-modifying complex to the centromere . Here , we demonstrate that transient depletion of components of the RNAi pathway that generate or bind small RNAs does not perturb heterochromatin assembly in our Tas3WG strain . Instead , transient depletion of the Clr-C complex blocks heterochromatin assembly , suggesting a critical role for continuous Clr-C activity during heterochromatin assembly in Tas3WG cells . We have directly tested whether Clr-C can target centromeres when expressed in cells deficient for RNAi and Clr-C . We find that RNAi–independent recruitment of Clr-C can occur and likely contributes to the critical initiating mechanisms of heterochromatin assembly .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"molecular",
"biology/histone",
"modification",
"molecular",
"biology/rna-protein",
"interactions",
"molecular",
"biology/centromeres",
"molecular",
"biology/chromosome",
"structure",
"genetics",
"and",
"genomics/chromosome",
"biology",
"genetics",
"and",
"genomics/epigenetics",
"molecular",
"biology/chromatin",
"structure"
] |
2010
|
Continuous Requirement for the Clr4 Complex But Not RNAi for Centromeric Heterochromatin Assembly in Fission Yeast Harboring a Disrupted RITS Complex
|
Recent information has revealed the functional diversity and importance of mitochondria in many cellular processes including orchestrating the innate immune response . Intriguingly , several infectious agents , such as Toxoplasma , Legionella , and Chlamydia , have been reported to grow within vacuoles surrounded by host mitochondria . Although many hypotheses have been proposed for the existence of host mitochondrial association ( HMA ) , the causes and biological consequences of HMA have remained unanswered . Here we show that HMA is present in type I and III strains of Toxoplasma but missing in type II strains , both in vitro and in vivo . Analysis of F1 progeny from a type II×III cross revealed that HMA is a Mendelian trait that we could map . We use bioinformatics to select potential candidates and experimentally identify the polymorphic parasite protein involved , mitochondrial association factor 1 ( MAF1 ) . We show that introducing the type I ( HMA+ ) MAF1 allele into type II ( HMA− ) parasites results in conversion to HMA+ and deletion of MAF1 in type I parasites results in a loss of HMA . We observe that the loss and gain of HMA are associated with alterations in the transcription of host cell immune genes and the in vivo cytokine response during murine infection . Lastly , we use exogenous expression of MAF1 to show that it binds host mitochondria and thus MAF1 is the parasite protein directly responsible for HMA . Our findings suggest that association with host mitochondria may represent a novel means by which Toxoplasma tachyzoites manipulate the host . The existence of naturally occurring HMA+ and HMA− strains of Toxoplasma , Legionella , and Chlamydia indicates the existence of evolutionary niches where HMA is either advantageous or disadvantageous , likely reflecting tradeoffs in metabolism , immune regulation , and other functions of mitochondria .
Mitochondria are highly dynamic organelles that , besides producing energy and regulating apoptotic and Ca2+ signals , play a key role in orchestrating several aspects of cell function and behavior including autophagy , apoptosis , and immune signaling [1]–[3] . Interestingly , recent evidence indicates that mitochondrial dynamics , including distribution , size , and shape , are linked to altered regulation of certain innate immune responses [4] , [5] . Given this diversity of function , it is no surprise that mitochondria have recently emerged as central to host–pathogen interactions , with an increasing number of viruses and bacteria having been shown to manipulate host mitochondria in their roles in apoptosis , energy production , and immune function . For example , the hepatitis C virus protease NS3/4a cleaves the N-terminal fragment of the mitochondrial antiviral signaling protein ( MAVS ) , rendering infected cells unable to induce interferon ( IFN ) production [6] , and Helicobacter pylori vacuolating cytotoxin ( VacA ) engages the machinery involved in regulation of mitochondrial fission to induce apoptosis [7] . In light of these newly established roles for mitochondria and their targeting by different microbial effector proteins , it is interesting that during infection with certain pathogens host mitochondria associate with and appear sequestered at the vacuole in which the microbes reside . This phenomenon has been previously reported during infection with the bacteria Legionella pneumophila [8] and Chlamydia psittaci [9] and the parasite Toxoplasma gondii [10] , and this occurs both in mammalian ( T . gondii , L . pneumophila , and C . psittaci ) and protozoan host cells ( L . pneumophila in Hartmannella vermiforis [11] ) . Although the mechanism and functional consequence of this association during infection with these bacteria remain unknown , it has been reported that this association is species-specific [9]—for example , C . psittaci , but not Chlamydia trachomatis , intimately associates with mitochondria . In the case of T . gondii , three predominant , clonal lineages have been described based on genotyping performed on isolates from human and animal infections in North America and Europe [12] . Referred to as type I , type II , and type III , these lineages differ widely in a number of phenotypes including migratory capacity [13] , cytokine production in infected cells [14] , [15] , and virulence in mice [16] . Whether these lineages also differ in their ability to recruit host mitochondria , however , has not previously been assessed . Since the initial discovery of host mitochondrial association ( HMA ) in macrophages infected with ( type I ) Toxoplasma tachyzoites in 1972 [10] , the question of what parasite factor mediates this association has been of intense interest . Previous work suggested that Toxoplasma HMA was mediated by Toxoplasma rhoptry protein 2 ( ROP2 ) [17] . However , subsequent work showed that parasites lacking ROP2 expression were indistinguishable from wild-type parasites in their ability to recruit host mitochondria , reopening the question of what factors mediated the association [18] . The functional significance of the recruitment of mitochondria to the Toxoplasma vacuole also remained a conundrum . Although HMA has long been assumed to reflect a crucial means by which the parasite acquires key metabolites , only limited evidence has emerged to support this hypothesis [19] , [20] . Given our expanded understanding of the diverse roles performed by mitochondria , however , we hypothesized that HMA confers a selective advantage to the various infectious agents where it has been described and that mitochondrial functions other than metabolism might also contribute to this advantage . Here , we report that HMA is strain-specific in Toxoplasma and that it is mediated by a novel , secreted parasite factor , mitochondrial association factor 1 ( MAF1 ) , which differs in sequence , gene copy number , and expression between the three canonical strains . We demonstrate that MAF1 is necessary and sufficient for HMA and go on to show that during Toxoplasma infection , HMA is associated , in vitro and in vivo , with substantially altered levels of cytokines . These results demonstrate that HMA is an additional means by which Toxoplasma tachyzoites interface with the immune signaling of the host and suggest that , in addition to possible metabolic roles , HMA may present a novel strategy for subversion of host immune signaling by a pathogen .
Given that the three predominant strains of Toxoplasma differ widely in a number of immune-related phenotypes [21] and that mitochondria have recently been shown to play a key role in orchestrating the cellular immune response to viral infection [22] , [23] , we hypothesized that HMA might differ between these canonical strains . To investigate this possibility , human foreskin fibroblasts ( HFFs ) were labeled with MitoTracker and infected with representative type I ( RH ) , type II ( Me49 ) , or type III ( CEP ) strains of T . gondii . At 4 h postinfection ( hpi ) , the type I and III tachyzoites showed the canonical , intimate , and extensive association of host mitochondria at the parasitophorous vacuole membrane ( PVM ) –host interface , whereas type II parasites showed little if any HMA ( Figure 1A–C ) . To quantify and expand on these findings , the percentage of the PVM associated with host cell mitochondria was measured in electron micrographs of HFFs 4 hpi with type I , II , and III parasites ( Figure 1D–F ) . The results showed that ∼35% and ∼18% of the PVM is tightly associated with host cell mitochondria in type I and III strains , respectively , in contrast to <2% of the PVM in infections with type II parasites ( Figure 1G ) . Although host endoplasmic reticulum ( ER ) has also been reported to associate with the PVM [24] , the abundance and size of the ER meant we were unable to confidently assess or quantify such association and so cannot comment on which strains of Toxoplasma do or do not exhibit this trait . To ensure the generality of the HMA− phenotype in type II strains , HMA was also assessed in an independent type II strain , Pru , which was isolated in France ( Me49 was isolated in the United States of America ) ; Pru was also found to be HMA− ( Figure S1 ) , demonstrating that the HMA− phenotype is not an isolate-specific property of Me49 . Similarly , to determine whether HMA differed among cell lines of different mammalian and tissue origins , we tested the phenotype in murine macrophages and Madin–Darby canine kidney cells and obtained the same results as in the HFFs: type II was consistently HMA− , whereas types I and III were HMA+ ( unpublished data ) . To determine if HMA occurs during in vivo infection , peritoneal exudate cells ( PECs ) were isolated from mice infected with type I and type II parasites and processed for electron microscopy analysis . Quantification of HMA in infected PECs ( Figure S2 ) confirmed that HMA occurs in animals infected with type I but not type II strains . These results show that recruitment of host mitochondria to the Toxoplasma PVM is a strain-specific property of Toxoplasma in vitro as well as in vivo . Previous work has utilized genetic crosses between the canonical Toxoplasma lineages to map the loci responsible for major phenotypic differences [25] . We sought to test our hypothesis that strain-specific HMA had a genetic basis by utilizing F1 progeny from a type II×III cross [26] , as these lineages differ in their ability to recruit mitochondria and their progeny have been previously genotyped at 135 different markers [27] . To facilitate HMA screening in the F1 progeny , we first determined whether co-infection with an HMA− strain and HMA+ strain could alter the HMA phenotype of either parasite . The results showed that co-infection of HFFs with strains that differ in HMA did not affect the phenotype: type II parasites were within HMA− vacuoles , whereas types I and III showed efficient HMA ( Figure S3 ) . To reliably determine the HMA phenotype of the F1 progeny , therefore , HFFs were co-infected with each of the 34 F1 progeny and a GFP expressing HMA+ control . This allowed for a direct comparison and qualitative assessment of the HMA phenotype between each progeny and the HMA+ control within the same cell . In all , 15 progeny were HMA− , 18 were HMA+ , and one was indeterminate; example HMA+ and HMA− progeny are shown in Figure 2A and Figure 2B , respectively . To identify the genomic region ( s ) mediating HMA , we searched for association between the previously mapped Toxoplasma genetic markers [27] and the HMA phenotype and found a highly significant co-segregation between HMA and chromosome II ( LOD ∼6 . 3; p<0 . 001 , Figure 2C ) ; only F1 progeny having type III alleles at chromosome II were HMA+ ( Figure S4 ) . To refine the preliminary mapping , we analyzed the genotypes of two progeny with recombinations on chromosome II , E7 and STG2 ( Figure S4 , boxed region ) , which indicated that the region containing the parasite factor that mediates HMA is in the interval between the left end of chromosome II and marker KT-L379A , a maximum size of ∼1 . 2 Mb . To choose among the list of 153 genes present in the region identified on chromosome II as possibly involved in HMA , we used a candidate gene approach based on properties predicted for the parasite factor . HMA occurs at the PVM and is strain-specific; thus , we used bioinformatics to first identify genes that encoded a predicted signal peptide and a transmembrane ( TM ) domain , and had different expression levels between HMA+ ( types I and III ) and HMA− ( type II ) strains . Of the 10 genes that fulfilled these criteria , the type I gene TGGT1_053770 ( denoted in ToxoDB v7 . 2; as TGME49_020950 in type II and TGVEG_051550 in type III; Figure 3A ) stood out as the most promising for multiple reasons: ( 1 ) This gene had been previously identified in proteomic studies as a secreted Toxoplasma antigen [28] , suggesting that it was capable of interacting with the host cell; ( 2 ) microarray data for 17 of the progeny phenotyped for HMA showed that this transcript was of generally higher abundance in the 9 F1 progeny that were HMA+ relative to the 8 F1 progeny that were HMA− ( Figure 3B ) [29]; ( 3 ) we had also observed in prior studies that the protein encoded by TGGT1_053770 is heavily phosphorylated in infected cells relative to extracellular parasites [30] , a further indication that it might be secreted into the PV or host cell , and we confirmed postinvasion phosphorylation of TGGT1_053770 by comparing electrophoretic migration after addition of phosphatase in intracellular and extracellular parasite lysates ( Figure 3C ) ; and ( 4 ) lastly , as with Toxoplasma ROP5 , which is crucial for type I virulence in mice [31] , there is precedent for gene family expansions playing a key role in interfacing with host immune defenses [32] , [33] and determining pathogen host range [34] . The TGGT1_053770 locus is expanded in T . gondii: based on raw genomic sequence coverage , type I ( GT1 ) has ∼6–10 copies , whereas types II ( ME49 ) and III ( VEG ) have ∼4–5 copies ( Figure 3D ) . To determine if the protein encoded by TGGT1_053770 traffics to the PVM , a property expected of a protein mediating HMA , HFFs were infected with type I parasites engineered to ectopically express an N-terminally HA-tagged version of the protein driven by its own promoter and with the HA-tag placed just downstream of the putative signal peptide . Using anti-HA antibody , the tagged protein was indeed found to localize at the PVM ( Figure 3E ) . This localization was confirmed by raising antibodies to recombinant TGGT1_053770 and showing that these antibodies likewise stained the PVM of cells infected with type I parasites ( Figure 3F ) . There was near-perfect co-localization of the anti-HA antibodies and anti-TGGT1_053770 antibodies in cells infected with type I parasites expressing TGGT1_053770 , showing that the N-terminal tagging did not detectably alter the protein's trafficking in the infected cells ( unpublished data ) . Given that the PVM and intracellular staining pattern was consistent with that of a dense granular protein , we also looked with antibodies to dense granule protein 7 ( GRA7 ) and anti-HA antibodies and found that there was co-localization between these two proteins in extracellular parasites expressing the TGGT1_053770 transgene ( Figure 3H ) ; hence , TGGT_053770 appears to encode a dense granule protein . To compare expression of this locus between the archetypal strains , Western blot analysis of extracellular lysates of type I , II , and III tachyzoites was performed using anti-TGGT1_053770 antibodies . The results showed that although expression levels were comparable in HMA+ type I and III parasites ( albeit the type III protein migrates slightly but reproducibly more slowly ) , type II parasites showed no detectable expression ( Figure 3G ) , consistent with the microarray data in Figure 3D . To determine if this failure to detect a signal in lysates from type II parasites might be due to a gross difference between the allele in this strain relative to TGGT1_053770 , we PCR-amplified and sequenced the locus from RH ( type I ) and Me49 ( type II ) genomic DNA . Although the tandem duplications of this locus make it difficult to determine the true extent of sequence variation within and between strains , the results showed a 94 . 5% predicted amino acid identity between the two types ( Figure S5 ) , making it very unlikely that polyclonal antisera raised to one would not react with the other . Collectively , the data above indicate that TGGT1_053770 has the expected properties of a parasite mediator of HMA in being a tandemly repeated locus that differs in sequence , copy number , and expression between strains and that encodes a novel dense granule phosphoprotein found at the PVM . To directly test the hypothesis that TGGT1_053770 mediates HMA , we infected HFFs with type II parasites or type II parasites engineered to express the N-terminally HA-tagged type I allele of this gene ( referred to here as type II:MAF1 ) . At 12 hpi , cells were fixed and stained with anti-TGGT1_053770 polyclonal sera along with antibodies to Translocase of the Outer Membrane 20 ( TOM20 ) , a marker of host mitochondria . The results ( Figure 4A ) showed that expression of TGGT1_053770 in type II tachyzoites is associated with a dramatic gain in HMA , with the recruited mitochondria being specifically associated with the portions of the PVM that stained with anti-TGGT1_053770 . Electron microscopy ( Figure 4B ) confirmed that the HMA+ phenotype of type II:TGGT1_053770 parasites results in a marked increase in the levels of mitochondrial recruitment; on average , ∼30% of the PVM surface appears associated with host mitochondria when infecting with these parasites compared to <2% with type II parasites ( Figure 4C ) . The results above show that TGGT1_053770 encodes a novel protein that is sufficient for recruitment of host mitochondria to the vacuole containing type II Toxoplasma parasites and confirm the role of TGGT1_053770 in mediating HMA . We will hereafter refer to TGGT1_053770 as MAF1 . To further explore the role of MAF1 in HMA , type I parasites lacking MAF1 were generated ( Δmaf1 ) and verified as deleted for MAF1 via PCR , Western blot analysis , and microscopy ( Figure S6 ) . Co-infection experiments using type I and type I:Δmaf1 parasites showed the predicted HMA dependency on MAF1 ( Figure 4D ) ; type I parasites exhibit HMA in contrast to the clear loss of HMA in type I:Δmaf1 parasites . Electron microscopy ( Figure 4E ) was used to quantify the difference in HMA: only ∼3% of the PVMs surrounding type I:Δmaf1 were associated with host mitochondria , in contrast to ∼35% for type I PVMs ( Figure 4C ) . These results indicate that MAF1 is necessary for HMA during type I infection . We next sought to investigate whether MAF1 is directly responsible for HMA or whether a parasite partner protein is involved . To do this , we imaged mouse embryonic fibroblasts ( MEFs ) retrovirally transfected with an expression construct encoding N-terminally HA-tagged MAF1 . IFA analysis of a transfected population showed a clear co-localization of MAF1 with TOM20 , a mitochondrial membrane marker ( Figure 5A ) , indicating that MAF1 is directly responsible for HMA . Although BLAST searches yielded no clues as to the coding function or region of MAF1 responsible for HMA , we hypothesized that during Toxoplasma infection the predicted TM domain anchors MAF1 at the PVM , allowing for the MAF1 C-terminal domain to be exposed to the host cell cytosol , where it tethers host mitochondria . To test this hypothesis , type II parasites were engineered to express MAF1 bearing a C-terminal HA epitope tag ( MAF1_CHA ) . The HA tag served a dual purpose: to allow for recognition of the expressed transgene as well as function as a steric block if an exposed C-terminus is necessary for HMA . The results showed that the type II:MAF1_CHA parasites remain HMA− , suggesting a native C-terminus is necessary for MAF1 function ( unpublished data ) . Surprisingly , in type I parasites engineered to express MAF1_CHA , the C-terminally HA-tagged MAF1 acts as a dominant negative; its expression results in a dramatic loss of HMA at 4 hpi even though it correctly localizes to the PVM ( Figure 5B and C ) . These data suggest that the C-terminal domain of MAF1 is key to its role in HMA , although we cannot exclude an indirect effect of the C-terminal epitope tag . The identification of MAF1 as the molecule that Toxoplasma uses to mediate HMA enabled us to next address the functional consequences of HMA in the host–pathogen interaction . HMA has been hypothesized to serve as a source of metabolites for the parasite , and previous work has shown that Toxoplasma scavenges host-derived mitochondrial lipoic acid [19] . To determine the role of HMA during tachyzoite growth within a host cell , type I versus type I:Δmaf1 proliferation was compared in a competitive growth assay in HFFs in vitro . At 7–10 d postinfection ( dpi ) , we observed no significant growth disadvantage in type I:Δmaf1 parasites due to the loss of HMA . A similar comparison of growth of type II versus type II:MAF1 parasites also yielded no reproducible change in their relative numbers 7–10 dpi ( unpublished data ) . Collectively , these data support the conclusion that HMA does not result in growth differences between type I and type II parasites in a nutritionally replete environment , although this does not exclude a metabolic role for HMA . As previously mentioned , HMA occurs during infection with the vacuolar pathogens C . psittaci and L . pneumophila . To address the consideration that HMA represents a conserved host response to vacuolar pathogens and might be the result of toll-like receptor ( TLR ) or IFN-dependent signaling , we infected bone marrow isolated from MyD88−/− , IFNAR−/− , and wild-type mice with type I and type II parasites ( MyD88 is a common part of the signaling pathway for many TLRs , and IFNAR is the shared receptor used by type I IFNs ) . Electron microscopy showed no differences in HMA in these mutant bone marrow cells relative to wild-type cells ( unpublished data ) . Given our growing understanding of how different Toxoplasma strains vary in their interaction with the host , we hypothesized that HMA might be involved in the strain-specific ability of Toxoplasma to alter innate immune signaling in infected host cells . Interestingly , early observations of HMA described the presence of “giant mitochondria” at the Toxoplasma PVM during infection [35] , and more recently , the proteins that regulate mitochondrial morphology have been found to play a crucial role in modulating cellular immune signaling during viral infection [36]–[38] . To determine if HMA results in morphological differences in the associated host mitochondria , the average sizes of unassociated ( cytosolic ) mitochondria and PVM-associated mitochondria in mouse bone-marrow-derived macrophages ( mBMDMs ) infected with type I parasites were measured in electron micrographs . The analysis was also done during type II:MAF1 infection for confirmation . The results showed that in mBMDMs infected with type I or type II:MAF1 parasites , the average cross-sectional area of a PVM-associated mitochondrion is approximately 3-fold greater than the average for cytosolic mitochondrion ( Figure 6A and B ) . These changes in morphology are consistent with the elongation of mitochondria observed during viral infection [37] and support a general role for alterations in mitochondrial morphology having an impact in microbial infection . Previous work has shown that signaling through the mitochondrion can alter transcription of proinflammatory cytokines [39] . To investigate the role of HMA in the immune response to Toxoplasma , microarrays were used to compare the transcriptional profiles of MEFs 8 hpi with type II and type II:MAF1 parasites . The microarray analysis revealed that the expression of several chemokines and IFN-gamma-induced genes differed significantly in a MAF1-dependent manner ( Figure 6C ) and that three of the greatest transcriptional changes were seen for the proinflammatory cytokines CXCL2 , CSF2 , and IL6 , with respective increases in expression in the type II:MAF1–infected cells of ∼4 . 6- , 4 . 8- , and 5-fold . Using Gene Set Enrichment Analysis ( GSEA ) , we found that canonical immune pathways , including JAK-STAT , NFAT , and IL10 , were among the most significantly enriched gene sets up-regulated during type II:MAF1 infection relative to type II wild type ( Figure S7 ) . Although we cannot exclude the possibility that MAF1 mediates these effects independent of its role in HMA , these data suggest that HMA in type II parasites has a significant impact on the host cell's transcriptional response to Toxoplasma , especially in terms of innate immune signaling . To explore this issue further , levels of 26 cytokines , chemokines , and growth factors were measured from supernatants of mBMDMs 3 hpi and 17 hpi with type II and type II:MAF1 parasites using the Luminex Mouse 26-plex ( Meso Scale Discovery , services rendered by Stanford Human Immune Monitoring Center ) . mBMDMs were used because monocytes have been shown to account for approximately 60% of infected cells during a peritoneal infection with Toxoplasma and are required for immunity to this parasite [40] . At 3 hpi , no significant differences were seen between type II and type II:MAF1–infected mBMDMs ( unpublished data ) , whereas at 17 hpi , we observed increased secretion of IL6 and IL10 and the proinflammatory chemokines RANTES ( CCL5 ) and the murine IL-8 homologue KC ( CXCL1 ) in cells infected with type II:MAF1 ( Figure 6D ) . Together , these results argue for a key role for HMA in modulation of the innate immune response during infection . We and others have previously shown that type I parasites generate a profoundly different immunological response compared with type II parasites , which activate NF-κB and result in increased IL-12 production [14] , [15] , [41] , [42] . To avoid the possibility that the impact of type II on other immune pathways might interfere with our interpretation of the results , we addressed whether the loss of HMA in its natural context ( type I background ) impacted infection in vitro and in vivo . To test this , WT MEFs were infected with type I and type I:Δmaf1 parasites , and at 8 hpi , the RNA was isolated and processed for microarray analysis . Interestingly , unlike the 231 genes differentially regulated between type II:MAF1 and type II infection , only eight genes were observed to be significantly induced in a MAF1-dependent manner during type I infection , including SerpinB2 and IL-11 ( Figure S8 ) ; the increased expression of SerpinB2 and IL-11 receptor alpha ( IL-11rap ) was also MAF1-dependent during type II:MAF1 infection ( Figure 6C ) and might represent part of a HMA-induced signaling pathway in host cells . To address whether the loss of HMA in the type I background impacted virulence in the murine host , we compared lethality following intraperitoneal infection of C57/BL6 mice with type I and type I:Δmaf1 tachyzoites and observed no differences ( Figure 7A ) . This was expected , as previous analyses of the same II×III F1 progeny used to map HMA , as well as crosses of type I with types II and III , showed no indication of a significant mouse-virulence trait on chromosome II , the MAF1-encoding chromosome [43]–[45] . Although our results confirm that MAF1 is not a major virulence determinant using the crude measure of mortality by intraperitoneal infection in mice , they did not address the possibility of other differences in the host response to infection by type I and type I:Δmaf1 parasites . To test for such effects , PECs were isolated from mice 5 dpi with type I or type I:Δmaf1 tachyzoites and their cytokine production assessed on the Luminex platform at 6 h and 12 h postisolation . The results showed differences ( p<0 . 05 ) in two cytokines , IL4 and eotaxin , and interestingly , both were present at lower levels in cultured supernatant of PECs from the type I:Δmaf1–infected mice compared to mice infected with type I ( Figure 7B ) parasites . To test the effect on the animal as a whole , sera from similarly infected mice were also analyzed 5 dpi . We observed that , as with IL4 and eotaxin secretion from isolated PECs , 14 of the 26 cytokines examined were present at lower levels in mice infected with type I:Δmaf1 relative to type I parasites ( Figure 7C ) . These differences included IL6 , IL10 , and RANTES ( CCL5 ) , which were increased in a MAF1-dependent manner during type II versus type II:MAF1 infections in mBMDMs ( Figure 5D ) . To determine if the differential cytokine production observed was a result of differences in parasite burden , mice were infected with RFP-expressing type I and type I:Δmaf1 tachyzoites . At 5 dpi , the same time point used for the sera analyses , PECs were isolated and analyzed for RFP expression using flow cytometry and shown to exhibit comparable parasite levels ( Figure 7D ) . Thus , although the MAF1 genotype appears to have no effect on lethality or parasite burden during the acute stages of an i . p . infection , these results suggest some role for MAF1 in modulating the immune response and support for the notion that HMA is a novel strategy by which an intracellular pathogen can interfere with host signaling .
It is well known that the three archetypal T . gondii strains that dominate in Europe and North America differ in several important host interactions including the ability to inactivate mouse immunity-related GTPases ( IRGs ) that otherwise attack the PVM [47] and to phosphorylate host signal-transducers and activators of transcription ( STATs ) [48] , [49] . We show that HMA is yet another dramatic way in which the three dominant genotypes differ and so is part of a complicated calculus in the parasite's evolution with the optimal solution ( HMA+ or HMA− ) likely depending on the interplay with other factors . For example , it could be that there are strain-specific differences in the need for host metabolites produced by the mitochondria ( e . g . , lipoic acid [19] ) and/or susceptibility to the damage mitochondrial metabolites can cause ( e . g . , reactive oxygen species [50] ) . Of note , type II strains but not types I and III encode a version of the soluble factor , GRA15 , that stimulates NF-κB production in infected cells [14] . It may be that a combination of HMA and NF-κB stimulation would be suboptimal to the parasite compared to either on its own , but the extraordinary success of type II in certain regions of the world clearly indicates that this strain is well optimized for at least some niche . It is also possible that the function served by HMA in types I and III is provided by a completely different mechanism in type II , but until that function is known , this is difficult to speculate about . Most models predict that different traits evolved in certain Toxoplasma strains because of their interaction with a particular host species or host state . IRGs , for example , are crucial parts of the immune armamentarium in mice but are functionally lacking in primates , possibly explaining the polymorphisms seen in the Toxoplasma effectors , ROP5 and ROP18 , that neutralize them . MAF1 , like ROP5 or ROP18 , may have evolved as a crucial defense to the immune response in a specific host . We note that the dramatic difference in murine fibroblast transcription observed in type II:MAF1 versus type II infection was not observed in type I versus type I:Δmaf1 infection . This is likely due to the very different host responses seen in cells infected with type I versus type II parasites . We saw no difference in HMA in the three species of host cells examined ( human , mouse , and dog ) , but we cannot exclude that some untested host exists where type II tachyzoites do show HMA , although this would likely require host-specific activation of transcription from the MAF1 locus given that MAF1 expression in type II parasites is essentially quiescent when grown in vitro . The genes responsible for the HMA seen during Legionella and Chlamydia infection have not yet been identified , but no MAF1 homologue is seen in their respective genomes by BLAST analysis ( unpublished data ) . Interestingly , HMA is not seen with all species of these two genera: L . pneumophila and C . psittaci induce the phenomenon , but Legionella micdadei and C . trachomatis do not [9] , [51] , suggesting the presence of a particular niche occupied by the HMA-inducing species of these bacteria where the phenomenon is advantageous . Although it seems likely that HMA is an example of convergent evolution in these bacteria and Toxoplasma , the biological implications of HMA in these bacteria will be difficult to dissect until HMA+ and HMA− versions are engineered in the same genetic background . Although these results demonstrate that MAF1 directly binds host mitochondria and is crucial for HMA , they do not make any prediction as to the mitochondrial binding partner and whether HMA is directly responsible for the observed alterations in the morphology of the associated mitochondria . No other molecule has been described with the ability to mediate HMA , and the sequence of MAF1 did not provide any obvious clues to mechanism . Previous models involving ROP2 as the mediator of HMA proposed that a mitochondrial import signal on the protein's N-terminus is recognized by the import machinery on the mitochondria , and an incomplete , stalled uptake of the protein results in HMA . This exact mechanism seems unlikely given that the MAF1 C-terminus appears to play a key role , as seen by the inability of the MAF1_CHA construct to mediate HMA , but our results would be consistent with a model involving uptake of C-terminal import signals as shown for the yeast DNA helicase Hmi1P [52] . Further understanding of the mechanism by which MAF1 mediates HMA will require identification of the host molecule ( s ) bound by MAF1 . Mitochondria act as a nexus for a wide range of host cellular processes and signaling networks including chemotaxis , calcium signaling , apoptosis , and immune signaling . It has been previously reported that during Toxoplasma infection , over one-third of all modulated host proteins are mitochondrial [53] , [54] , a significant enrichment considering mitochondrial proteins are predicted to account for only 4 . 25% of the cell proteome [54] . This is consistent with HMA having a profound effect on mitochondria during a Toxoplasma infection , similar to what has been reported during bacterial and viral infections where substantial changes are observed in mitochondrial morphology and outer mitochondrial membrane composition [50] , [55] . Our results on the increased cross-sectional area of PVM-associated mitochondria also suggest HMA affects their physiology , although we have not directly assessed this in our experiments here . The identification of the key protein mediating HMA allowed us to show that this phenomenon is associated with a significant alteration of transcription of host cell immune genes . We do not know the mechanism by which HMA results in the transcriptional responses or cytokine responses observed , or whether it is directly responsible for these changes . It is well known , however , that during viral infection , signaling through MAVS results in recruitment of cytosolic factors to the outer mitochondrial membrane; activation of different transcription factors including NF-kB , IRF3 , and IRF7; and subsequent production of type I IFNs and inflammatory cytokines [39] . Although there was a significant enrichment for genes with NF-kB binding sites in their promoters by Transcription Binding Site analysis of the microarray data during type II:MAF1 infection of mBMDMs ( unpublished data ) , there was no evidence for IRF3 or IRF7 activation . Interestingly , although the absence of MAVS did not appear to affect HMA in type I–infected cells , five of the eight genes whose induction or repression was MAF1-dependent in MEFs during type I infection were also MAVS-dependent ( Figure S8 ) . These include SerpinB2 and TSG14 ( tumor necrosis factor–inducible gene 14 ) but not , however , IL11 , suggesting MAVS may play some role in MAF1-induced immune changes in the host cell . Although we cannot exclude the possibility that MAF1 is mediating the observed transcriptional and immune responses independently of HMA , we hypothesize that type I and III Toxoplasma co-opt mitochondrial signaling using HMA to generate a cytokine milieu conducive to parasite survival and efficient dissemination . Interestingly , the cytokines whose levels were increased in HMA+ strains during mBMDM infection in vitro and in PECS ex vivo , including IL6 , CCL5 , and IL10 , are not those that would be expected to change based on interference with known mitochondrial signaling pathways ( i . e . , MAVS , NF-kB , IRF3/7 [22] , [23] ) . These data suggest that HMA might be functioning through a yet unidentified pathway linking mitochondria and innate immune signaling , further highlighting the connection between mitochondria and immunity . It is possible that the observed changes in mitochondrial morphology reflect this manipulation , as it has previously been shown that the machinery that regulates mitochondrial fission and fusion ( e . g . , Mitofusin 1 and 2 ) are also necessary in the signaling cascades that result during viral infection [37] , [38] . Future work will be required to map out the pathways that link HMA to the observed differences in cytokine responses , as well as to determine how the consequences of HMA impact the host–parasite interaction . Our identification of strain-specific differences in HMA during infection with Toxoplasma and the molecular basis underlying those differences provides a model that can be used to dissect the biological role of HMA and the possibly novel ways by which mitochondria engage immune-signaling pathways .
T . gondii parasites of the type I ( RHΔhxgprt ) , type II ( ME49Δhxgprt ) , and type III ( CEPΔhxgprt ) strains [deleted for the hypoxanthine–xanthine–guanine phosphoribosyl transferase ( HXGPRT ) gene] were maintained by serial passage in HFF monolayers . The GFP-expressing strain of RH ( RHgfpluc ) [56] and the F1 recombinant progeny derived from type II ( ME49 ) and type III ( CEP ) [57] crosses used for the co-infections have all been described previously . mBMDMs were obtained from female C57BL/6 mice as previously described [14] and cultured in 20% M-CSF , 10% heat-inactivated FBS , 2 mM L-glutamine , 1 mM sodium pyruvate , 1× DMEM ( Dulbecco's modified Eagle's medium; Invitrogen ) plus nonessential amino acids , and 50 µg/ml each of penicillin and streptomycin . MEFs ( a gift from Z . J . Chen , University of Texas Southwestern Medical Center at Dallas ) and HFFs were grown in complete DMEM ( cDMEM ) supplemented with 2 mM glutamine , 100 U/ml penicillin , and 100 µg/ml streptomycin and 10% heat-inactivated fetal calf serum ( FCS ) . Genomic DNA from either RH88 ( type I ) or ME49 ( type II ) was used as a template in PCR reactions using Platinum Taq HiFi polymerase ( Invitrogen; Life Technologies ) with primers 5′ AGGGATACGAACAACTCGCTTAT 3′ ( forward ) and 5′ GTCCAGCATGCTAGCCAGATACGT 3′ ( reverse ) . PCR was performed under standard conditions , except extension times for each cycle were carried out for 7 min at 68°C to minimize chimera formation between different MAF1 copies . PCR products were gel-purified and cloned into the PCR2 . 1 Topo vector and completely sequenced . Inserts ranged in size from 2 , 995 to 3 , 119 bp and were spliced in silico based on the annotated splice sites for TGGT1_053770 . Sequences were aligned using ClustalW ( http://www . ebi . ac . uk/Tools/msa/clustalw2/ ) and visualized using Jalview . For generation of an N-terminally hemagglutinin ( HA ) -tagged MAF1 expression construct ( pMAF1 ) , the promoter ( ∼1 . 5 kb sequence upstream of the start codon ) was cloned into the pGra-HA_HPT vector [58] at the HindIII and NsiI sites using forward primer CCAAGCTTCTGCGACGTGATCGTGGCAA and reverse primer CCATGCATCGCGTAGTCCGGGACGTCGTACGGGTAACCGGCGGTCAG [reverse primer encodes an HA tag ( bolded ) fused immediately downstream of the region encoding the predicted signal peptide] . The coding region downstream of the signal peptide of MAF1 up to the stop codon was cloned into the NsiI and PacI sites with forward primer CCATGCATGGTGGTCTAGGCAGTCAGATGTCGG [introduces two glycine residues ( bolded ) after the signal peptide] and reverse primer CCTTAATTAATC AGTCCAGCATGCTAGCCAG . Transgenic parasite strains were made by electroporation of the parental type I and II T . gondii strains with 20 µg of linearized pMAF1 and applying mycophenolic acid and xanthine selection to isolate and clonally expand parasites positive for HXGPRT expression [59] . Type I strains transfected with pMAF1 are referred to as type I:TGGT1_053770 or type I:MAF1; type II strains transfected with pMAF1 are referred to as type II:TGGT1_053770 or type II:MAF1 . A construct for knocking out the endogenous MAF1 locus ( pMAF1LKO ) was created from the parental pTKO vector [60] . Briefly , an ∼850 bp region of genomic DNA corresponding to sequences upstream of the TGGT1_053770 start codon ( 5′ homology region or 5′ HR ) and a ∼720 bp region of genomic DNA corresponding to sequences upstream of TGGT1_ 064860 ( 3′ HR ) were cloned into NotI/EcoRV and HpaI/ApaI restriction sites that flank the T . gondii HXGPRT gene and GRA2 3′-UTR in the pTKO vector , respectively . Primers for amplification of the 5′ insert were AAGCGGCCGCCGAGACAGAGAGCAGTGCCAA and AAGATATCGGATGTGGCTCCACTGGTGAAT; for the 3′ insert , they were CCGTTAACGTGAGAGCTCCATCATCGACTCCTT and AAGGGCCCGGTGTGCCCGCTTGGATCAA . The type I:Δmaf1 parasite strain was made by electroporating type I ( ΔhxgprtΔku80 ) parasites with 25 µg of linearized pMAF1KO and selecting for HXGPRT-positive parasites as previously described [30] . HFFs were grown to confluency in 24-well plates . For staining with MitoTracker ( Invitrogen ) , medium was replaced with prewarmed DMEM containing MitoTracker at a concentration of 50 nM . After 30 min of incubation at 37°C , cells were washed and then infected with parasites in prewarmed cDMEM . At 4 hpi , cells were washed in PBS and fixed in prewarmed cDMEM containing 3 . 7% formaldehyde for 15 min . Otherwise , parasites were allowed to invade confluent HFF monolayers on coverslips for 6 and 12 h , fixed in 4% formaldehyde , permeabilized with 0 . 1% Triton-X100 for 20 min , and blocked in PBS supplemented with 3% BSA . Coverslips were then incubated with 3F10 ( anti-HA ) antibody ( Roche , Palo Alto , CA ) , antibodies specific for TOM20 ( SCBT , Santa Cruz , CA ) , or mouse polyclonal sera to MAF1 for 1–3 h at room temperature ( RT ) . Fluorescent secondary antibodies ( Invitrogen/Molecular Probes , Carlsbad , CA ) were used for antigen visualization . Coverslips were mounted in VectaShield ( Vector ) . Phase and fluorescence images were captured with a Hamamatsu Orca100 CCD on an Olympus BX60 ( 100× ) and FV1000 Olympus confocal scope ( DIC ) and processed using Image-Pro Plus 2 . 0 ( MediaCybernetics ) . For MAF1 immunoblotting , cells were lysed in sodium dodecylsulfate sample buffer , and samples were resolved on 8% SDS-polyacrylamide gels and blotted to PVDF membranes ( Millipore , Billerica , MA ) . Following gel transfer , membranes were blocked with PBS-0 . 05% Tween 20 ( PBS-T ) and 0 . 05% nonfat dry milk for an hour and then incubated in mouse anti-MAF1 sera at 1∶250 for 3 h . Following incubation , blots were washed three times in PBS-T and then incubated with horseradish peroxidase ( HRP ) -conjugated anti-mouse antibodies at a 1∶2 , 500 dilution . The membrane was stripped ( Invitrogen stripping buffer ) and reprobed with rabbit anti-SAG1 antibody at a 1∶100 , 000 dilution in blocking solution for 1 h . After three PBS-T washes , the blot was incubated with HRP-conjugated anti-rabbit antibodies at a 1∶3 , 000 dilution and developed using a chemiluminescence system ( Pierce Chemical Co . , Rockford , IL ) . For assessing phosphorylation , lysates from extracellular type I parasites and type I–infected HFFs ( intracellular ) were treated with and without calf intestinal phosphatase ( CIP ) for 1 h and loaded in each lane . The membrane was probed with anti-HA antibody conjugated to peroxidase and developed as stated above . Recombinant MAF1 protein ( amino acids 134–443 ) was expressed with a C-term His-6× tag as previously described [61] and was a gift of M . Tonkin and M . Boulanger ( University of Victoria ) . Blood of naïve female Balb/c mice was drawn and tested against recombinant MAF1I to establish baseline reactivity . Each mouse was then immunized intraperitoneally ( IP ) with 0 . 1 mg of MAF1 antigen in a 1∶1 mixture with RIBI ( Corexia ) in a final volume of 200 ml . Three IP injections of 50 mg in a 1∶1 mixture with RIBI ( final volume 200 ml ) were administered on day 14 , 35 , and 56 . The final bleed was taken at week 13; collected sera were verified by IFA and Western blot analysis . Monolayers of BMMs grown on glass coverslips were synchronously infected with parasites as described above and 6 h later fixed in Karmovsky's fixative—2% glutaraldehyde and 4% paraformaldehyde in 0 . 1 M sodium cacodylate pH 7 . 4—for 1 h at RT . Following this , the cultures were stained with 1% osmium tetroxide for 1 h at RT , washed 3× with ultrafiltered water , then stained in 1% uranyl acetate overnight . After a series of ethanol washes and propylene oxide incubations , the samples were placed into molds with labels and fresh 100% EMbed-812 resin in a 65°C oven overnight . Sections were taken between 75 and 90 nm , picked up on formvar/Carbon coated 75 mesh Cu grids , and stained for 20 s in 1∶1 saturated uranyl acetate ( 7 . 7% ) in acetone followed by staining in 0 . 2% lead citrate for 3 to 4 min . Samples were observed in the JEOL 1400 TEM at 80 kV , and photos were taken using a Gatan Multiscan 791 digital camera . For quantification of HMA , 30 images were taken of PVs in infected cells and 20 ( or as indicated ) were randomly selected for further analysis by ImageJ . Image J was used to measure the percentage of the PVM perimeter closely associated with host mitochondria ( PVM associated with mitochondria/total PVM×100 ) . Raw sequence reads for T . gondii ( strains GT1 , ME49 , and VEG ) were downloaded from the NCBI trace archive in FASTA format . Reads were aligned to the T . gondii ME49 genome ( version 7 . 0; ToxoDB ) using BLAT with the following parameters: -fastMap -minIdentity = 95 -minScore = 90 [62] . Following conversion of the blat output file ( psl format ) to the SAM format using the psl2sam . pl script within the Blat distribution , the SAM file was converted to a sorted BAM file using Samtools [63] . Sequence coverage was determined in each 500 bp window using coverageBed , distributed with BEDtools [64] . Output was uploaded in R to generate sequence coverage plots of the expanded locus encompassing TGGT1_053770 plus 20 Kb of upstream and downstream sequence . The average read coverage per 500 bp window in the 20 Kb sequence upstream of the expanded locus was used to normalize the coverage plots for all three strains . For generation of MAF1 expression in MEFs , the coding region for the full-length MAF1 was PCR-amplified from pMAF1 using the forward primer TTC TCG AGC ACC ATG GCC GGT TAC CCG TAC GAC G and reverse primer TTG CGG CCG CTC AGT CCA GCA TGC TAG CCA G and cloned into the XhoI and NotI sites of the MSCV2 . 2 vector ( gift from the Barton Lab , University of California , Berkeley ) downstream of the CMV promoter . The forward primers included the Kozak consensus sequence ( CACC ) before the start codon . The MAF1 transgene starts at the signal peptide cleavage site ( amino acid 24 ) , has an N-terminal HA tag after the signal peptide cleavage site , and runs through to the stop codon . Phoenix cells were transfected with 15 µg of pMSCV_MAF1 using Lipofectamine LTX ( Invitrogen ) . About 24 hpi transfection , the supernatant was removed and replaced with fresh cDMEM and the cells transferred to 32°C . The following day the supernatant was filtered with a 22–45 µm filter , polybrene at a concentration of 5 µg/ml was added , and the mix transferred to MEFs at a confluency of 50% . MEFs were incubated for one day at 32°C and then returned to 37°C for expansion . Cells were analyzed for MAF1 expression via immunofluorescence using anti-HA or polyclonal anti-MAF1 antibodies . For mouse arrays , WT and MAVS−/− MEFs were grown in a six-well plate to confluency . Parasite strains were syringe-lysed and washed once with PBS . MEFs were infected with type I , type I:Δmaf1 , type II expressing luciferase ( Me49:LUCΔhxgprt ) , or type II:MAF1 parasites at an MOI of 5 or mock-infected for 8 h after which RNA was isolated using TRIzol ( Invitrogen ) followed by labeling and hybridization to a mouse Affymetrix array ( Mouse 430A 2 . 0 ) according to the manufacturer's protocol . Probe intensities were measured with the Affymetrix GeneChip Scanner 3000 7G and were processed into image analysis ( . CEL ) files with GeneChip Operating Software ( Affymetrix ) . Data were normalized using the Robust Multi-Array Average normalization algorithm . GSEA was used to identify canonical pathways and candidate transcription factors modulated by MAF1 [65] , [66] . Statistical significance for differential gene expression was determined using SAM [67] . The delta parameter was adjusted to achieve a false discovery rate ( FDR ) nearest to 10% , or as indicated , and this delta value was used to select significantly regulated genes . All microarray data have been deposited in the Gene Expression Omnibus ( http://www . ncbi . nlm . nih . gov/geo/ ) . Primary mBMDMs were isolated from female B6 mice and plated to confluency in 24 wells . BMDMs ( three wells per condition ) were infected with type II ( Me49:LUCΔhxgprt ) or type II:MAF1 parasites at an MOI of 3 or mock-infected in equal volumes . At 3 hpi and 17 hpi , supernatants from the three wells for each condition were pooled , pushed through a 5 µm filter , and stored at −80°C until analysis . Supernatants were analyzed in the Human Immune Monitoring Core ( Stanford , CA ) by Luminex Mouse 26-plex kits purchased from Affymetrix and used according to the manufacturer's recommendations with modifications as described below . Briefly , samples were mixed with antibody-linked polystyrene beads on 96-well filter-bottom plates and incubated at RT for 2 h followed by overnight incubation at 4°C . Plates were vacuum-filtered and washed twice with wash buffer , then incubated with biotinylated detection antibody for 2 h at RT . Samples were then filtered and washed twice as above and resuspended in streptavidin-PE . After incubation for 40 min at RT , two additional vacuum washes were performed , and the samples resuspended in Reading Buffer . Each sample was measured in duplicate . Plates were read using a Luminex 200 instrument with a lower bound of 100 beads per sample per cytokine . Data shown are averaged from three independent experiments and are from three biological replicates ( different days ) ± s . e . m . Asterisks denote significantly different MFIs between mBMDM cells infected with type II versus type II:MAF1 , * p<0 . 05 , *** p<0 . 001 , **** p<0 . 0001 using a two-factor ANOVA analysis . Euthanized C57BL/6 mice ( n = 4 per parasite strain ) were injected with 6 ml cold 1× PBS using a 26 gauge needle . Mice were palpated for 1–2 min , following which fluid contents were aspirated out of the peritoneum . Content was spun down at 1 , 000 rpm for 5 min , washed , and resuspended in DMEM to a concentration of 2×106 per ml . We incubated 500 µl of this material for 6 or 12 h , passed it through a 0 . 2 µm filter , and the filtrate was stored at −80°C until Luminex analysis . Samples were analyzed for 26 cytokines using the Luminex platform ( see above ) . Data were reported if p<0 . 05 for an unpaired comparison of the averages . Terminal bleeds were performed 5 dpi on type I and type I:Δmaf1–infected female C57BL/6 ( n = 9 per parasite strain ) and kept at 4°C for 4 h . Bleeds were then spun down at 14 , 000× rpm , and sera were isolated , aliquoted into fresh tubes , and stored at −80°C until analysis . PECs were washed twice with FACS buffer ( 3% FBS in 1×PBS ) and stained for 30 min with surface marker antibody CD11b-APC-Cy7 ( BD Biosciences ) . After staining , the cells were washed with FACS buffer and run on an LSR II flow cytometer ( Becton Dickinson ) . Cells were sorted on a DIVA cell sorter ( BD Biosciences ) and analyzed using FlowJo software ( Tree Star ) . Cells were measured as a percentage of total intact cells ( determined by forward and side scatter measurements ) . Infected cells were identified by virtue of the red fluorescence of the RFP protein expressed by the engineered Toxoplasma .
|
Recent discoveries have revealed the remarkable functional diversity of mitochondria in roles other than energy production , including an integral role for mitochondria and their dynamics in the regulation of the innate immune response . Interestingly , host mitochondria are recruited to the membranes that surround certain intracellular bacteria and parasites during infection . To date , how and why this phenomenon occurs has been a mystery , although it has been proposed to provide a metabolic benefit to the microbes . Here we identify mitochondrial association factor 1 ( MAF1 ) as the parasite protein that mediates the association between the protozoan pathogen Toxoplasma and host mitochondria during infection . We show that MAF1 is needed to recruit host mitochondria to the Toxoplasma-containing vacuole and that this process is associated with changes in the immune response in infected cells and animals . These findings show that recruitment and association with host mitochondria is an important means by which intracellular pathogens interface with their host . We also find that the cost–benefit outcome of altering mitochondrial function might differ between strains depending on the precise niche in which they evolved; for infectious agents , these differences likely reflect different host organisms .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"model",
"organisms",
"cell",
"biology",
"evolutionary",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"immunology",
"microbiology",
"molecular",
"cell",
"biology",
"research",
"and",
"analysis",
"methods"
] |
2014
|
Toxoplasma Effector MAF1 Mediates Recruitment of Host Mitochondria and Impacts the Host Response
|
Chagas disease is a neglected tropical disease . About 6 to 8 million people are chronically infected and 10% to 15% develop irreversible gastrointestinal disorders , including megaesophagus . Treatment focuses on improving symptoms , and isosorbide and nifedipine may be used for this purpose . We conducted a systematic review to evaluate the effectiveness of pharmacological treatment for Chagas’ megaesophagus . We searched MEDLINE , Embase and LILACS databases up to January 2018 . We included both observational studies and RCTs evaluating the effects of isosorbide or nifedipine in adult patients with Chagas’ megaesophagus . Two reviewers screened titles and abstracts , selected eligible studies and extracted data . We assessed the risk of bias using NIH ‘Quality Assessment Tool for Before-After ( Pre-Post ) Studies with No Control Group’ and RoB 2 . 0 tool . Overall quality of evidence was assessed using GRADE . We included eight studies ( four crossover RCTs , four before-after studies ) . Three studies evaluated the effect of isosorbide on lower esophageal sphincter pressure ( LESP ) , showing a significant reduction ( mean difference −10 . 52mmHg , 95%CI −13 . 57 to−7 . 47 , very low quality of evidence ) . Three studies reported the effect of isosorbide on esophageal emptying , showing a decrease in esophageal retention rates ( mean difference −22 . 16% , 95%CI −29 . 94 to −14 . 38 , low quality of evidence ) . In one study , patients on isosorbide reported improvement in the frequency and severity of dysphagia ( moderate quality of evidence ) . Studies evaluating nifedipine observed a decrease in LESP but no effect on esophageal emptying ( very low and low quality of evidence , respectively ) . Isosorbide had a higher incidence of headache as a side effect than nifedipine . Although limited , available evidence shows that both isosorbide and nifedipine are effective in reducing esophageal symptoms . Isosorbide appears to be more effective , and its use is supported by a larger number of studies; nifedipine , however , appears to have a better tolerability profile . PROSPERO CRD42017055143 . ClinicalTrials . gov CRD42017055143 .
Chagas disease , also known as American trypanosomiasis , is an infectious zoonosis caused by the protozoan parasite Trypanosoma cruzi . Once confined to Latin America , it has now spread to other continents . It is estimated that about 6 to 8 million people are infected worldwide . Infection is lifelong and can be life threatening , killing more than 10 000 people every year . Despite its relevance , it is considered by the World Health Organization a neglected tropical disease [1 , 2] . Chagas disease is transmitted by triatomine vectors ( popularly known as kissing bugs ) , vertically ( from mother to fetus ) , orally ( by ingestion ) , by blood transfusion , by organ transplants , and by laboratory accidents . During the first weeks or few months , the disease presents in its acute form , which has no or only mild symptoms , such as fever , fatigue , and headache; thus , infection often goes unnoticed . The host’s immune system then controls parasite replication , and patients enter the chronic phase . Most patients remain asymptomatic , but approximately 30% of infected people develop medical complications from Chagas disease over the course of their lives , usually several years or even decades after the initial infection . The disease mainly affects the heart , digestive system , and nervous system [3 , 4] . Digestive disorders are the second most common manifestation of Chagas disease , affecting about 10% to 15% of patients [5] . There are geographical variations in the prevalence and severity of this condition , which may be associate with the distribution of different T . cruzi strains . They are more frequent in central and southern South America ( including Argentina , Bolivia , Brazil , Chile , Paraguay and Uruguay ) and very rare in other regions [6–8] . Gastrointestinal symptoms are the result of an irreversible enteric nervous system impairment caused by the parasite T . cruzi . Any organ of the digestive system can be affected , but the esophagus and colon are most often injured , causing megaesophagus and megacolon [6] . The esophagus is usually the first affected organ and megaesophagus is the most prevalent manifestation of gastrointestinal Chagas disease [6 , 9] . The most common symptoms of megaesophagus are dysphagia , odynophagia , and esophageal regurgitation . Even though these symptoms do not usually lead to death , they are associated with an increased risk of cancer and may impact quality of life [4 , 6 , 10] . The diagnosis of Chagas’ megaesophagus is based mainly on clinical history , symptoms , barium esophagogram and manometry . Further details about clinical aspects of Chagas’ megaesophagus are presented in Table 1 . The changes in esophageal function caused by T . cruzi are similar , but not equal , to those caused by idiopathic ( primary ) achalasia . Although differences have been described , both conditions are treated similarly [15 , 16] . Treatment always includes dietary adjustments , but they may not be sufficient for some patients . In these cases , pharmacological , endoscopic and surgical interventions can be recommended with the main purpose of decreasing lower esophageal sphincter pressure ( LESP ) , improving esophageal emptying , and relieving symptoms of dysphagia [17 , 18] ( Table 1 ) . Regarding pharmacological treatment , the most commonly used medications are isosorbide dinitrate and nifedipine [6] . Isosorbide dinitrate ( 2 . 5-5mg sublingually , 15 minutes before meals ) releases nitric oxide , which activates the enzyme guanylate cyclase , leading to smooth muscle relaxation [13 , 19] . Nifedipine ( 10mg sublingually , 30 minutes before meals ) is a calcium channel blocker that prevents calcium-dependent myocyte contraction , also leading to muscle relaxation [13 , 19] . These drugs appear to be effective in relieving symptoms , but their use is controversial because of the high incidence of side effects and no change in the course of the disease [6 , 18] . Only a few small studies have evaluated the use of these medications in patients with Chagas disease , and there is no systematic review on this topic . Therefore , the objective of this systematic review was to evaluate the effectiveness of isosorbide and nifedipine versus no treatment for esophageal manifestations of Chagas disease in adult patients and to determine the frequency of side effects .
This systematic review is reported according to the PRISMA Statement [20] ( S1 Appendix ) and was conducted following the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions [21] . The study protocol was registered with the International Prospective Register of Systematic Reviews ( PROSPERO ) , under the registration number CRD42017055143 ( S2 Appendix ) . We searched MEDLINE ( via PubMed ) , Embase and LILACS databases to retrieve potentially relevant articles from inception to January 2018 . We also screened the reference lists of identified publications for additional studies and contacted authors for further information as needed . We conducted two independent searches , one for each intervention ( isosorbide and nifedipine ) . Search terms included “Chagas disease” , “Trypanosoma cruzi” , “isosorbide” , and “nifedipine” . Keywords related to outcomes of interest and publication type were not included to enhance the sensitivity of the search . No language or publication date restrictions were imposed . Search terms were tailored to each database , and the complete search strategies are shown in S3 Appendix . We included studies that met the following criteria: ( A ) observational studies or clinical trials; ( B ) studies that assessed the effects of isosorbide or nifedipine on esophageal symptoms or esophageal function in patients with Chagas disease; and ( C ) individuals aged >18 years with digestive or cardiodigestive form of Chagas disease , according to original studies definitions . We excluded reviews , letters , and editorials . The outcomes of interest were ( A ) esophageal symptoms ( e . g . dysphagia and regurgitation ) , ( B ) esophageal function ( e . g . LESP and esophageal emptying ) , and ( C ) adverse events . We did not include other gastrointestinal symptoms caused by Chagas disease , because isosorbide and nifedipine are used only for esophageal symptoms . In order to screen and select eligible studies , we combined the results from both searches . All identified citations were entered into a software for reference management , and duplicates were excluded . Two independent reviewers ( CBM and CS ) screened the titles and abstracts of all potentially relevant articles identified by the searches , and studies not meeting the eligibility criteria were excluded . The same reviewers assessed the full-text articles of selected abstracts for inclusion according to the pre-specified eligibility criteria . If the study was reported in duplicate , the study published earlier or the one that provided more information was included . Independently , the same reviewers extracted data from the full text of included studies using a pre-designed data extraction form . Data extracted included study characteristics and outcomes of interest . When needed , data were extracted from figures or graphs using WebPlotDigitizer [22] . Disagreements regarding study eligibility or data extraction were discussed between the two reviewers . If consensus was not reached , a third reviewer ( VC ) arbitrated . Two reviewers ( CBM and CS ) independently assessed the methodological quality of included studies . We used the NIH ‘Quality Assessment Tool for Before-After ( Pre-Post ) Studies with No Control Group’ [23] for all observational before-after studies and the RoB 2 . 0 tool [24] for all crossover clinical trials . Disagreements regarding the methodological quality of the studies were discussed between the two reviewers . If consensus was not reached , a third reviewer ( VC ) arbitrated . The overall quality of evidence was assessed using GRADE [25] . Where possible , data were pooled using a meta-analytic approach . A random-effects model , with DerSimonian and Laird’s variance estimator , was used , and the results were presented as mean difference or pooled prevalence , with 95% confidence intervals ( 95%CI ) . A P value ≤ 0 . 05 was considered statistically significant . Statistical heterogeneity among studies was assessed using Cochran’s Q test and the I2 statistic . All meta-analyses were performed using the R statistical software version 3 . 3 . 3 , with meta package version 4 . 8–1 [26 , 27] . When a study did not report the standard deviation ( SD ) , one of the following three strategies was used: estimation of individual patient data from the study’s graphs and calculation of mean and SD; calculation of SD from P value; or input of the highest SD found in other studies for the same outcome . Studies not included in the meta-analysis were presented descriptively .
A total of 66 studies were retrieved for ‘isosorbide’ and 40 for ‘nifedipine’ . Of these 106 studies , eight met the eligibility criteria and were included in our review . Fig 1 shows the flow diagram of study selection , and Table 2 shows the main characteristics of included studies . The quality of all before-after studies was rated as fair [28–31] . All of them presented issues concerning lack of information about eligibility criteria , sample size calculation , blinding , and loss to follow-up . However , all clearly stated the objectives , interventions , and outcomes , statistically analyzed the results and presented P values for the analysis . Among crossover trials , only one study was rated as having a low risk of bias , with no major concerns [33] . The other three studies were rated as having a high risk of bias due to concerns related to randomization , allocation concealment , and blinding of patients and assessors [32 , 34 , 35] . Risk of bias assessment of included studies is summarized in S4 Appendix . Publication bias was not assessed due to the small number of studies .
This review evaluated the effect of isosorbide and nifedipine on esophageal symptoms in patients with Chagas disease . Studies investigating the effect of isosorbide showed that this medication decreased LESP and esophageal retention; moreover , it improved the severity and frequency of dysphagia as reported by patients . Among the studies , mean baseline LESP was 17 . 4mmHg . The reduction of 10 . 52mmHg correspond to 1 . 33 standard deviation , which is considered a large magnitude of effect , indicating the intervention is highly effective in reducing LESP [36] . However , about 30% of patients had headache , and up to 10% reported faintness and palpitation . Studies evaluating nifedipine showed a reduction in LESP but no effect on esophageal retention . The most common side effect of nifedipine was headache , reported by about 10% of patients . Even though digestive symptoms are a well-known manifestation of Chagas disease , recommendations for their treatment are often neglected . In a review of published guidelines for the management of Chagas disease , 10 documents were found , but only two of them provided recommendations for the pharmacological treatment of megaesophagus [37] . Both isosorbide and nifedipine led to improvement in esophageal symptoms . However , only one study , involving 11 patients , directly compared the two drugs of interest [34] . By indirectly comparing the effects of these medications on esophageal function , based on the results of our meta-analysis , isosorbide was superior compared to nifedipine , showing faster onset and longer duration of effects . Moreover , the body of evidence was more consistent for isosorbide than for nifedipine , with seven studies ( 146 patients ) on isosorbide and only two studies ( 26 patients ) on nifedipine . Although current evidence potentially indicates greater certainty that isosorbide is more effective , it is important to consider potential side effects when choosing treatment . In our systematic review , from 30 to 50% of patients on isosorbide reported side effects such as headache , faintness , and palpitation , which may impact long-term adherence to the medication . The rate of side effects of nifedipine was lower , a result similar to that reported in studies of patients without Chagas disease [38 , 39] . Although nifedipine is usually better tolerated , it should be avoided in patients with severe cardiomyopathy , a common manifestation of Chagas disease , due to the risk of hypotension and hydrosaline retention . To our knowledge , this is the first systematic review to evaluate the effects of isosorbide and nifedipine in patients with Chagas disease . We performed a comprehensive literature search without language or date restrictions and systematically evaluated the risk of bias and quality of evidence for the proposed interventions . Included studies applied diagnostic and staging methods currently in use , enhancing the external validity of our findings . Additionally , we performed a meta-analysis of several outcomes of interest , thereby increasing statistical power and precision and making it possible to assess the consistency of the findings . Our systematic review has some limitations , mostly due to the characteristics of included studies . First , the number of included studies is small , a consequence of the sparse number of publications in the field . Besides , all studies were conducted in Brazil , most of them in the same city . This may limit generalizability of the findings , which may be an issue of concern especially nowadays when Chagas disease has spread globally . Moreover , the studies were conducted at least 23 years ago , and since then the healthcare provided to patients has probably changed . Furthermore , all included studies were very small and had methodological limitations . Regarding outcome evaluation , most studies evaluated only surrogate outcomes for symptom improvement , and these measurements may not directly reflect clinical improvement . Besides , no study assessed the long-term effects of these medications; thus , our conclusions provide direct evidence only to short-term outcomes , and the long-term effects for these interventions are still unknown . It is important to note that our systematic review evaluated only studies of patients with Chagas disease; for the assessment of side effects , we did not include information from further studies conducted in other relevant fields , such as cardiology . Although including the results of patients without Chagas disease would increase the heterogeneity and indirectness of our findings , it could give more precision to the prevalence estimates of some side effects that may be similar in patients with and without Chagas disease . Based on the findings of our systematic review , while also acknowledging the lack of rigorous studies such as long-term clinical trials , isosorbide and nifedipine are effective in the treatment of esophageal manifestations of Chagas disease . Isosorbide appears to be more effective , and its use is supported by a larger number of studies . However , nifedipine appears to have a better tolerability profile . Both drugs seem to be valid alternatives , and the decisions about pharmacological treatment should be tailored to each patient .
|
Chagas disease is a chronic neglected tropical disease that has increased in prevalence in the last decade . About 10% of chronically infected patients develop the digestive form of the disease . Megaesophagus is a common manifestation , and symptoms include difficulty or discomfort in swallowing and regurgitation . Treatment approaches include dietary interventions , medications , and endoscopic and surgical interventions . Regarding pharmacological treatment , only a few small studies have evaluated the effects of isosorbide and nifedipine , mainly on surrogate outcomes . According to our systematic review , the use of isosorbide reduced lower esophageal sphincter pressure , esophageal retention after meal ingestion , and the frequency and severity of dysphagia , while nifedipine reduced esophageal retention after meal ingestion . In light of this , both medications are effective in the treatment of symptoms associated with Chagas disease . Isosorbide appears to be more effective , and its use is supported by a larger number of studies; nifedipine , however , appears to have a better tolerability profile . Both drugs are valid alternatives , and the decisions about pharmacological treatment should be tailored to each patient .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"statistics",
"laryngology",
"metaanalysis",
"tropical",
"diseases",
"parasitic",
"diseases",
"headaches",
"dysphagia",
"physiological",
"processes",
"mathematics",
"signs",
"and",
"symptoms",
"pharmaceutics",
"neglected",
"tropical",
"diseases",
"pharmacology",
"research",
"and",
"analysis",
"methods",
"mathematical",
"and",
"statistical",
"techniques",
"adverse",
"reactions",
"protozoan",
"infections",
"otorhinolaryngology",
"ingestion",
"research",
"assessment",
"chagas",
"disease",
"diagnostic",
"medicine",
"systematic",
"reviews",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"drug",
"therapy",
"statistical",
"methods"
] |
2018
|
Isosorbide and nifedipine for Chagas' megaesophagus: A systematic review and meta-analysis
|
Brucellosis is endemic in the bovine population in India and causes a loss of US$ 3·4 billion to the livestock industry besides having a significant human health impact . We developed a stochastic simulation model to estimate the impact of three alternative vaccination strategies on the prevalence of Brucella infection in the bovine populations in India for the next two decades: ( a ) annual mass vaccination only for the replacement calves and ( b ) vaccination of both the adult and young population at the beginning of the program followed by an annual vaccination of the replacement calves and , ( c ) annual mass vaccination of replacements for a decade followed by a decade of a test and slaughter strategy . For all interventions , our results indicate that the prevalence of Brucella infection will drop below 2% in cattle and , below 3% in buffalo after 20 years of the implementation of a disease control program . For cattle , the Net Present Value ( NPV ) was found to be US $ 4·16 billion for intervention ( a ) , US $ 8·31 billion for intervention ( b ) and , US $ 4·26 for intervention ( c ) . For buffalo , the corresponding NPVs were US $ 8·77 billion , US $ 13·42 and , US $ 7·66 , respectively . The benefit cost ratio ( BCR ) for the first , second and the third intervention for cattle were 7·98 , 10·62 and , 3·16 , respectively . Corresponding BCR estimates for buffalo were 17·81 , 21·27 and , 3·79 , respectively . These results suggest that all interventions will be cost-effective with the intervention ( b ) , i . e . the vaccination of replacements with mass vaccination at the beginning of the program , being the most cost-effective choice . Further , sensitivity analysis revealed that all interventions will be cost-effective even at the 50% of the current prevalence estimates . The results advocate for the implementation of a disease control program for brucellosis in India .
Brucellosis is an important zoonotic disease causing infertility , repeat breeding , retention of placenta and abortion in cattle . Humans in contact with animals usually get infected by coming in direct or indirect contact with reproductive secretions and excretions from infected animals . The disease is quite painful among humans and causes undulant fever , chills , fatigue , joint and muscle pain . If not treated , the disease can last for months and years and can cause orchitis , epididymis and endocarditis . Successful implementation of disease control programs have resulted in the eradication of brucellosis from domestic livestock in most of the developed countries [1] . However , the disease is still prevalent and classified as a neglected zoonosis in many parts of the developing world [2] . The disease is endemic in most of the production animals in India [3 , 4] . With the reported disease seroprevalence of 9 . 3% in cattle [5] and 16·4% in buffalo populations [6] , brucellosis is a serious economic concern for the cattle and buffalo industry [7] . Recent studies in India demonstrate that brucellosis in livestock populations results in a median loss of US$ 3·43 billion , with more than 95% of the losses occurring in the cattle and buffalo industry [7] . Brucellosis can be successfully controlled using appropriate intervention policies . Lack of resources to compensate farmers and a ban on cow slaughter in most parts of the country means that test and slaughter policy to control brucellosis cannot be implemented in India . There is no treatment for the disease in animals . Therefore , vaccination of cattle and buffalo population remains the sole alternative for the prevention and control of brucellosis in livestock populations in the country . However , information of benefits and costs of implementing intervention strategies to control the disease in India are largely unknown . This study aims to assess the costs and benefits of alternative control strategies for brucellosis in India . Initially , a stochastic simulation model was developed to project the course of Brucella infection for the national cattle and buffalo herd , over the next twenty years , under two different vaccination schemes . Subsequently , we performed a cost-benefit analysis to quantify the expected benefits of the proposed alternatives . We anticipate that this study would help policy makers to adopt the best available long-term intervention policy to prevent and control the occurrence of brucellosis in livestock and human populations in India .
Firstly , we developed a stochastic simulation model to estimate the impact of alternative vaccination strategies on the prevalence of Brucella infection for the cattle and buffalo populations in India , for the next two decades . The considered alternatives were based on published literature [8–10] . For the first intervention , we assumed a planned annual livestock mass vaccination campaign using Brucella abortus S19 for the female bovine ( cattle and buffalo ) replacement calves . For the second intervention , we assumed that all the adult and young female bovine populations will be vaccinated at the beginning of the program , followed by an annual vaccination of only replacements . The third intervention considered the annual mass vaccination of replacements for a decade followed by a decade of a test and slaughter strategy . We quantified the expected benefits and gains of the proposed control programs and performed a benefit-cost analysis to calculate the overall net expected benefit for each intervention . To estimate the expected benefits from the alternative vaccination strategies , the following dynamic , synchronous , discrete time event stochastic simulation model was setup . First , animals were generated within herds . The time step ( t ) for this model was one year . At each time step , ( a ) the life stage ( i . e . age ) of each animal was determined by a dynamic component that is based on data about the age distribution and the age-specific replacement rate for the cattle and the buffalo populations; and ( b ) the infection stage for each animal was based on the expected prevalence of brucellosis for each species . Prevalence ( P ) was simulated at the herd level and animals within the same herd were assumed to attain the same risk of getting infected . For each herd , P was simulated for the first year and for each of next years it was based on the mean prevalence estimate of the previous year Pt-1 . Animals that got infected were assumed to remain infected for life . Replacements and animals that were not infected were assumed to attain a yearly risk ( YRt ) of getting infected that depended on Pt-1 and the expected mean duration ( D ) of the disease in the infected animals: YRt=1−e−Pt−1 ( 1−Pt−1 ) D The model was allowed to run for a “burn-in” phase of 50 years and then each of the alternative interventions was considered: ( a ) annual mass vaccination only for the replacement calves and ( b ) vaccination of both the adult and young population at the beginning of the program followed by an annual vaccination of the replacement calves and ( c ) annual mass vaccination of replacements for a decade followed by a decade of a test and slaughter strategy . Vaccination was assumed to provide complete protection from infection although we allowed for a small rate of vaccination failures and we also considered different realistic vaccination coverage rates that in real life affect vaccine efficacy . At each time step , a number of parameters was recorded among which prevalence ( P ) , replacement rate , the number of vaccinations and for the third scenario ( i . e . vaccination of replacements followed by a test and cull strategy ) the number of tested animals and the number of culled animals . Estimates were based on the summaries of 4000 simulations of all animals and herds that were run for 40 times , for each species and intervention . A detailed description of the data sources and the input parameters of the model follows . Input parameters and the corresponding distributions are also summarized in Table 1 [5–7 , 11–22] . For each intervention we evaluated the effect of varying input parameters on the expected benefits . Specifically , we assessed the impact of ( a ) reducing the initial prevalence of Brucella infection by fifty percent , ( b ) reducing the vaccination coverage to 50% and ( c ) having herds that are consistently unvaccinated ( i . e . herds that were more likely to remain unvaccinated the next year ) . We also assessed the additional benefits of expanding the intervention strategies beyond the twenty year period . The interventions were considered for t = 20 years and at a discount rate ( r ) of 5% . Initially , we predicted the annual costs ( Ct ) and benefits ( Bt ) for each strategy and subsequently calculated the net present value ( NPV ) by applying the discount rate: NPV=∑t=1TBt−Ct ( 1+r ) t Further , for each intervention the benefit cost ratio ( BCR ) was estimated as the discounted value of the incremental benefits divided by the discounted value of the incremental costs: BCR=∑t=1TBt ( 1+r ) t∑t=1TCt ( 1+r ) t The costs included vaccine costs , service costs of vaccination ( transportation , cold chain , and veterinarian fee ) , animal identification costs ( ear tagging ) , service costs for surveillance and diagnostics , and costs for health education program ( Table 1 ) . The averted losses were considered as benefits for implementing the control programs [27] . Based on our previous study [7] , the losses occurring due to brucellosis per infected animal were estimated by dividing total losses for each species with the number of infected animals for that species ( Table 1 ) . Due to lack of data , the health and economic burden of human brucellosis could not be accounted into the overall benefits of the control programs . The analyses were conducted using R-statistical program ( R statistical package version 2 . 12 . 0 , R Development Core Team , http://www . r-project . org ) and we run Monte Carlo simulations for 10 , 000 iterations so as to determine confidence limits for these estimates .
For each intervention , our results indicate that the prevalence of Brucella infection will drop below 2% in cattle after 20 years of the implementation of disease control program ( Fig 1 ) for the cattle population . For buffaloes , a similar trend was observed . However , due to the higher initial prevalence of infection , it only drops below 3% after the twenty year implementation of all interventions ( Fig 2 ) . The NPV during the first 20 years of the program for cattle for scenario 1 , 2 and , 3 are presented in Table 2 . For cattle , the NPV was found to be US $ 4·16 billion ( 95% CI: US $ 3·16; 5·39 billion ) for the scenario 1 , US $ 8·31 ( 6·40; 9·87 ) billion for the scenario 2 and , US $ 4·26 ( 3·26; 5·61 ) for the third scenario ( Table 2 ) . The results indicate that first 20 years of the programme will be cost-effective for all scenarios with the second intervention ( vaccination of replacements with mass vaccination at the beginning of the program ) being a significantly more cost-effective choice . The BCR for the first , second and the third intervention for cattle were 7·98 ( 6·29; 10·09 ) , 10·62 ( 8·33; 12·5 ) and , 3·16 ( 2·66; 3·83 ) , respectively . Similar results were obtained for buffaloes ( Table 3 ) . The NPV for the 50% prevalence estimates during the first 20 years of the program for cattle for scenario 1 , 2 and , 3 are presented in S1 Table . For cattle , NPV was found to be US $ 1·78 billion ( 95% CI: US $ 1·07; 2·79 billion ) for the scenario 1 , US $ 3·27 ( 2·18; 4·23 ) billion for scenario 2 and , US $ 0·87 ( -0·24; 1·71 ) for scenario 3 ( S1 Table ) . For buffaloes , the NPV for the 50% prevalence estimates was found to be US $ 3·69 billion ( 95% CI: US $ 2·74; 4·54 billion ) for the scenario 1 , US $ 5·96 ( 4·40; 7·27 ) billion for the scenario 2 and , US $ 2·09 ( 1·02; 3·20 ) for the third scenario ( S2 Table ) . Further , sensitivity analysis revealed that for either species , disease prevalence will further reduce to less than 1% after 50 years of implementation for either intervention and will virtually lead to eradication of the disease after 100 years of the implementation programme . The long time to eradicate infection is based on the fact that we only considered realistic vaccination coverage rates . Our primary analysis , assumed a vaccination coverage of 70% . Reduction of the vaccination rate led to reduced NPV and BCR values . The same impact had the assumption that herds that were not covered were more likely to remain uncovered the next year .
This is the first systematic analysis of a brucellosis control program interventions for bovine brucellosis in India . Bovine brucellosis is highly prevalent in India and causes significant losses to the livestock industries . The results suggest that all of the three approaches investigated for controlling the disease would be beneficial as the prevalence of Brucella infection will drop below 2% in cattle and 3% in buffalo after 20 years of the implementation of disease control program . All programs had positive NPVs and >1 BCRs indicating the benefits from all programs are higher than their respective costs . The best BCR was obtained in the second intervention , i . e . vaccination of both the adult and young population at the beginning of the program followed by an annual vaccination of the replacement calves . It leads to a significant drop in prevalence at the beginning of the program and hence the risk of transmitting the disease in the subsequent years is lower . This is thus the most cost-effective approach for control of brucellosis in India . Overall , the results advocate the implementation of a disease control program for brucellosis in India . The results of sensitivity analyses indicated that a positive effect for all interventions and a net benefit of billions of dollars for any intervention remains even after considering significantly reduced initial prevalence and vaccination coverage . This suggests that the control program would be beneficial even if some of the assumptions used in the model are changed , further supporting the implementation of a control program for the disease . It must be noted that we only considered economic benefits of the control programs for the livestock populations . The benefits such as disability-adjusted life years ( DALYs ) and social losses averted due to the control programs could not be accounted . Similarly , the extra costs due to increased livestock numbers ( feed costs ) , or unintentional consequences ( abortion due to vaccinating a female cow ) were not estimated . However , we believe that this will not have a major impact on the results of the current study . Although the third approach–i . e . annual mass vaccination of replacements for a decade followed by a decade of a test and slaughter strategy–was also found to be cost-effective , it is less likely to be adopted in India because it is a Hindu majority country and Hindus consider cows to be sacred . As a result , cow slaughter is banned in most states of India . Thus it would be difficult to get community support for a strategy involving animal slaughter although it drastically reduces prevalence of the disease if implemented after a decade of vaccination . It will also be more expensive as it would involve testing of animals which would include sample collection , transport and laboratory testing . Also , there would be additional costs involved for culling infected animals . Therefore , this may not be the preferred strategy in the Indian situation . Moreover , in calculating losses for test and slaughter , we assumed that animals will be consumed after slaughter in accordance with the WHO guidelines [28] . However , it may not be feasible to do so or may increase the risk of spread of infection . Therefore , it would be more sensible to adopt a ‘test and euthanasia’ strategy in which the infected animals are euthanized and their carcasses burnt or buried and not consumed . This strategy is likely to have a greater acceptance among the community which is very essential for the success of any control program . However , this would increase the cost of the test and slaughter program as it will result in a complete loss of slaughtered animal instead of just a loss of 20% considered in the scenario . Thus the actual cost of the third scenario may be higher than we estimated . In this study , BCRs for three inventions for cattle were estimated to range from 3·16 to 10·62 and for buffaloes from 3·79 to 21·27 . Similar estimates have been obtained in some other studies conducted around the world . The strategy of vaccinating 3–6 month old female bovine , male ovine and female ovine followed by compulsory slaughter after attaining the target prevalence have been advocated in Turkey [10] , where BCR was estimated to be 2·26 [10] . A BCR of 3·2 has been estimated for control of brucellosis in Nigeria [29] . A national serological survey and risk based vaccination using S19 and an awareness program was found to have a BCR of 6·8 for control of brucellosis in Nepal [9] . Note that we only considered scenarios for control of the disease; eradication was not considered feasible in the current circumstances . The disease is highly prevalent and endemic in India; therefore , it would be unrealistic to achieve eradication . Further , India is a vast country with significant movement and intermixing of animals . Moreover , eradication would definitely require test and slaughter but religious and cultural beliefs would impede implementation of any such program due to limited community support . However , once the prevalence reduces below 2% after 20 years , there may be greater community support for eradication as well as test and slaughter/euthanasia as discussed before . Therefore , it would be wise to revisit this question sometime in the future . In this work realistic inputs of vaccination coverage aimed to also adjust for the reduced vaccine efficacy due to vaccine failure as well as problems associated with cold chains , which were not directly accounted for . Assumed vaccination rates ranged from 70% to 50% to cover different vaccine efficacies that have been used in previous studies [25 , 30] . Undoubtedly , real-time data of vaccine efficacy could further improve the predictive ability of our model . However , in our sensitivity analysis we considered the realistic fact that herds that were not covered once were more likely to remain uncovered due to issues associated with inability to reach them or farmers’ will to cooperate . However , even better NPV and BCR will be achieved if the vaccination coverage/efficacy is improved . It has been reported that 53 . 6% of the bovine ( cattle and buffalo ) population receives foot and mouth disease ( FMD ) vaccination in India [31 , 32] . Therefore , our assumption of 50% vaccination coverage is quite realistic . However , there will be pockets of low ( <10% ) and high coverage ( 90%+ ) areas . Many factors such as poor infrastructure , lack of knowledge and veterinary personnel availability are responsible for poor adoption of vaccines in India [33] . A farmer’s perceptions such that vaccination could lead to decrease in milk yield , swelling and fever also decrease vaccine coverage [33] . Low community acceptance , vaccine stock outs at the local level and timeliness of vaccine also affect the vaccine coverage [34] . These factors could affect the benefits but could not be accounted in the current study . The advantage of using the S19 vaccine is that immunity induced is long-lasting and has been reported to be effective till fifth pregnancy [10 , 35] . However , there are a number of concerns with using this vaccine . The major concern is the common occurrence of the needle stick injuries and the accidental inoculation in veterinary personnel while participating in Brucella vaccination programs [36] . The rates of accidental exposure ranging from 6·7% to 46% have been reported [37] . The needle stick injuries have been reported to cause a low virulence human brucellosis [38] . To avoid needle stick injuries , research should be conducted in the use of a safety vaccinator as used for other vaccines such as Gudair® vaccination in Australia [39] and animals should be properly restrained before vaccination . Additionally , the S19 vaccine could interfere with the recommended diagnostic tests and may cause abortion in the pregnant animals [35 , 40] . Moreover , the vaccine cannot be used for male [41] or infected animals [42] . Therefore , there is a need for the development of a better vaccine that can differentiate infected from vaccinated animals ( DIVA ) . RB51 vaccine can be used instead of S19 as it allows serological differentiation between naturally infected and vaccinated animals but it is not currently available in India and is considered to have a lower efficacy than S19 . It is worth mentioning here that the cost and benefit analyses evaluated in this manuscript only pertain to the effect of the disease on the domestic animal population . The benefits to the human population would be over and above the benefits discussed here but were beyond the scope of this study . It is well known that humans get infected while handling infected animals . Therefore , various studies have shown that the disease is prevalent among occupational groups such as veterinary personnel , laboratory workers , livestock farmers and abattoir workers in India [43–45] . We have recently shown that the disease causes a loss of 177 601 ( 95% UI 152 695–214 764 ) DALYs at the rate of 0 . 15 ( 95% UI 0 . 13–0 . 17 ) DALYs per thousand persons every year [46] and an annual median loss of Rs 627 . 5 million ( US $ 10 . 46 million ) in India [46] . Complete eradication of the disease will save these losses but further studies are required to investigate the real impact of the control strategies discussed in this manuscript on the human population .
|
Brucellosis is an endemic zoonosis in India and recent studies demonstrate that the disease results in a median loss of US$ 3 . 43 billion in livestock populations . Lack of resources to compensate farmers and a ban on cow slaughter means that test and slaughter policy to control brucellosis cannot be implemented in India . This is the first systematic analysis of a brucellosis control program interventions for bovine brucellosis in India . The cost-benefit analysis was successfully conducted and indicated benefits of implementing the intervention policies . For each intervention , our results indicate that the prevalence of Brucella infection will drop below 2% in cattle after 20 years of the implementation of disease control program although some strategies were better than others . The expected net present value ( NPV ) was found to range from US $ 4·16 to $ 8·31 billion for cattle and from $ 7·66 to $ 13·42 billion in buffalo for the three strategies investigated . The benefit cost ratio ( BCR ) ranged from 3·16 to 10·62 for cattle and from 3·79 to 21·27 for buffalo . The results advocate for the implementation of a disease control program and will help development of an official health policy for the control of brucellosis in India .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"united",
"states",
"livestock",
"medicine",
"and",
"health",
"sciences",
"immunology",
"tropical",
"diseases",
"geographical",
"locations",
"india",
"brucellosis",
"north",
"america",
"preventive",
"medicine",
"bacterial",
"diseases",
"vaccines",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"vaccination",
"and",
"immunization",
"veterinary",
"science",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"veterinary",
"diseases",
"zoonoses",
"agriculture",
"people",
"and",
"places",
"asia",
"biology",
"and",
"life",
"sciences"
] |
2018
|
Cost-benefit analysis of intervention policies for prevention and control of brucellosis in India
|
Noma ( cancrum oris ) , a neglected tropical disease , rapidly disintegrates the hard and soft tissue of the face and leads to severe disfiguration and high mortality . The disease is poorly understood . We aimed to estimate risk factors for diagnosed noma to better guide existing prevention and treatment strategies using a case-control study design . Cases were patients admitted between May 2015 and June 2016 , who were under 15 years of age at reported onset of the disease . Controls were individuals matched to cases by village , age and sex . Caretakers answered the questionnaires . Risk factors for diagnosed noma were estimated by calculating unadjusted and adjusted odds ratios ( ORs ) and respective 95% confidence intervals ( CI ) using conditional logistic regression . We included 74 cases and 222 controls ( both median age 5 ( IQR 3 , 15 ) ) . Five cases ( 6 . 5% ) and 36 ( 16 . 2% ) controls had a vaccination card ( p = 0 . 03 ) . Vaccination coverage for polio and measles was below 7% in both groups . The two main reported water sources were a bore hole in the village ( cases n = 27 , 35 . 1%; controls n = 63 , 28 . 4%; p = 0 . 08 ) , and a well in the compound ( cases n = 24 , 31 . 2%; controls n = 102 , 45 . 9%; p = 0 . 08 ) . The adjusted analysis identified potential risk and protective factors for diagnosed noma which need further exploration . These include the potential risk factor of the child being fed pap every day ( OR 9 . 8; CI 1 . 5 , 62 . 7 ) ; and potential protective factors including the mother being the primary caretaker ( OR 0 . 08; CI 0 . 01 , 0 . 5 ) ; the caretaker being married ( OR 0 . 006; CI 0 . 0006 , 0 . 5 ) and colostrum being given to the baby ( OR 0 . 4; CI 0 . 09 , 2 . 09 ) . This study suggests that social conditions and infant feeding practices are potentially associated with being a diagnosed noma case in northwest Nigeria; these findings warrant further investigation into these factors .
Noma or cancrum oris , is a poorly understood , rapidly progressing gangrenous infection of the oral cavity , associated with a high mortality rate [1] . It mostly affects children under the age of five years [2] . It is estimated that up to 90% of noma cases die [3] , and those who survive have severe facial disfigurements [2] . These can result in multiple physical impairments such as difficulty speaking , swallowing , eating , seeing and breathing which can lead to stigmatization in their communities [4] . Noma is thought to be most prevalent along the noma belt which stretches from Senegal to Ethiopia [4] , however noma cases have recently been reported in the United Kingdom [5] , United States [6] , Afghanistan [7] , South Korea [8] and Laos [9] . A northwest Nigerian based study concluded that the incidence of noma is estimated to be 6 . 4 per 1000 children [10] , and the World Health Organization ( WHO ) estimates that 140 000 children contract noma each year globally [2] . Little is understood about noma as most cases live in underserved , difficult to reach locations , many cases go undiagnosed and the mortality rate is so high . Previous observational studies have suggested that risk factors for development of noma include malnutrition , low birthweight , absence of breastfeeding , poor oral hygiene , co-morbidities , proximity of livestock to area of residence , large family size , access to unsafe drinking water and living in a village with a high prevalence of acute necrotising gingivitis [4 , 11–13] . Recently , an increased incidence of noma has been reported in higher resource settings in patients with immunosuppressive diseases such as human immunodeficiency virus ( HIV ) [5 , 6 , 8 , 14 , 15] . Since 2015 , Médecins Sans Frontières ( MSF ) has collaborated with the Nigerian Ministry of Health ( MoH ) to treat noma patients identified across the northwest of Nigeria , in Sokoto . The programme provides nutritional , psychosocial ( for the patient and their families ) and medical care to prepare noma patients for the required surgical interventions at the Noma Children’s Hospital . These are conducted by MoH and MSF surgical teams on a routine basis; since August 2015 , the programme has treated 227 noma patients . We conducted a case-control study to identify risk factors for diagnosed noma in terms of demographic characteristics , medical history , socio-economic-behavioural aspects and access to health care in order to better guide existing prevention and treatment strategies for this neglected disease .
The study was conducted in Sokoto and Kebbi states , which are located in northwest Nigeria . Cases were defined as patients with diagnosed noma admitted to the Noma Children’s Hospital between May 2015 and June 2016 who were under 15 years of age at self-reported onset of the disease . Controls were individuals matched to cases by village of residence , current age ( +/- 2 years ) and sex . Our aim was to include all cases enrolled at the Noma Children’s Hospital in the year before data collection , and we calculated that with a sample size of 67 cases and 200 controls ( three controls per case ) , we would be able to estimate an odds ratio ( OR ) of 2 . 5 for suspected risk factors with a power of 80% and 60% of controls being exposed to that risk factor . Controls were selected from houses neighbouring those in which the cases and their families live . Seventy-eight percent of cases and controls were younger than 18 years of age , their parents or caretakers were asked to participate in the interviews using a structured questionnaire . The questionnaire covered the sociodemographic characteristics of the cases and controls ( age , gender , education , employment , total household members ) , their current living conditions ( water source , proximity to livestock , material of houses ) and their vaccination history ( read on vaccination card if available ) . Additionally , parents and/or caretakers were asked to respond to questions pertaining to the duration of breastfeeding after the cases and controls were born and other nutrition-related practices during the neonatal period and current practices . The health status , access to health care and healthcare seeking behaviour for the case or control in the previous 12 months were also assessed . Finally , all cases and controls aged less than five years at the time of interview had a mid-upper arm circumference ( MUAC ) measurement taken at the time of the interview . The questionnaire was formatted in Kobo Collect ( http://www . kobotoolbox . org/ ) and uploaded to tablets for mobile data collection purposes . Completed questionnaires were uploaded daily to a secure MSF server through an internet connection . The study coordinator verified all completed questionnaires on a daily basis for data consistency and quality . We calculated the frequencies and respective proportions for all categorical variables and used chi-square tests for comparison of these variables between cases and controls . For continuous variables , we calculated means with standard deviation or medians and interquartile ranges ( depending if approximately Normally distributed ) for cases and controls separately , and used t-tests to compare Normally distributed variables , and Kruskal Wallis tests for non-Normally distributed variables . Food variables for current feeding practices were categorised as animal products ( meat , milk , egg ) , grains ( fura , mashed rice , millet , corn , bread ) and vegetables ( sweet potato , beans , bean cake , moringa leaf with ground nut cake , cassava ) . Respondents could answer with ‘Yes’ , ‘No’ and ‘Don’t know’ . We grouped all responses for ‘No’ and ‘Don’t know’ into a single category . To investigate the impact of this grouping , we conducted a sensitivity analysis for each of these variables . As the results showed that the direction of association remained the same , we retained this grouping as the reference category . We estimated risk factors for being a diagnosed noma patient by comparing odds of exposure in cases and controls using unadjusted and adjusted conditional logistic regression to calculate Odds Ratios ( ORs ) and their respective 95% confidence intervals ( CI ) and p-values . The adjusted conditional logistic regression model was constructed using all risk factors that had a p-value of <0 . 2 in the unadjusted analysis and sufficient outcomes in each category . Variables were eliminated from the adjusted model using a manual backwards stepwise approach [16] , and adjusted models were compared using the likelihood ratio test , any variable with a p-value under 0 . 2 was kept in the model . All data analyses were conducted with Stata 14 ( StataCorp , College Station , TX , USA ) . The MSF Ethical Review Board approved the study protocol ( study 1710 ) , as did the Usmanu Danfodiyo University Teaching Hospital ( UDUTH ) Health Research and Ethics Committee in Nigeria ( UDUTH/HREC/2017/No . 595 ) and the Ministry of Health in both Sokoto ( SKHREC/032/017 ) and Kebbi ( MOH/SUB/4027/Vol . I/14 ) states . All interviewees were over the age of 18 and written informed consent was provided by each participant ( for participants who were illiterate , the consent form was read aloud to them and a thumb print was then requested ) . All participants were assured that there was limited risk of harm from participation in this study , and that they were free to withdraw at any point .
Out of the 112 noma patients who had sought care in the programme between May 2015 and June 2016 , we identified 87 who lived in Kebbi and Sokoto states and were eligible for inclusion in the study . Of these , 10 could not be located , and we managed to interview 77 cases ( 88 . 5% ) . We were unable to reach the village of three identified cases for logistical reasons . Thus , the final analysis included 74 noma cases and 222 controls . Six of the cases had passed away in the time between discharge and the interview; the interviews were still conducted with their caretakers . At the time of first admission to the hospital , 17 of these cases had acute noma , 57 had inactive noma , two had trismus and one had no diagnosis noted at time of admission to the hospital . Twenty one cases were hospitalized at the time of interview and the remaining 56 were interviewed in their home villages . As expected , there were no significant differences in matching variables ( sex and age of child ) between cases and controls . The control group differed from the case group in that caretakers were younger , their family sizes were smaller ( and houses in the compound fewer ) and most of their houses were made of mud ( which might be a proxy for higher socio-economic status ) ( Table 1 ) . Also , the respondents for the control group were more frequently the mother of the child ( Table 1 ) . The self-reported median age of onset of noma amongst cases was 2 . 0 ( IQR 2 . 0 , 3 . 0 ) . Cases between 6 months and 5 years old had lower mean MUAC measurements ( mean 134; SD 20 ) than controls ( mean 142; SD 11; p = 0 . 002 . Six percent of cases ( n = 5 ) had a vaccination card available at the time of interview , compared with 16% ( n = 36; p = 0 . 03 ) of controls . The vaccination status ( based on the vaccination card was ) for polio ( cases n = 3 , 3 . 9%; controls n = 13 , 5 . 9%; p = 0 . 07 ) and measles ( cases n = 2 , 2 . 6%; controls n = 15 , 6 . 8%; p = 0 . 1 ) was very low for both cases and controls . Parents and caretakers reported that most cases ( n = 63 , 81 . 8% ) and controls ( n = 158 , 71 . 2%; p = 0 . 22 ) had been breastfed between one to two years after birth . With respect to the main source of drinking water , parents reported that the two main water sources were a bore hole in the village ( cases n = 27 , 35 . 1%; controls n = 63 , 28 . 4%; p = 0 . 08 ) and a well in the family compound ( cases n = 24 , 31 . 2%; controls n = 102 , 46 . 0%; p = 0 . 08 ) . In terms of animal ownership , the proportion of case and control families reporting owning donkeys , dogs , sheep , goats and chickens were similar . Case families , however , were less likely to have cows in their compound compared with controls ( cases n = 36 , 46 . 8%; controls n = 128 , 57 . 7%; p = 0 . 09 ) . The unadjusted analysis suggested that the likelihood of being a diagnosed noma case increased when the household was large ( >10 people ) , the child was heavy at birth ( self-reported weight ) , breastfeeding occurred for >1 year , first solid food was given after the age of 12 months , the child ate pap ( type of porridge staple made from maize , sorghum , or millet ) , and grains each day . Protective factors against acquiring noma included: the mother being the primary caretaker , the caretaker being married , having a well in the compound , colostrum being given to the child , child first given water over the age of 7 months , all children in the household being alive and the child having taken medication ( traditional or biomedicine ) in the year preceding the interview ( Table 2 ) . Due to collinearity , the following variables were not included in the adjusted analysis: the food group variable “Grains” ( fura , mashed rice , millet , corn , bread ) , material the walls of the house was made out of , wealth score , measles , polio , one or more vaccination being reported on a vaccination card . In the adjusted analysis , eating pap every day remained strongly associated with being a risk factor for diagnosed noma ( OR 9 . 8; CI 1 . 5 , 62 . 7; p = 0 . 02 ) as did a later age of first solid food ( >12 months ) ( OR 5 . 07; CI 0 . 9 , 26 . 08; p = 0 . 07 ) . Variables that were protective against being a case included the mother being the primary caretaker ( OR 0 . 08; CI 0 . 01 , 0 . 5; p = 0 . 007 ) , the caretaker being married ( OR 0 . 006; CI 0 . 0006 , 0 . 5; p = <0 . 001 ) and colostrum being given to the baby ( OR 0 . 4; CI 0 . 09 , 2 . 09; p = 0 . 07 ) ( Table 2 ) .
In conclusion , our case control study suggests that infant and current feeding behaviours as well as caretaker demographics may affect the risk of developing noma . Malnutrition and low vaccination coverage , high morbidity of infectious diseases along with low access to health care are all likely contributing factors . We recommend that further research is implemented to determine the true burden of noma , and that prospective studies are implemented to better understand the sequence of events contributing to the development of noma . Only with these sets of indicators will it be possible to better formulate and target prevention programmes .
|
Noma or cancrum oris is an orofacial gangrene that rapidly disintegrates the hard and soft tissue of the face . Little is known about noma as most cases live in underserved , difficult to reach locations . There is a dearth of literature on the risk factors for the development of noma . Médecins Sans Frontières ( MSF ) in collaboration with the Nigerian Ministry of Health runs projects at the Noma Children’s Hospital in Sokoto . A case control study was conducted in northwest Nigeria to explore exposures associated with diagnosed noma using unadjusted and adjusted conditional logistic regression models . Potential risk and protective factors for diagnosed noma were identified and these findings need further exploration . The study identified that feeding pap to the child every day was a potential risk factor for diagnosed noma ( possibly a proxy for poor variety in the diet ) . The following potential protective factors for diagnosed noma were identified: the mother being the primary caretaker , the caretaker being married , and colostrum being given to the baby . Noma is a neglected disease , and current risk factors suggest that intervention efforts could be more effective by focussing on access to health care , the benefits of breastfeeding and a varied diet . However , more research is needed in order to better understand the pathogenesis of this disease in order to improve prevention , early detection and treatment .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"death",
"rates",
"neonatology",
"children",
"medicine",
"and",
"health",
"sciences",
"maternal",
"health",
"immunology",
"tropical",
"diseases",
"health",
"care",
"pediatrics",
"preventive",
"medicine",
"bacterial",
"diseases",
"age",
"groups",
"women's",
"health",
"infants",
"neglected",
"tropical",
"diseases",
"population",
"biology",
"vaccination",
"and",
"immunization",
"families",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"health",
"risk",
"analysis",
"noma",
"child",
"health",
"people",
"and",
"places",
"population",
"metrics",
"breast",
"feeding",
"biology",
"and",
"life",
"sciences",
"population",
"groupings",
"mothers"
] |
2018
|
Risk factors for diagnosed noma in northwest Nigeria: A case-control study, 2017
|
The US2-11 region of human and rhesus cytomegalovirus encodes a conserved family of glycoproteins that inhibit MHC-I assembly with viral peptides , thus preventing cytotoxic T cell recognition . Since HCMV lacking US2-11 is no longer able to block assembly and transport of MHC-I , we examined whether this is also observed for RhCMV lacking the corresponding region . Unexpectedly , recombinant RhCMV lacking US2-11 was still able to inhibit MHC-I expression in infected fibroblasts , suggesting the presence of an additional MHC-I evasion mechanism . Progressive deletion analysis of RhCMV-specific genomic regions revealed that MHC-I expression is fully restored upon additional deletion of rh178 . The protein encoded by this RhCMV-specific open reading frame is anchored in the endoplasmic reticulum membrane . In the presence of rh178 , RhCMV prevented MHC-I heavy chain ( HC ) expression , but did not inhibit mRNA transcription or association of HC mRNA with translating ribosomes . Proteasome inhibitors stabilized a HC degradation intermediate in the absence of rh178 , but not in its presence , suggesting that rh178 prevents completion of HC translation . This interference was signal sequence-dependent since replacing the signal peptide with that of CD4 or murine HC rendered human HCs resistant to rh178 . We have identified an inhibitor of antigen presentation encoded by rhesus cytomegalovirus unique in both its lack of homology to any other known protein and in its mechanism of action . By preventing signal sequence-dependent HC translocation , rh178 acts prior to US2 , US3 and US11 which attack MHC-I proteins after protein synthesis is completed . Rh178 is the first viral protein known to interfere at this step of the MHC-I pathway , thus taking advantage of the conserved nature of HC leader peptides , and represents a new mechanism of translational interference .
Human cytomegalovirus ( HCMV ) is a widespread pathogen which is mostly asymptomatic in immune competent individuals , but pathogenic in the immune compromised such as post-transplant or AIDS patients [1] . Following primary infection , HCMV establishes a latent infection for life which is largely controlled by the cellular immune system . Immune control of HCMV requires enormous immunological resources with often more than 10% of the T cell pool being CMV-specific , a number that might further increase with age [2] . However , these immunological efforts are unable to eliminate the virus and do not prevent super-infection [3] . Thus , HCMV is a master in surviving in the face of a constant immunological onslaught . As one of the largest human viruses , with well over 200 open reading frames ( ORFs ) , HCMV uses only about a third of its coding potential for “essential” functions whereas the majority of its genes are non-essential for growth in vitro [4] , [5] . Many of these “non-essential” genes encode modulators of innate or adaptive immune responses including inhibitors of apoptosis , interferon-induction , T cell and NK cell recognition [6]–[9] . However , the importance of these immune modulators for viral pathogenesis and immune escape in vivo is not known since HCMV does not infect immunocompetent experimental animals . Such restricted species specificity is a hallmark of CMVs and , as a result , CMVs have co-evolved with their hosts [10] . Chimpanzee CMV is most closely related to HCMV [11] . However , chimpanzees are a protected species and unsuitable as an animal model . Although more distantly related to humans , rhesus macaques ( RM ) are readily available for experimentation . Sequence analysis of rhesus CMV ( RhCMV ) revealed that approximately 60% of the open reading frames ( ORFs ) are homologous to HCMV ORFs including most of the aforementioned immune modulators [12] , [13] . In order to study the importance of some of the immune regulatory functions in vivo , we have begun to characterize several of the conserved immune modulators of RhCMV . The US2-US11 genomic region of HCMV encodes multiple proteins that interfere with several MHC and MHC-like molecules . Among the best studied of these is the US6-family which contains four genes that inhibit MHC class I ( MHC-I ) -mediated antigen presentation to T cells: US2 , US3 , US6 and US11 [14]–[16] . These proteins are type I transmembrane glycoproteins that reside in the endoplasmic reticulum and show clear homology to each other and structural features resembling the IG-superfamily fold [17] . Despite these structural similarities , each protein interferes in its own unique way with the assembly of MHC-I with peptides at a post-translational level . Upon completion of heavy chain ( HC ) translation and translocation into the lumen of the ER , but prior to assembly with the light chain β2-microglobulin ( β2-m ) , US2 and US11 mediate the retro-translocation of MHC-I molecules to the cytosol [18] . There , the HC is deglycosylated by N-glycanase and degraded by the proteasome [19] . US6 inhibits peptide translocation by the TAP thus preventing the MHC-I heterodimers from obtaining viral peptides [16] . Finally , US3 prevents ER exit of peptide-loaded MHC-I molecules [15] , both by directly interacting with MHC-I molecules and by interfering with tapasin and protein-disulfide isomerase , both chaperones of the peptide loading complex [20] . We previously demonstrated that the US2-11 orthologues of RhCMV are also functionally equivalent in that Rh182 ( US2 ) and Rh189 ( US11 ) mediate proteasomal destruction of MHC-I , Rh183 ( US3 ) retains MHC-I and Rh185 ( US6 ) inhibits TAP [21] . Thus , it seemed likely that eliminating the genomic region spanning RhUS2-11 from RhCMV would restore MHC-I assembly and transport in RhCMV-infected cells as previously observed for US2-11-deleted HCMV [22] . Surprisingly however , we discovered that in addition to these conserved mechanisms , RhCMV contains an additional ORF , rh178 , that targets the MHC-I assembly pathway . Interestingly , this ORF does not display any homology to the US6 gene family and acts by a novel mechanism that operates post-transcriptionally , but prior to completion of translation/translocation .
Deletion of the genomic region encoding US2-US11 restores MHC-I expression in HCMV-infected cells [22] . To determine if deletion of the homologous region in the RhCMV genome would likewise restore MHC-I expression we created a recombinant RhCMV lacking RhUS2-11 using a RhCMV-derived bacterial artificial chromosome ( BAC ) ( Protocol S1 ) [23] . Similar to recombinant HCMV lacking US2-11 , a growth defect was not observed for ΔRhUS2-11 [22] . However , unlike US2-11-deleted HCMV , ΔRhUS2-11 retained some ability to reduce MHC-I steady state levels in infected TRFs as shown by immunoblot ( Fig . 1A ) . At 48 hours post-infection , MHC-I was markedly reduced in ΔRhUS2-11-infected TRFs . To determine whether the reduced steady state levels were due to interference with newly synthesized MHC-I , we immunoprecipitated MHC-I from radiolabeled TRFs infected with wild-type ( WT ) or ΔRhUS2-11 . When cells were labeled for one or two hours , we recovered dramatically less MHC-I from RhCMV-infected cells despite the use of polyclonal antiserum K455 recognizing all forms of MHC-I ( Fig . 1B ) [24] . Compared to WT there was an increase in HC recovery from ΔRhUS2-11-infected cells . Such residual HC was also observed in pulse-chase experiments , when ΔRhUS2-11-infected TRFs were pulsed for 10 min and chased from 30 min up to 90 min ( Fig . S1A ) . However , compared to mock-treated cells , radiolabeled HC was drastically reduced at all time points either during pulse or chase . In contrast to HC , expression of control proteins such as Transferrin-receptor or vimentin was unaffected in RhCMV-infected cells ( Fig . 1D ) . Also , we did not observe a general shut-off of host protein expression or a dramatic decrease of glycoprotein recovered with the lectin concanavalin A ( data not shown ) . Moreover , expression of the light chain β2-m was much less affected by RhCMV compared to HC , particularly in short pulse/chase experiments ( Fig . 1C ) . These data suggested that in addition to RhUS2-11 inhibiting MHC-I assembly , RhCMV specifically interferes with expression of HC . The residual HC recovered from ΔRhUS2-11-infected cells indicate that this viral inhibition of HC expression ( VIHCE ) was either incomplete or VIHCE did not equally affect all MHC-I alleles present in TRFs . Since only minimal amounts of HC are detectable during ΔRhUS2-11 infection , we wanted to examine if VIHCE caused rapid degradation of HCs . In cells infected with HCMV , HC is initially synthesized but then rapidly degraded as shown by pulse-chase ( Fig . 1C ) . This observation is consistent with previous reports and is due to the reverse translocation of MHC-I mediated by US2 and US11 followed by proteasomal destruction of MHC-I [19] . In contrast , during infection with both WT RhCMV and ΔRhUS2-11 only minimal amounts of HC were detectable after a 10-min radiolabel , and remained low during a 30-min chase ( Fig . 1C ) . Furthermore , during a radiolabel for only 1-min HC synthesis was markedly reduced during RhCMV infection ( Fig . S1B ) . To rule out that HC was not recovered due to epitope masking by a viral protein or because HC was in a complex with NP40-insoluble proteins , we lysed cells in SDS to disrupt protein complexes and denature the HC prior to IP . Using either a monoclonal antibody that recognizes only free HC ( HC-10 ) [25] or K455 , we were unable to recover increased amounts of HC under these conditions ( Fig . 1E ) . Taken together these data suggest that RhCMV either prevents complete HC synthesis or degrades HC prior to complete protein synthesis . Since co-translational degradation is mediated by proteasomes [26] we wanted to determine whether HC translation could be completed in the presence of proteasome inhibitors . TRFs were infected with ΔRhUS2-11 and treated with the proteasomal inhibitor MG132 . However , no significant increase in HC recovery was observed either when total MHC-I was immunoprecipitated with K455 from NP40-lysates or with HC-10 from SDS-lysates ( Fig . 1F ) . In contrast , HC was stabilized in cells transduced with Adenovirus expressing HCMV US11 . The proteasomal inhibitors Lactacystin and ZL3VS also failed to stabilize HC in ΔRhUS2-11-infected cells ( data not shown ) . Taken together these data strongly suggest that RhCMV inhibits expression of HC prior to or during polypeptide synthesis . Since this phenotype is observed in the absence of RhUS2-11 and is not present in HCMV , we further conclude that RhCMV contains one or more unique gene ( s ) encoding VIHCE . Since VIHCE seems to be specific to RhCMV , but absent from HCMV , we examined the RhCMV genome for potential candidate genes . The genomic region spanning ORFs Rh158 to rh180 , corresponding to the region between IE1/IE2 ( UL123/UL122 ) and US1 in HCMV , contains a large number of genes that are either specific to RhCMV or are homologous to genes frequently deleted in laboratory strains of HCMV [12] , [27] . To examine whether this region contains the VIHCE gene , we deleted Rh158–180 using the BAC-recombination strategy shown in Fig . 2A . Interestingly , Δ158–180 did not show any obvious growth defects despite such a large deletion ( data not shown ) . Moreover , pulse-chase labeling of Δ158–180-infected TRFs revealed initial synthesis of MHC-I followed by degradation ( Fig . 2B ) . This degradation could be inhibited by the proteasome inhibitor MG132 ( Fig . 2C ) . MG132 also stabilized a smaller , presumably deglycosylated , degradation intermediate ( * ) which is also observed in cells transfected with RhUS2 [21] . Thus , it seemed likely that Δ158–180 lacked VIHCE , and that in the absence of VIHCE HC was now degraded by the RhCMV homologues of US2 and US11 . To examine whether the combined deletion of RhUS2-11 and VIHCE would restore HC expression in RhCMV-infected cells , we created a recombinant lacking both Rh158–180 and RhUS2-11 ( Fig . 2A ) . As expected from the single deletions , the resulting double-deletion virus Δ158–180 , ΔRhUS2-11 did not display a growth defect in vitro ( not shown ) . When TRFs were infected with Δ158–180 , ΔRhUS2-11 , HC expression was similar to Mock-infected cells indicating that this recombinant virus no longer interfered with MHC-I expresson ( Fig . 2B ) . Taken together , these data indicate that the VIHCE gene is located within the Rh158–180 region of RhCMV . Furthermore , the fact that HC synthesis is observed in the absence of VIHCE supports our conclusion that VIHCE acts prior to the ER-associated degradation caused by the US2-US11 homologs . To identify the gene ( s ) coding for VIHCE we systematically deleted fragments of decreasing size within the Rh158–180 region in an iterative fashion ( Fig . 3A; Table S1 ) . We took advantage of the fact that HC is initially synthesized in cells infected with VIHCE-deleted virus but then degraded by US2 and US11 to distinguish between recombinants encoding or lacking VIHCE . Initially , viruses carrying deletions approximately spanning the left or right half of the Rh158–180 region were generated ( Fig . 3A ) . TRFs were infected with recombinants Δ158–168 and Δ167–180 and pulse-chase was performed . Since HC was expressed in TRFs infected with Δ167–180 and not in TRFs infected with Δ158–168 , we concluded that VIHCE was located in the Rh167–180 region . Similarly , HC was expressed in TRF infected with viruses Δ175–180 , Δ175–178 , Δ176–178 , and Δ177–178 , but not Δ167–174 , Δ179–180 , and Δ175–177 ( Fig . 3A ) . These data suggested that rh178 encodes VIHCE . The region encoding rh178 overlaps with several predicted ORFs and with a previously identified large intron of the US1-homologue Rh181 [28] ( Gene Bank Accession: AF474179 ) . To exactly determine the mRNAs encoding VIHCE we mapped the transcriptional start and stop sites of the rh178 ORF and generated additional , smaller deletions and point mutants within the rh178 coding region ( Figs . 3–4 ) . We performed 5′ and 3′ RACE as well as Northern blot analysis . Sequence analysis of the 5′ RACE product identified a transcription start site downstream of the originally predicted rh178 start codon ( Fig . 3D ) . The identified transcript is predicted to encode a shorter version of rh178 . 3′RACE and cDNA cloning further revealed additional splice products in this region: a shorter splice product lacking most of the rh178 protein encoding region ( rh178 . 4; Note that Rivailler et al . , ( 2006 ) have detailed additional predicted ORFs upstream of rh178 and denoted them rh178 . 1 , rh178 . 2 , and rh178 . 3 ) and the above mentioned large Rh181-transcript which does not contain rh178 since it is removed by splicing . These three transcripts share the same polyadenylation signal and 3′ terminus ( Fig . 3B ) . Northern blot analysis using the predicted rh178 coding region as probe revealed two transcripts ( Fig . 3C ) . A larger predominant transcript of approximately 1600bp corresponds to the expected size of rh178 . The smaller transcript may correspond to rh178 . 4 , a shortened rh178 , or an unidentified transcript of the opposite sense . These data confirm the expression of the rh178 transcript during infection and correct the prediction of its protein coding region . Kinetic analysis indicates that VIHCE is an early gene that is already expressed within 4 hours of infection ( Fig . S3 ) . To determine whether the protein encoded by rh178 is responsible for VIHCE , we created a frameshift in the 5′-end of the predicted coding region ( rh178FS ) ( Fig . 4A ) . Since the primer-directed mutagenesis strategy also caused deletion of a portion of the 5′-UTR we generated a control virus ( rh178FSCtrl ) containing the same modification of the predicted 5′-UTR of rh178 but no frameshift ( Fig . 4A ) . While rh178FSCtrl inhibited HC expression similar to WT ( Fig . 4B ) , HC was synthesized in rh178FS-infected TRFs ( Fig . 4C ) . Thus , VIHCE is mediated by the rh178-encoded protein . The rh178 protein ( Fig . 5A ) , with a molecular weight of approximately 24 kDa , does not display significant homology with non-RhCMV sequences in the genomic database . A stretch of highly hydrophobic amino-acids beginning at amino acid 14 is predicted to represent a non-cleaved amino-terminal signal anchor ( Fig 5B ) . Thus , the most likely topology for this protein is that of a type 1b transmembrane protein , i . e . a large cytoplasmic C-terminus following the signal-anchor . Immunofluorescence analysis of epitope-tagged rh178 indicates that the protein localizes to the ER , suggesting that rh178 is anchored in the ER-membrane ( Fig . 5E ) . To obtain better expression of rh178 for further analysis , we constructed replication-defective adenovirus vectors expressing either wild type rh178 ( Ad178 ) or HA-tagged rh178 ( Ad178-HA ) . While there is a predicted glycosylation site at position N101 , digestion of whole cell lysate from Ad178-HA transduced cells with peptide:N-Glycosidase F ( PNGase ) failed to cause a shift in rh178 migration , while a shift was seen with MHC-I HC ( Fig . 5C ) . Thus , rh178 does not appear to be glcosylated and this is a further indication that the C-terminus of rh178 is located in the cytosol . To determine if rh178 by itself was capable of VIHCE , we transduced TRFs and performed pulse-chase analysis . Cells transduced with Ad178 exhibited reduced expression of HCs while β2-m was unaffected ( Fig . 5D ) , similar to the HC inhibition observed in RhCMV-infected cells ( Fig . 1 ) . MHC-I HC in cells transduced with a control adenovirus vector , AdTrans , was not affected . Thus , rh178 is both necessary and sufficient for VIHCE . Our data suggest that VIHCE prevents expression of the majority of HCs prior to completion of protein synthesis . Residual , VIHCE-resistant HCs are eliminated by RhUS2-11 . The dramatic reduction of newly synthesized HC observed even in the presence of proteasome inhibitors further suggests that VIHCE either blocks transcription of HC mRNA , completion of HC protein synthesis , or causes HC degradation in a proteasome-independent manner . However , the levels of HC mRNA did not change upon RhCMV-infection as shown by Northern blot ( Fig . 6A ) and by quantitative RT-PCR ( data not shown ) . Additionally , the size of the HC mRNA was unaltered in RhCMV-infected cells suggesting that mRNA is not cleaved , alternative spliced or degraded by RhCMV . We further determined whether HC mRNA is polyadenylated and exported into the cytoplasm by isolating nuclear , cytoplasmic , and polyadenylated RNA fractions from infected cells . We did not observe a significant difference in any of these fractions compared to Mock-infected cells ( data not shown ) . These data indicate that HC mRNA transcription , poly-adenylation , splicing and export to the cytosol is not affected by RhCMV . To determine whether the association of HC mRNA with ribosomes is inhibited we analyzed the polyribosome distribution of HC mRNA [29] . When sucrose-gradient fractions from lysates of Mock-infected or RhCMV-infected TRFs were analyzed by Northen blots , HC mRNA sedimented to the polyribosome fractions 12 and 13 in both Mock- and RhCMV-infected cells ( Fig . 6B ) . Small shifts in polyribosome density were observed in RhCMV infection for both HC and GAPDH mRNA , suggesting virus infection causes a slight reduction of ribosomal occupancy on cellular transcripts . Therefore , it seems that VIHCE does not inhibit the association of polyribosomes with HC mRNA . While sedimentation to the polyribosome fraction indicates the association of HC mRNA with ribosomes , it was possible that the ribosomes were not active . In order to determine if the ribosomes associated with the HC mRNA are actively translating we incubated cells with puromycin . Puromycin is a polypeptide chain terminator that requires an active peptidyl transferase to cause ribosome dissociation from transcripts . A short ( 4 min ) incubation with puromycin caused a shift in the polyribosome profile of HC mRNA in both RhCMV and Mock-infected cells , indicating ribosome dissociation ( Fig . 6C ) . This result indicates that the ribosomes bound to the HC mRNA are actively translating and not simply stalled on the transcript . Taken together these data suggested that HC mRNA is transcribed normally in RhCMV-infected cells and that protein translation is not inhibited at the level of initiation or elongation . However , since full-length HC protein cannot be recovered it seems most likely that HC translation is not completed . Observations similar to VIHCE were reported for translation inhibition by microRNAs that bind to the 3′-UTR of target transcripts . Similar to VIHCE , mRNAs that are targeted by a given microRNA are found in an active polyribosomal complex but a translated polypeptide intermediate can not be recovered even in the presence of proteasome inhibitors [30] . To examine the possibility that VIHCE targets the 3′-UTR of HCs we tested the ability of VIHCE to block synthesis of HC with or without its native 3′-UTR . Since antibodies to rhesus HCs are not available , and VIHCE is able to block expression of human HCs ( Fig . 7A ) , we chose to examine VIHCE function on HLA-A3 . To determine whether the 3′-UTR was required for this inhibition we transiently expressed HLA-A3 with or without its native 3′-UTR in TRFs . Following transfection we infected cells with either RhCMV containing VIHCE ( ΔRhUS2-11 ) or RhCMV lacking VIHCE ( Δrh178 , ΔRhUS2-11 ) . Expression of both HLA-A3 carrying the native 3′-UTR and a heterologous vector-derived 3′-UTR sequence was reduced by VIHCE ( Fig . 7B ) . The 5′-UTR was vector-derived in both constructs . Therefore , we conclude that VIHCE does not target the UTRs of HC mRNA . Translation of type I transmembrane proteins such as HC is dependent upon an N-terminal signal peptide ( SP ) that mediates translocation across the ER membrane . Upon translation initiation , the SP is recognized by the signal-recognition particle ( SRP ) which binds to the SP and arrests translation . This is followed by docking of the translation complex to the SRP-receptor which aids the transfer of the ribosomal/mRNA/nascent polypeptide complex to the SEC61 translocon [31] . Translation then resumes and the nascent polypeptide chain is imported into the lumen of the ER . The fact that VIHCE requires the HC coding sequence suggested that the HC protein might be at least partially translated and that VIHCE acts on the nascent polypeptide . Compared to human HC , we observed that the murine MHC-I molecule H2-Kb was more resistant to VIHCE ( data not shown ) . We hypothesized that this resistance was encoded in the amino-terminus of H2-Kb , specifically the SP . To test this hypothesis we replaced the SP of HLA-A3 with that of H2-Kb . As a further control , we also introduced the SP of CD4 which is more divergent from the HLA-A3 SP ( Fig . 7C ) . In both instances we observed that expression of the chimeric protein was much less reduced by virus expressing VIHCE compared to native HLA-A3 . Remarkably , the SP of Kb is quite similar to that of HLA-A3 ( Fig . 7C ) yet HLA-A3 expression was restored to almost the same levels as observed for the CD4 SP ( Fig . 7D ) . Therefore , we conclude that the SP of primate MHC-I is required for VIHCE to inhibit HC translation . The fact that VIHCE requires the MHC-I SP further suggests that VIHCE interferes with SP-dependent translocation which would lead to translation arrest and rapid , co-translational destruction of the resulting protein fragments . We next examined if the MHC-I SP is sufficient for VIHCE recognition . To test this we created a chimeric CD4 molecule with the HLA-A3 signal peptide in place of the native CD4 signal peptide ( A3/CD4 ) . When either wild type CD4 or A3/CD4 was expressed in TRFs , neither molecule was significantly affected by the presence of VIHCE , whereas the endogenous MHC-I HC was decreased ( Fig . 7E , 7D ) . This indicates that while the MHC-I SP is necessary for recognition by VIHCE , it is not entirely sufficient .
We report here that the ORF rh178 of RhCMV encodes a novel immune modulatory function , viral inhibitor of heavy chain expression ( VIHCE ) , which prevents the translation of HC in a signal-peptide dependent , but not sufficient , manner . This finding is surprising because RhCMV additionally expresses the HCMV US2-US11 homologs that also interfere with MHC-I stability and assembly . The VIHCE-encoding rh178 is so far unique to RhCMV suggesting that rh178 represents an adaptation to the evolutionary pressure of the non-human primate MHC system . Our previous observations [21] suggested that the immune evasion mechanisms encoded by the US2-US11 region predate the separation of human and old-world non-human primates which is assumed to have taken place about 25 million years ago [32] . Recent sequence analysis of the MHC-I locus in RM revealed that the MHC-I has undergone a tremendous change since then . Whereas a typical human or ape haplotype contains “only” six active MHC-I genes , as many as 22 different MHC-I genes are expressed in rhesus . Moreover , the sequence divergence was estimated to be 10-fold higher and genes have been duplicated at an approximately three times greater rate than in humans [33] , [34] . Thus , it is conceivable that the additional MHC-I genes forced RhCMV to evolve additional countermeasures . It is known that polymorphic MHC-I proteins are differentially affected by US2 and US11 of HCMV [35] , [36] , although the exact rules of this discrimination still need to be determined . Moreover , each of the US6-family viral immune modulators interferes at a distinct step during the assembly cascade [37] . Allele-specificity has also been reported for MCMV which contains three genes [38] , unrelated to either the US6-family or VIHCE , and each of three MCMV-gene products interferes with a different step of MHC-I assembly [39] . Thus , it seems that CMVs optimize their interference mechanisms , both within a given organism by sequentially attacking MHC molecules during assembly and within a given population by broadening the allele-specificities of these attacks . This conclusion is also supported by our finding that RhCMV lacking either rh178 or RhUS2-11 only partially suppressed MHC-I assembly and transport compared to WT RhCMV . This is either due to differences in allele-specificity within a given animal or an incomplete elimination of all alleles . The finding that RhCMV has a larger number of gene products interfering with MHC-I assembly than either HCMV or MCMV thus correlates with the observation that RM have a larger number of active MHC-I alleles than either human or mouse . The extracellular domains of MHC-I , particularly the peptide-binding regions , are highly polymorphic and evolve rapidly . In contrast , the cleaved signal peptide is highly conserved among different MHC-I alleles including the RM MaMu and the human HLA genes [33] . Many signal peptides for MaMu-I , MaMu-3 and MaMu-A show less than 3 amino-acids difference to either HLA-A , B or C alleles and some MaMu-SPs are identical to HLA-SPs [40] . A possible reason for the high conservation of HLA signal peptide sequences is the fact that a conserved nona-peptide ( VMAPRTLLL in the HLA-A3 sequence ) is presented by the non-polymorphic HLA-E molecule to the negative signaling receptor CD94/NKG2A or C of NK cells [41] . This system seems to be conserved in RM , although some alleles start at the methionine within the peptide [33] . Interestingly , the SP of the HCMV UL40 glycoprotein contains this nona-peptide which is presented by HLA-E in HCMV-infected cells in a TAP-independent fashion [42] , [43] . By loading the decoy peptide onto HLA-E , HCMV is thought to prevent the “missing self” stimulation of NK cells by MHC-I downregulation . Importantly , this nona-peptide is also encoded within the SP of Rh67 of RhCMV which otherwise shares only 19% identity with UL40 [12] . Since VIHCE requires polypeptide sequence beyond the SP in MHC-I HCs , the Rh67 protein is likely resistant to VIHCE despite containing a similar SP sequence . The MHC-I SP mimic contained in UL40 sets precedence for CMV taking advantage of the highly conserved SP to escape the cellular immune response . Different from UL40 however , rh178 does not mimic the SP , but seems to rely at least in part on this conserved sequence to broadly eliminate HCs . VIHCE is clearly different from any other previously described immune modulatory mechanism since the ER-localized protein rh178 interferes with HC expression after the onset but prior to the completion of translation . One possible mechanism is that rh178 inhibits translation at a step that occurs after the SRP targets the nascent polypeptide/ribosomal complex to the ER membrane-localized SRP receptor . During this process , translation is arrested until SRP is released upon GTP hydrolysis and SEC61 binding [31] , [44] . A possible scenario is that rh178 interacts with the SRP/nascent polypeptide/ribosome complex at the ER-membrane thus prolonging translational arrest . Alternatively , rh178 could prevent this complex interaction with the SEC61 translocon in ER-membrane . Conceivably , rh178 could also interfere with the translocation of HC in a manner similar to cotransin , a small molecule translocation inhibitor , which specifically interferes with binding of certain SPs to a SEC61 subunit [45] . The ensuing translocational stalling results in co-translational degradation by the proteasome , a process that involves cytosolic chaperones [46] . For non-stop RNA it was recently also shown that translational arrest results in protein fragments that are rapidly degraded by the proteasome [47] . Therefore , it seems likely that HC translation intermediates are degraded by the proteasome despite the fact that we were unable to detect a degradation intermediate in the presence of proteasome inhibitors . Possible reasons why such breakdown products were not identified are their potentially small and heterogenous size and their extremely rapid degradation . HC-derived intermediates might also lack the epitopes recognized by the HC-specific antibodies used in this study . Targeted disruption of protein translation by a viral protein has so far not been described as an immune evasion strategy . However , it was recently shown that the microRNA miR-UL112 of HCMV inhibits the translation of MHC-I-related chain B ( MICB ) , a ligand for the activating NK cell receptor NKG2D [48] . Thus , CMVs seem to interfere at multiple levels and by multiple strategies with translation of immune stimulatory genes . The virus might thereby employ or mimic cellular pathways of translational or translocational regulation . Further elucidation of the molecular events of VIHCE might thus reveal previously unrecognized host cell mechanisms of translational and translocational control .
Telomerized rhesus fibroblasts ( TRFs ) [49] and telomerized human fibroblasts ( THFs ) were obtained from Jay Nelson and maintained in Dulbecco's modified eagle's medium ( DMEM ) with 10% fetal bovine serum , 100U/mL penicillin and 100ug/mL streptomycin . RhCMV strain 68 . 1 was obtained from Scott Wong [12] and propagated in TRFs . Recombinant RhCMVs were created as described in the supplemental methods using the RhCMV BAC obtained from Peter Barry [23] . Recombinant rh178 adenoviruses were created using the AdEasy vector system according to the manufacturers protocol ( Stratagene ) . Adenoviruses AdTrans and AdUS11 were obtained from David Johnson . HLA-A3 and CD4 constructs were expressed from a modified version of pCDNA3 . 1 ( - ) ( Invitrogen , Carlsbad , CA ) in which the CMV promoter was replaced with the EF1α promoter ( obtained from Jay Nelson ) to create pEF1α . HLA-A3 was obtained by PCR from Jurkat T-cell cDNA using the forward primer 5′ctggaattcatggccgtcatggcgccccgaac and the reverse primer 5′gtcggatcctcacactttacaagctgtgag to amplify the coding region only or the reverse primer 5′gtcggatccttaggaatcttctcc to include the 3′UTR . pEF1α expression plasmids were electroporated into TRFs using the AMAXA Nucleofector II ( AMAXA Biosystems , Gaithersburg , MD ) using cell line solution L and the T-030 program . 1e6-2e6 TRFs were resuspended in 100 µl AMAXA solution and 2 µg expression plasmid . After electroporation cells were recovered in 500 µl RPMI for 45min at 37°C , and then plated in prewarmed complete DMEM . Transfection efficiency was monitored with a GFP reporter and was consistently >90% . Infections with RhCMV were performed 24 hours after electroporation . Cells were starved for 30-min , except where noted , using DMEM without serum , methionine ( Met ) or cysteine ( Cys ) . Labeling was performed for indicated times using Pro-mix 35S-Met/Cys ( GE Healthcare ) at 400 µCi/mL . To chase the label , cells were washed 3× in phosphate buffered saline ( PBS ) followed by incubation at 37°C in DMEM with 10% FBS containing 90 µg/mL Met and 188 µg/mL Cys . For NP-40 lysis , cells were lysed for 30 minutes at 4°C in 1% NP-40 in PBS with complete protease-inhibitor cocktail ( Roche ) . For SDS lysis , cells were lysed for 10 minutes at 25°C in 0 . 6% SDS in PBS with complete protease-inhibitor cocktail , then diluted in 3× volume of 1 . 2% triton X-100 in PBS prior to immunoprecipitation . For glycosidase treatment , PNGase was obtained from NEB and used according to the manufacturers protocol after NP-40 lysis . Polyclonal sera K455 recognizes both chains of the MHC-I heterodimer , assembled and unassembled ( obtained from Per Peterson ) [24] . HC-10 only recognizes free MHC-I heavy chains [25] . HLA-A3 antibody was purified from the GAP A3 hybridoma , obtained from ATCC ( HB-122 ) . Antibodies to Calreticulin , Transferrin , Vimentin , HA and FLAG were obtained , respectively , from Stressgen ( Victoria , BC ) , Zymed ( S . San Francisco , CA ) , Biomeda ( Burlingame , CA ) , Santa Cruz , and Sigma . Human CD4 antibody ( AHS0412 ) was obtained from Invitrogen . Secondary Alexa Fluor-conjugated antibodies 594 goat anti-rabbit and 488 goat anti-mouse were obtained from Invitrogen . Approximately 5×106 TRFs were either Mock-infected or RhCMV-infected for 24 hours . Fresh media was placed on the cells for 45-minutes , and cells were placed on ice and washed 2× with cold PBS containing 0 . 1 mg/ml cycloheximide ( Sigma ) . All subsequent steps were performed at 4°C . Cells were lysed for 10 min using 600 µl of polysome lysis buffer ( 15mM Tris , pH 7 . 4 , 15mM MgCl2 , 0 . 3M NaCl , 1% Triton x-100 , 0 . 1 mg/mL cycloheximide , 1 mg/mL heparin ) . Lysates were cleared at 12 , 000× g for 10 min . The supernatant was layered onto the top of a 10–50% sucrose gradient composed of sucrose in polysome lysis buffer excluding Triton x-100 . The gradients were centrifuged at 35 , 000 rpm in a Sorvall SW-41 rotor for 3 hours . 750 µl fractions were collected from the top of the gradient . After adding 4 . 25ml of 5 . 65M guanidine HCl , each fraction was ethanol precipitated ( −20°C overnight ) . RNA was pelleted at 15 , 000× g for 30 min , washed with 70% ethanol , dried at 25°C , and resuspended in 400 µl RNAse-free water . RNA was then re-precipitated by adding 40 µl 0 . 3M sodium acetate and 900 µl 100% ethanol , washed with 70% ethanol and resuspended in 50 µl RNAse-free water . For Northern blotting , 10 µl of each fraction was separated on a denaturing 1% agarose gel containing 1× MESA ( Boston BioProducts , Worcester , MA ) and 3 . 7% formaldehyde and transferred to Immobilon-Ny+ nylon membrane ( Millipore ) by capillary blotting in 20×SSC . RNA was fixed by air drying at 25°C for 30 min and baking at 80°C for 2 hours . Radiolabeled probes were generated by random priming . After denaturing at 100°C for 10 min , the probe was chilled on ice and added to 5mL ExpressHyb hybridization solution ( Clontech ) for hybridization . Membranes were pre-hybridized for 30 min at 68°C followed by probe hybridization for 2 hours , rinsed and washed twice with 2× SSC , 0 . 05%SDS followed by two washes in 0 . 1× SSC , 0 . 1% SDS . Transfected cells were fixed with 3 . 7% formaldehyde for 40 minutes , washed twice with PBS , quenched with 50mM NH4Cl for 10 min , washed twice with PBS , and permeabilized with 0 . 1% Triton X-100 in PBS for 7 min prior to staining . Total RNA from TRFs infected with WT RhCMV ( or RhCMV lacking rh175–178 as a negative control ) for 24 hours was used . For 3′ RACE , cDNA was synthesized using an oligo-dT anchor ( 5′gaccggatccgaattcgtcgacttttttttttttttttv ) . PCR was performed from cDNA using a PCR anchor primer ( 5′-gaccggatccgaattcgtcgac ) and a gene specific primer . For 5′ RACE , cDNA was synthesized with a gene specific primer ( rh178 5′-catttgcatgcagctgtgcg ) . 10 µg cDNA was then treated with terminal deoxynucleotidyl transferase and 0 . 5mM dATP at 37°C for 30 min , followed by purification and PCR using a nested gene specific primer ( rh178 5′-gcgcgaaacacgcgtttgc ) and the oligo-dT anchor .
|
To avoid immune detection by cytotoxic T lymphocytes , viruses interfere with antigen presentation by major histocompatibility complex class I ( MHC-I ) molecules . We have discovered a unique cytomegaloviral protein that interferes with the biosynthesis of MHC-I heavy chains and was thus termed viral inhibitor of heavy chain expression ( VIHCE ) . We show that VIHCE does not affect transcription of MHC-I mRNA or the formation of poly-ribosomes . Surprisingly , however , very little MHC-I protein is detected , even when proteasomal protein degradation is inhibited , suggesting incomplete protein translation . Interestingly , VIHCE requires the proper MHC-I signal peptide , suggesting that CMV takes advantage of the high conservation of MHC-I signal peptides and interferes with protein translation by inhibiting signal sequence-dependent protein translocation . This is the first description of a viral protein that specifically targets the translation of a cellular immuno-stimulatory protein .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"immunology/antigen",
"processing",
"and",
"recognition",
"immunology/immunomodulation",
"virology/animal",
"models",
"of",
"infection",
"virology",
"virology/immune",
"evasion",
"molecular",
"biology/translational",
"regulation"
] |
2008
|
Signal Peptide-Dependent Inhibition of MHC Class I Heavy Chain Translation by Rhesus Cytomegalovirus
|
Salt stress is one of the major abiotic factors that affect the metabolism , growth and development of plants , and soybean [Glycine max ( L . ) Merr . ] germination is sensitive to salt stress . Thus , to ensure the successful establishment and productivity of soybeans in saline soil , the genetic mechanisms of salt tolerance at the soybean germination stage need to be explored . In this study , a population of 184 recombinant inbred lines ( RILs ) was utilized to map quantitative trait loci ( QTLs ) related to salt tolerance . A major QTL related to salt tolerance at the soybean germination stage named qST-8 was closely linked with the marker Sat_162 and detected on chromosome 8 . Interestingly , a genome-wide association study ( GWAS ) identified several single nucleotide polymorphisms ( SNPs ) significantly associated with salt tolerance in the same genetic region on chromosome 8 . Resequencing , bioinformatics and gene expression analyses were implemented to identify the candidate gene Glyma . 08g102000 , which belongs to the cation diffusion facilitator ( CDF ) family and was named GmCDF1 . Overexpression and RNA interference of GmCDF1 in soybean hairy roots resulted in increased sensitivity and tolerance to salt stress , respectively . This report provides the first demonstration that GmCDF1 negatively regulates salt tolerance by maintaining K+-Na+ homeostasis in soybean . In addition , GmCDF1 affected the expression of two ion homeostasis-associated genes , salt overly sensitive 1 ( GmSOS1 ) and Na+/H+ exchanger 1 ( GmNHX1 ) , in transgenic hairy roots . Moreover , a haplotype analysis detected ten haplotypes of GmCDF1 in 31 soybean genotypes . A candidate-gene association analysis showed that two SNPs in GmCDF1 were significantly associated with salt tolerance and that Hap1 was more sensitive to salt stress than Hap2 . The results demonstrated that the expression level of GmCDF1 was negatively correlated with salt tolerance in the 31 soybean accessions ( r = -0 . 56 , P < 0 . 01 ) . Taken together , these results not only indicate that GmCDF1 plays a negative role in soybean salt tolerance but also help elucidate the molecular mechanisms of salt tolerance and accelerate the breeding of salt-tolerant soybean .
Soil salinity is a global ecological issue that has severely affected plant growth and development and decreased agricultural production . A high-salt environment causes various damages to crops , such as plant water loss , high osmotic stress , and homeostasis and ion imbalances in plant cells [1 , 2] . The cultivation of salinity-tolerant plants and improvements in their adaptability to saline-alkali soils are urgently needed . Understanding the mechanism of salt tolerance in plants is the most crucial basic knowledge needed for the breeding of salt-tolerant plants [3 , 4] . Soybean is a traditional edible leguminous crop that provides abundant protein , high-quality vegetable oils and a variety of physiologically active substances to human beings . In addition , soybean is a moderately salt-tolerant crop , and the yield of soybean is significantly reduced if the soil salinity exceeds 5 dS/m [5] . Plant salt tolerance is a complex quantitative trait that is affected by numerous genetic and nongenetic factors [6 , 7] . In recent years , both forward and reverse genetic approaches have been applied to reveal the functions of key salinity response genes in soybean . Biparental quantitative trait locus ( QTL ) mapping and genome-wide association studies ( GWAS ) have been used as effective and precise tools to detect QTLs associated with salt tolerance , and a number of QTLs for salt tolerance have been detected in previous studies [8–14] . For example , a major QTL was detected and mapped near the single sequence repeat ( SSR ) marker Sat_091 on chromosome 3 , and this QTL explained 41% and 60% of the total phenotypic variation observed under greenhouse and field conditions , respectively , in an F2:5 population derived from a cross of the cultivars S-100 ( salt-tolerant line ) and Tokyo ( salt-sensitive line ) [11] . Other researchers recently confirmed this QTL in a different mapping population [8 , 9 , 13–15] , and a gene at this major locus , Glyma03g32900 , was identified and cloned [10 , 15–18] . Although the major QTLs related to salt tolerance on chromosome 3 were consistently mapped in several studies , some reports indicated that additional QTLs/genes contribute to salt tolerance in the soybean genome . For example , a previous study showed that 45 SNPs mapped on chromosomes 2 , 3 , 7 , 8 , 10 , 13 , 14 , 16 , and 20 , including 31 SNPs on chromosome 3 mapped at or near the previously reported major salt tolerance QTL , are significantly associated with salt tolerance [19] . Through both field and greenhouse experiments , other researchers identified a major salt tolerance-related QTL on chromosome 18 in F7:11 recombinant inbred lines ( RILs; 184 ) [20] . Our previous study found that eight SNPs and 13 suggestive SNPs are associated with salt tolerance indices and verified that five candidate genes located on chromosomes 8 , 9 and 19 are associated with the response to salt stress at the soybean germination stage by association mapping [21] . A novel QTL associated with the leaf sodium content ( LSC ) , which showed a high logarithm of odds ( LOD ) value ( 4 . 56 ) and R2 ( 11 . 5% ) , was identified on chromosome 13 using the soybean physical map of 132 F2 plants [13] . The above studies largely focused on salt tolerance at the soybean seedling stage but rarely investigated salt tolerance at the germination stage . However , seed germination is critical for ensuring new seedlings and enhancing yield during the plant growth cycle . It has been suggested that soybean germplasms have different degrees and mechanisms of salt tolerance at different developmental stages [5] . Although QTLs for salt stress at the germination stage have been identified by linkage mapping [22 , 23] and association analysis [21 , 24] , there is little information regarding the genetic mechanisms of salt tolerance at the germination stage . Understanding the genetic mechanism of salt tolerance at the seed germination stage is very important for improving the salt tolerance of soybean . In this study , linkage mapping and genome-wide association study were performed to dissect the genetic architecture of salt tolerance at the soybean seed germination stage , and a major QTL , qST-8 , was found to be significantly associated with salt tolerance . Furthermore , an investigation combining whole-genome sequence , bioinformatics , and gene expression analyses as well as plant transformation demonstrated that a cation diffusion facilitator ( CDF ) , GmCDF1 , negatively regulates salt tolerance through Na+-K+ homeostasis in soybean . Additionally , haplotype and candidate-gene association analyses in 31 natural soybean varieties confirmed that GmCDF1 plays a negative regulatory role in salt tolerance .
The means , standard deviations and ranges of four germination-related traits—imbibition rate ( IR ) , germination index ( GI ) , germination potential ( GP ) and germination rate ( GR ) —and four salt tolerance indices—the ratio of the imbibition rate under salt conditions to the imbibition rate under no-salt conditions ( ST-IR ) , the ratio of the germination index under salt conditions to the germination index under no-salt conditions ( ST-GI ) , the ratio of the germination potential under salt conditions to the germination potential under no-salt conditions ( ST-GP ) and the ratio of the germination rate under salt conditions to the germination rate under no-salt conditions ( ST-GR ) of RILs and natural populations—are presented ( S1 Table ) . The mean values of the four germination-related traits obtained in the presence of 150 mM NaCl were lower than found under normal conditions , which indicated that salt stress depressed the growth and development of soybean at the germination stage . Moreover , the level of salt tolerance-related traits , with the exception of ST-IR , varied widely in both the RILs and natural populations . The mean ST-IR , ST-GI , ST-GP and ST-GR varied from 0 . 85 to 0 . 99 , from 0 . 25 to 1 . 02 , from 0 . 20 to 0 . 88 and from 0 . 41 to 1 . 00 , respectively , in the RILs population and from 0 . 86 to 0 . 97 , from 0 . 21 to 0 . 88 , from 0 . 00 to 0 . 92 and from 0 . 24 to 1 . 09 , respectively , in the natural population . Two parents of the RILs population , Kefeng No . 1 and Nannong 1138–2 , showed different tolerances to salt stress ( S1 Table ) . The ST-GR and ST-GI of Kefeng No . 1 were higher than those of Nannong 1138–2 , indicating that Nannong 1138–2 was more sensitive to salt stress than Kefeng No . 1 at the germination stage . An analysis of variance ( ANOVA ) showed that the genotype , environment and the genotype-by-environment interaction significantly influenced the four salt tolerance indices ( P<0 . 001 ) in the two populations ( S1 Table ) . In addition , significant ( P<0 . 001 ) genetic variations were found for the four germination-related traits and the salt tolerance indices in the two populations across four and three environments , respectively . Moreover , the phenotypic frequencies of ST-IR , ST-GI , ST-GP and ST-GR in the two populations approximately fit a normal continuous distribution , indicating that these four salt tolerance indices are quantitative traits controlled by multiple factors ( S1 Fig ) . Pearson’s correlations among the four salt tolerance indices were analyzed based on the means of two populations ( S2 Table ) . For the two populations , ST-IR was significantly negatively correlated with ST-G , ST-GP and ST-GR ( P<0 . 01 ) , whereas ST-GI was significantly positively correlated with ST-GP and ST-GR , and ST-GP was strongly positively correlated with ST-GR ( P<0 . 01 ) . Four salt tolerance indices of the RILs population in four different environments were used for QTL mapping . A total of 25 QTLs associated with four salt tolerance indices during the soybean germination stage were detected on chromosomes 1 , 2 , 7 , 8 , 10 , 15 , 17 and 18 , with LOD values ranging from 2 . 50 to 17 . 06 ( Table 1 ) . Three of the five QTLs for ST-IR , two of the four QTLs for ST-GI , three of the nine QTLs for ST-GP and two of the seven QTLs for ST-GR were significantly associated with salt tolerance and located on chromosome 8 . With the exception of qSTGP-8-2 , other QTLs related to salt tolerance and located on chromosome 8 ( named qST-8 ) that showed largely overlapping confidence intervals ( CIs ) were considered the same QTL . This QTL closely linked with the marker Sat_162 was detected mostly for the four salt tolerance indices in E1 , E2 , E3 , and E4 , contained the physical genetic region between the markers BE820148 and AW132402 ( Fig 1A ) and explained 6 . 25%–46 . 82% of the phenotypic variation . The marker BE820148 was detected for ST-GR once in E1 with significant LOD ( 8 . 85 ) and R2 ( 12 . 84% ) values close to the marker Sat_162 . This result suggested that the candidate gene might be located within the region between markers BE820148 and Sat_162 or closer to marker Sat_162 . A GWAS was conducted to detect SNPs associated with salt tolerance across three environments with 207 , 608 SNPs [minor allele frequency ( MAF ) >0 . 05] obtained from the NJAU 355K SoySNP array [25] , and the Manhattan and quantile-quantile plots for the GWAS results are shown in Fig 2 and S2 Fig . The 18 SNPs significantly ( with a significance threshold of -log10 ( P ) ≥5 . 32 ) associated with salt tolerance indices are listed in Table 2 . In addition , 74 SNPs with suggestive thresholds ( 4 . 5≤-log10 ( P ) <5 . 32 ) were also identified in the GWAS ( S3 Table ) . These SNPs are located on chromosomes 1 , 8 , 11 , 13 , 14 , 15 , 16 , 18 and 19 , indicating that the salt tolerance of soybean at the germination stage is controlled by multiple genes . Moreover , we found that 17 out of 18 significant SNPs and 48 out of 74 suggestive SNPs were located on chromosome 8 , forming a cluster flanked by the SNP markers AX-93912074 and AX-93634504 ( -log10 ( P ) ≥5 . 32 ) with a physical position of 7716458–8268861 bp . Interestingly , this cluster was located in qST-8 , which was identified by the previous linkage mapping in four environments ( Fig 1B ) , indicating that this cluster is critical for salt tolerance at the germination stage of soybean and that a candidate gene for salt tolerance can likely be identified . According to gene annotation on Phytozome 12 ( https://phytozome . jgi . doe . gov/pz/portal . html ) , 70 gene models are located within the above-described cluster . For fine mapping , we performed whole-genome sequencing on Kefeng No . 1 and Nannong 1138–2 , which are the parents of the RILs population used in this study , and the SNP density distribution within the soybean genome is shown in S2 Fig . We compared the whole genome of Kefeng No . 1 to that of Nannong 1138–2 and found that 273 SNPs were located on chromosome 8 between the SNP markers AX-93912074 and AX-93634504 . Among these SNPs , 42 nonsynonymous SNPs were located in the exons of 21 genes , and 15 SNPs were located in the 2 . 0-kb promoter regions of 11 genes , including three identical genes ( S4 Table ) . We performed a quantitative real-time PCR ( qRT-PCR ) analysis to investigate whether the expression of these 29 genes ( S4 Table ) in Kefeng No . 1 and Nannong 1138–2 was affected by salt stress . The results demonstrated that the expression of six genes ( Glyma . 08g101300 , 102200 , 103000 , 106100 , 106200 and 106400 ) was too low to be detected , whereas that of 16 genes did not change significantly in response to salt stress ( S3 Fig ) , and that of seven genes could be induced by salt stress ( S4 Fig ) at the soybean germination stage . Among these seven genes , six were induced by salt stress in both Kefeng No . 1 and Nannong 1138–2 , and only Glyma . 08g102000 was dramatically upregulated in Nannong 1138–2 but not in Kefeng No . 1 under salt stress . In fact , the expression level of Glyma . 08g102000 in Nannong 1138–2 was nearly 30-fold higher than that in Kefeng No . 1 after treatment with 150 mM NaCl , whereas only a 1-3-fold change was found for the other six genes . Thus , Glyma . 08g102000 , which is located within the QTL qST-8 detected in our study , might be a candidate gene involved in the regulation of salt tolerance in soybean . Glyma . 08g102000 encodes an 817-amino-acid protein , has a length of 2457 bp and is a member of the CDF family ( named GmCDF1 ) . A phylogenetic analysis indicated a close relationship between Glyma . 08g102000 and AtMTP12 ( S6A Fig ) . The TMHMM program ( http://www . cbs . dtu . dk/services/TMHMM/ ) predicted that GmCDF1 possesses 14 transmembrane domains ( TMs ) with cytosolic N and C termini , similarly to AtMTP12 ( S6B Fig ) . A previous report showed that AtMTP12 forms a functional complex with AtMTP5t1 to transport Zn into the Golgi and thereby increases tolerance to Zn stress in Arabidopsis [26] . However , another study found that in rice , Os08g32650 and Os01g03914 , two homologous genes of GmCDF1 , are responsive to salt stress , and the expression levels of these two genes are lower in two salt-tolerant mutant lines of rice than in sensitive wild-type plants under salt stress [27] . These results suggest that GmCDF1 might encode a cation diffusion facilitator and respond to salt stress . We performed a qRT-PCR analysis to explore the expression pattern of GmCDF1 in six representative soybean accessions , including three salt-tolerant accessions ( Kefeng No . 1 , NJAU_C051 and NJAU_C204 ) and three salt-sensitive accessions ( Nannong 1138–2 , NJAU_C071 and NJAU_C136 ) , at the germination stage under normal and salt stress conditions ( Fig 3A and 3B ) . As shown in Fig 3C , the expression of GmCDF1 in Kefeng No . 1 was not significantly enhanced after exposure to salt stress . In contrast , the expression of GmCDF1 in NJAU_C051 and NJAU_C204 , the other two salt-tolerant soybean accessions , was upregulated from 48 h and reached a peak value at 72 h . In Nannong 1138–2 and NJAU_C136 , the expression of GmCDF1 was upregulated after NaCl treatment for 24 h and reached a peak value at 48 h , and in NJAU_C071 , GmCDF1 was upregulated after NaCl treatment for 24 h and reached a peak value at 72 h . Interestingly , higher fold-changes in the expression of GmCDF1 after treatment with salt for 48 and 96 h were observed in the salt-sensitive accessions than in the salt-tolerant accessions ( Fig 3C ) . To further confirm the candidate gene , we investigated the expression patterns of GmCDF1 in different soybean tissues , and our results showed that GmCDF1 was expressed constitutively in most soybean tissues . The highest level of GmCDF1 transcript was detected in flowers , seeds and roots , whereas GmCDF1 was weakly expressed in leaves , pods and stems ( S7 Fig ) . The high expression level found in roots suggests that the function of GmCDF1 could be investigated using the soybean hairy root transformation system [28] . To investigate the role of GmCDF1 under salt stress , two constructs ( pMDC83-GmCDF1 and pBI-GmCDF1 ) were generated for overexpression ( GmCDF1-OE ) and RNA interference ( GmCDF1-RNAi ) analyses , respectively . Transgenic soybean hairy roots were produced using to the Agrobacterium rhizogenes-mediated hairy root transformation system [28] . The average expression level of GmCDF1 in GmCDF1-OE hairy roots was 31 . 8-fold higher than that in the wild-type strain K599-generated ( harboring the empty vector pMDC83 ) control hairy roots ( Control 1 ) , whereas the expression level of GmCDF1 in the GmCDF1-RNAi hairy roots was 53% lower compared with that in the control hairy roots ( Control 2 ) generated by strain K599 [harboring the empty vector pB7GWIWG2 ( II ) ] ( S8A and S8B Fig ) . In the presence of 0 mM NaCl , no apparent difference was found between the transgenic hairy roots and their controls , indicating that the overexpression or silencing of GmCDF1 had little impact on the growth of soybean hairy roots under normal conditions ( Fig 4A ) . However , after exposure to 75 mM NaCl for four days , obvious differences were observed between the transgenic plants and their controls . The GmCDF1-OE plants exhibited more sensitivity to salt stress than the Control 1 plants ( Fig 4A ) . The average fresh weight of the Control 1 hairy roots and shoots was significantly heavier than that of the GmCDF1-OE hairy roots ( Fig 4B and 4C ) . The GmCDF1-OE plants exhibited unhealthier leaves with lower average chlorophyll contents ( soil plant analysis development , SPAD ) than the Control 1 plants ( Fig 4D ) . Moreover , the expression of GmCDF1 was higher in the GmCDF1-OE hairy roots than in the Control 1 roots after treatment with NaCl for four days ( S8A Fig ) . Both the fresh weight of the hairy roots and shoots and the average SPAD value of the leaves of the GmCDF1-RNAi plants were higher than those of the Control 2 plants after treatment with 75 mM NaCl for four days ( Fig 4E–4G ) . Under salt stress , the expression of GmCDF1 was substantially suppressed in the GmCDF1-RNAi hairy roots compared with that in the Control 2 plants ( S8B Fig ) . These overexpression and silencing experimental results suggest that GmCDF1 negatively regulates salt tolerance in soybean . Excessive accumulation of salt usually leads to ion toxicity , which disrupts the metabolism of plants under salt stress . To assess the potential differences in the ion contents between the transgenic hairy roots and their corresponding control hairy roots , the ion contents of Na+ and K+ were determined by ICP-OES in this study . The results showed that in the absence of salt stress , the transgenic hairy roots showed little variation in the average Na+ and K+ contents compared with the corresponding control hairy roots , and the non-transgenic shoots exhibited similar results ( Fig 5A–5D ) . However , after treatment with 75 mM NaCl for four days , the average Na+ contents in the GmCDF1-OE hairy roots and non-transgenic shoots were significantly higher than those found in the Control 1 hairy roots and shoots ( Fig 5A ) , whereas the GmCDF1-RNAi hairy roots and non-transgenic shoots accumulated less Na+ than the Control 2 hairy roots and shoots ( Fig 5C ) . Moreover , the GmCDF1-RNAi hairy roots accumulated less K+ than the Control 1 roots under salt stress ( Fig 5B ) , whereas the GmCDF1-RNAi hairy roots accumulated more K+ than the Control 2 roots ( Fig 5D ) . To further investigate the role of GmCDF1 in salt stress adaptation in soybean , we analyzed the expression levels of salt stress-related genes with or without NaCl treatment . The relative expression of GmSOS1 in the GmCDF1-OE hairy roots was significantly lower than that in the Control 1 hairy roots , under both normal and salt stress conditions . Similar to GmSOS1 , the expression of GmNHX1 in the GmCDF1-OE hairy roots was significantly lower than that in the Control 1 hairy roots under both normal and salt stress conditions ( Fig 6B ) . In contrast , higher transcript levels of GmSOS1 and GmNHX1 were found in the GmCDF1-RNAi hairy roots compared with their control hairy roots , regardless of the presence of salt stress ( Fig 6C and 6D ) . These results suggest that the overexpression or silencing of GmCDF1 might affect the expression of GmSOS1 and GmNHX1 in soybean . Because the expression level of GmCDF1 in Nannong 1138–2 was nearly 30-fold higher than that in Kefeng No . 1 after treatment with 150 mM NaCl for 48 h , differences might exist between the promoter regions of GmCDF1 in Kefeng No . 1 and Nannong 1138–2 . Thus , the 2 . 0-kb promoter regions of GmCDF1 upstream of the start codon were cloned and sequenced , and the results showed that the promoter of GmCDF1 in Kefeng No . 1 was 747 bp shorter than that in Nannong 1138–2 . In fact , seven deletions were identified in the promoter region of GmCDF1 in Kefeng No . 1 , and these were located at -130 bp , -315~-319 bp , -736~-962 bp , -968~-985 bp , -994~-1318 bp , -1311~-1463 bp and -1469~-1686 bp ( upstream of the start codon ) . These deletions in the promoter regions of GmCDF1 might be the reason for the dramatic upregulation of GmCDF1 in Nannong 1138–2 but not in Kefeng No . 1 under salt stress . We also sequenced the GmCDF1 gene , an approximately 4 . 6-kb genomic region including the 2 . 0-kb promoter region of GmCDF1 upstream of the start codon and the 2 . 6-kb region of GmCDF1 from the 5'-UTR to 3'-UTR , in a subset of 31 soybean accessions representing varieties with high salt tolerance , moderate salt tolerance and low salt tolerance . The sequencing analysis identified 11 indels and 15 SNPs ( MAF>0 . 05 ) ( S5 Table ) that were retained for the subsequent association analysis . After sequencing , five of the 15 SNPs and three of the 11 indels were identified as nonsynonymous mutations , and the remaining 10 SNPs and eight indels were found to be synonymous mutations ( S5 Table ) . Furthermore , these 11 indels and 15 SNPs exhibited strong linkage disequilibrium ( LD ) and could form three LD blocks , as demonstrated by a LD analysis ( Fig 7A ) . Furthermore , a GmCDF1-based association analysis was performed to investigate the relationship between the allelic variation of GmCDF1 and salt tolerance . The results showed that only two SNPs , S-671 ( located 671 bp upstream of the start codon ) and S605 ( located 605 bp downstream of the start codon ) , were significantly associated with ST-GR ( Fig 7A ) , contributing to 20 . 17% and 32 . 50% of the phenotypic variations for ST-GR in the representative subset , respectively . The sequencing of GmCDF1 revealed that S-671 and S605 are located in the promoter region and exon of GmCDF1 , respectively . Based on these 11 indels and 15 SNPs , the 31 soybean genotypes were classified into ten haplotypes ( Hap1-Hap10 ) ( Fig 7B ) . Haplotype 1 ( Hap1 , n = 12 ) formed the largest group , Hap2 ( n = 8 ) was the second largest group , and the other eight haplotype classes constituted minor groups , each comprising one or two soybean accessions ( Fig 7A ) . The soybeans carrying Hap1 had significantly smaller ST-GR values than those carrying Hap2 ( p = 8 . 4×10−4 ) ( Fig 7C ) , which indicated that Hap1 was more sensitive to salt stress than Hap2 . With the comparison of these two haplotypes , only two different SNPs , S-671 and S605 , which are located in the promoter region and exon of GmCDF1 , respectively ( Fig 7B ) . However , S605 did not result in an amino acid change . As it is known , the promoter always plays a central role in transcriptional regulation , and the relative expressions of GmCDF1 were detected in seeds from these 31 soybean accessions treated or not treated with 150 mM NaCl for 48 h . Association mapping was performed using the relative expressions of GmCDF1 , and 15 polymorphic sites were significantly associated with the relative expressions of GmCDF1 in these 31 soybean accessions ( S7 Table ) . Except S605 , S-671 was the most significant polymorphic sites explaining 21 . 16% of the phenotypic variation among these polymorphic sites . Moreover , the soybean accessions carrying Hap1 showed higher GmCDF1 expression than those carrying Hap2 ( Fig 7D ) . These results suggested that S-671 , might be the SNP responsible for the difference in relative expressions of GmCDF1 , leading to different salt tolerance in soybean eventually . Additionally , it was found that the expression of GmCDF1 was negatively correlated with ST-GR in these 31 soybean accessions ( r = -0 . 56 , P < 0 . 01 ) . All these results suggested that the expression of GmCDF1 can partially explain the phenotypic variation in soybean salt tolerance .
Genes encoding members of the CDF family have been cloned from bacteria , yeast , plants , and animals [31–35] and play critical roles in cation accumulation , cation tolerance , signal transduction cascades and oxidative stress resistance [36–39] . In addition , plant CDF transporters usually play an important role in metal homeostasis and tolerance [40] . AtMTP1 overexpression enhances zinc resistance and accumulation in Arabidopsis [41] , and AtMTP3 and AtMTP11 enhances zinc tolerance and manganese tolerance , respectively [42 , 43] . In rice , OsMTP8 . 1 and OsMTP11 are involved in the uptake and translocation of manganese [38 , 44] . Two Beta vulgaris MTP members , BmMTP10 and BmMTP11 , and a cucumber CsMTP8 have also been identified as manganese transporters that confer increased tolerance to manganese [45 , 46] . Obviously , these above-mentioned studies of CDF/MTP proteins mainly focused on metal tolerance , such as Zn or Mn tolerance , in plants . In contrast to CDF proteins that are known to improve resistance to metal stress in plants , GmCDF1 overexpression led to sensitivity to salt stress , whereas the silencing of GmCDF1 enhanced tolerance to salt stress in soybean . As shown in the present study , GmCDF1 negatively contributes to the salt adaptation of soybean . To the best of our knowledge , this study provides the first demonstration that GmCDF1 is negatively associated with salt tolerance in soybean . Notably , the results showed a slight increase in the Zn2+ concentration of the GmCDF1-OE transgenic soybean hairy roots and a significant reduction in the Zn2+ concentration of GmCDF1-RNAi transgenic soybean hairy roots compared with their control roots under normal conditions ( S9A and S9D Fig ) , indicating that GmCDF1 might be involved in Zn2+ transport in soybean . In addition , Mn is not involved in GmCDF1-regulated salt tolerance because no significant differences in the Mn concentration were found between the transgenic soybean hairy root and non-transgenic soybean hairy roots under either normal or salt stress conditions ( S9B and S9E Fig ) . The maintenance of ion homeostasis is an important salt tolerance mechanism in soybean . In fact , maintaining an adequate K+ concentration and a high K+/Na+ ratio has been shown to be necessary for plant survival and growth under salt stress [47] . The salt overly sensitive ( SOS ) signaling pathway has been well characterized for salt tolerance , and AtSOS1 is indispensable for driving Na+ efflux from xylem parenchyma cells to root xylem under salt stress to maintain a relatively low Na+ concentration [48] . The overexpression of GmSOS1 in A . thaliana reduces Na+ accumulation in both the roots and shoots and enhanced tolerance to salt stress at the seed germination and seedling stages [49] . In addition , it has been reported that A . thaliana Na+/H+ exchanger 1 ( AtNHX1 ) is a Na+ , K+/H+ antiporter in Arabidopsis [50 , 51] . In tomato , the overexpression of AtNHX1 induces the accumulation of K+ in vacuoles as well as the transport of K+ from roots to shoots [52 , 53] . Moreover , the overexpression of GmNHX1 in Lotus corniculatus results in lower Na+ and K+ contents , a higher K+/Na+ ratio , and a higher salt tolerance compared with those of wild-type plants under salt stress [54] . The above-mentioned genes are all positively associated with salt tolerance in plants . However , we found a novel gene , GmCDF1 , that negatively regulated salt tolerance through Na+-K+ homeostasis in soybean . The overexpression of GmCDF1 enhanced Na+ absorption and depressed the accumulation of K+ under salt stress , which led to a higher Na+ content and lower K+ content in soybean hairy roots compared with those found in the Control 1 roots ( Fig 4A and 4B ) , and the opposite results were obtained with the silencing of GmCDF1 in soybean hairy roots ( Fig 4C and 4D ) . These results suggest that GmCDF1 might facilitate the accumulation of Na+ and depress the absorption of K+ , ultimately increasing the ionic toxicity caused by salt stress . Soybean salt tolerance is a complex quantitative trait affected by numerous genetic and non-genetic factors . GmSOS1 and GmNHX1 indirectly contribute to soybean tolerance to salt stress [49 , 54] . Moreover , the silencing of SlSOS1 in tomato ( Solanum lycopersicum ) results in hypersensitivity to salt stress [55] , and the nhx1 mutation reduces the establishment of A . thaliana seedlings compared with the wild-type protein under salt stress , indicating that salt tolerance is depressed if SOS1 or NHX1 expression is reduced in plants . However , a qRT-PCR analysis of for ion homeostasis-associated genes in transgenic hairy roots showed a negative correlation between the expression of GmCDF1 and two genes , GmSOS1 and GmNHX1 . GmSOS1 and GmNHX1 exhibited lower expression in the Control 1 roots than these in the GmCDF1-OE hairy roots under salt stress , respectively ( Fig 6A and 6B ) . In contrast , the expression levels of GmSOS1 and GmNHX1 were 1 . 8-fold and 2 . 0-fold higher , respectively , in the GmCDF1-RNAi hairy roots than in the Control 2 hairy roots under salt stress ( Fig 6C and 6D ) . These results suggest that the existence of crosstalk between GmCDF1 and two salt tolerance-related genes , GmSOS1 and GmNHX1 , and the gene expression data indicate that GmCDF1 negatively regulates salt tolerance in soybean . In addition to GmSOS1 and GmNHX1 , other salt tolerance-related genes , such as GmSALT3 , GmHKT1;4 and GmNcl , were detected in soybean hairy roots under normal and salt conditions , and no significant differences were detected in the expression of these three genes in the GmCDF1-OE hairy roots or in the GmCDF1-RNAi hairy roots when exposed to salt stress ( S10 Fig ) . Thus , these three genes might not be affected by GmCDF1 . In addition , published experimental evidence proves the importance of Ca2+ for salt adaptation [56 , 57] . In this study , the Ca2+ concentration in transgenic soybean hairy roots and non-transgenic shoots was not significantly changed compared with that of the control plants under either normal or salt stress conditions ( S9C and S9D Fig ) , indicating that GmCDF1 might not affect the transportation of Ca2+ in soybean . However , the mechanism through which GmCDF1 regulates the Na+-K+ balance directly remains to be elucidated in future studies . In recent years , CRISPR/Cas9 technology has been widely used to introduce targeted mutations for studying gene function in plants [58–60] , and this powerful tool will be used to explore this gene in our future study .
The linkage mapping population consisting of 184 RILs ( designated as NJRIKY ) was derived from a cross between Kefeng No . 1 and Nannong 1138–2 and was developed by single-seed descent at the National Center for Soybean Improvement of China [61 , 62] . A natural population including 211 cultivated soybean accessions was used for the GWAS ( S1 Table ) . Seeds of the RILs population were collected from four environments: the Jiangpu Experimental Station of the Nanjing Agricultural University ( 32 . 12° N 118 . 37° E ) , Nanjing , China , in 2012 ( E1 ) , 2013 ( E2 ) and 2014 ( E3 ) and the Experimental Farm of the Jiangsu Yanjiang Institute of Agricultural Sciences ( 31 . 58° N 120 . 53° E ) , Nantong , China , in 2015 ( E4 ) . The seeds used for the GWAS were obtained from the following three environments: the Jiangpu Experimental Station of Nanjing Agricultural University ( 32 . 12°N 118 . 37°E ) in Nanjing , China , in 2012 and 2013 ( E1 and E2 , respectively ) and the Experimental Farm of the Jiangsu Yanjiang Institute of Agricultural Sciences ( 31 . 58°N 120 . 53°E ) in Nantong , China , in 2015 ( E4 ) . Prior to germination , the seeds were sterilized with a chlorine gas atmosphere to minimize the danger of microbial contamination and infection . Forty uniform healthy weighed seeds were then placed on two sheets of filter paper ( in sterilized Petri dishes ) and treated with 15 mL of water or 150 mM NaCl . The seeds were incubated in a growth chamber at 25±1°C in the dark for 6 days . Twenty-four hours later , the imbibed seeds were weighed to calculate the seed IR . Subsequently , the seeds were placed into new dishes with filter paper , and 5 mL of NaCl solution ( 0 or 150 mM ) was added . After the seeds were rinsed , the number of germinated seeds was counted to calculate the GI , germination potential and GR every day for the next 5 days . Soybean seeds were considered to be germinated when the radicle and plumule length of the soybean seed were greater than the seed length . Three replications were conducted in this study . The evaluated germination traits included IR [IR ( % ) = ( W2–W1 ) /W1×100% , where W1 represents the dry seed weight before imbibition and W2 represents the seed weight after imbibition for 24 h] , GI [GI = Σ ( Gt/Dt ) , where Gt is the accumulated number of germinated seeds on day t and Dt indicates the time corresponding to Gt in days] , GP [GP ( % ) = N3/N*100 , where N3 indicates the number of germinated seeds on day 3 and N represents the total number of experimental seeds] , and GR [GR ( % ) = Nt/N*100% , where Nt indicates the number of germinated seeds on day t and N represents the total number of experimental seeds] . The ST was defined as the ratio of the germination-related traits ( IR , GI , GR and GP ) under salt conditions to the same traits under salt-free conditions [21] . The mean values of all phenotypic data obtained for the RILs population in the four environments and for the natural population in the three environments were utilized for descriptive statistics and correlation analysis . ANOVAs for all traits were performed using SAS 9 . 0 software ( SAS Institute 1999 ) , and Pearson’s correlations between traits were assessed using SPSS 20 software ( SPSS Statistics 20 ) . The frequency histograms of the four salt tolerance indices were generated with Origin 8 . 0 software . A genetic linkage map was constructed from the 184 F7:11 lines of RILs using 221 SSR markers , three EST-SSRs and one R gene ( resistance to soybean mosaic virus ) [61] . The constructed linkage map covered 2 , 625 . 9 cM of the soybean genome with an average distance of 11 . 8 cM between markers . Composite interval mapping ( CIM ) was employed for QTL mapping with WinQTLCart version 2 . 5_011 ( http://statgen . ncsu . edu/qtlcart/ ) . The control marker number and window size were set to 5 and 10 cM , respectively . The forward and backward regression method was selected , and empirical thresholds were computed using permutation test analyses ( 1000 permutations , overall error level of 5% ) [63] . The QTLs considered to be significant were those with LOD peaks that exceeded the genome-wide threshold of 2 . 5 [64] . Confidence intervals were defined as the map interval corresponding to a 1-LOD decline on either side of the LOD peak . We employed 207 , 608 SNPs with MAF>5% acquired from the NJAU 355 K SoySNP array to genotype the 211 soybean accessions used in the GWAS performed in this study [25] . The LD decay rate , defined as the chromosomal distance where the LD decays to half of its maximum value , was 130 kb in the 211 cultivated soybeans [25] . The mean values of all phenotypic data from E1 , E2 , and E3 were used for the GWAS . The GWAS was conducted with an R package called Genome Association and Prediction Integrated Tool ( GAPIT ) [65] using a compressed mixed linear model ( CMLM ) and controlling for relatedness and population structure [66] . The threshold for a significant association was set to 1/n ( n is the number of markers , P≤4 . 82×10−6 or -log10 ( P ) ≥5 . 32 ) [67] . In addition , SNPs within the threshold of 4 . 5≤-log10 ( P ) <5 . 32 were defined as suggestive SNPs [64] . Whole-genome sequencing was performed on the two parents of the 184 RILs population , Kefeng No . 1 and Nannong 1138–2 , using the Illumina HiSeq 4000 sequencing platform . The genome of the cultivated soybean Williams 82 was used as a reference sequence . Genome Analysis Toolkit ( GATK ) was used for SNP calling to genotype these two soybean accessions . After obtaining nucleotide polymorphism information from Kefeng No . 1 and Nannong 1138–2 , we screened the SNPs to obtain candidate genes for salt tolerance . A gene model was considered a candidate gene for salt tolerance if the gene satisfied the following conditions: ( 1 ) the SNP was located within a QTL that was found to be significantly associated with salt tolerance in this study and ( 2 ) the SNP was located within the coding region of the gene model and resulted in an amino acid exchange or was located in the promoter region of the gene model . The coding sequence of GmCDF1 was cloned from NJAU_204 , which is tolerant to salt stress , and subsequently subcloned into the vector pMDC83 ( containing double CaMV 35S promoter ) to produce the pMDC83-GmCDF1 overexpression vector . Specific primers were employed to amplify a 365-bp fragment from cDNA of NJAU_204 to be ligated into the vector pB7GWIWG2 ( II ) , and this ligation yielded the pBI-GmCDF1 RNAi vector . The primers used to construct these vectors are listed in S4 Table . pMDC83-GmCDF1 , pBI-GmCDF1-RNAi and their respective empty vectors were independently transformed into Agrobacterium rhizogenes strain K599 for hairy root transformation [28] . One-week-old seedlings were injected with transformed K599 and transferred to a temperature-controlled germination chamber with a 12-h light/12-h dark cycle , a day/night temperature of 28°C/25°C and high humidity . Approximately two to three weeks later , specifically when the hairy roots were approximately 2–10 cm near the infection site where the hairy roots formed , the primary root was cut off . The hairy roots of the seedlings were immersed in 1/2 Hoagland nutrient solution for five days and then treated with water or 75 mM NaCl for four days . In addition , the chlorophyll concentrations of the top secondary fully expanded leaves were measured three times using a chlorophyll meter ( Konica Minolta SPAD502 ) and are expressed as SPAD values , and the fresh weights of the hairy roots from all the transgenic line were noted prior to PCR confirmation . Subsequently , the soybean hairy roots and shoots were harvested separately and used for gene expression or ICP-OES analysis . Negative soybean hairy roots and their respective shoots were discarded as soon as their phenotypic data were collected , and the samples for ICP-OES analysis were rinsed three times with deionized water . To analyze the expression of candidate genes , seeds of 31 soybean accessions including the two parents ( Kefeng No . 1 and Nannong 1138–2 ) of the RILs population were sterilized with a chlorine gas atmosphere to minimize the danger of microbial contamination and infection and treated with 0 or 150 mM NaCl as in the above-described seed germination experiment . After treatment with or without salt stress for 48 h , 15 quarters of the embryos from three replications were sampled and stored at -80°C for the isolation of total RNA . To analyze the expression of GmCDF1 in different soybean tissues , we sampled different tissues from the roots , stems , leaves , and flowers during the full-blossom period , pod walls on the 15th day after flowering ( DAF ) , and seeds at 15 DAF . Total RNA was isolated using the RNAsimple Total RNA Kit ( TIANGEN Beijing , China ) , and first-strand cDNA was reverse-transcribed using a TaKaRa Primer Script RT reagent kit with gDNA Eraser . Gene expression was determined by RT-PCR using an ABI 7500 system ( Applied Biosystems , Foster City , CA , USA ) with the SYBR Green Real-time Master Mix ( Toyobo ) , and the data were analyzed using ABI 7500 Sequence Detection System ( SDS ) software version 1 . 4 . 0 . The normalized expression was calculated for each sample as ΔΔCT = ( CT , Target-CT , Tubulin ) genotype- ( CT , Target-CT , Tubulin ) calibrator , and the fold change was calculated as 2-ΔΔCT [68] . The transcript level of tubulin ( GenBank accession number: AY907703 ) was used as a quantitative control . The primers used in the present study are listed in S2 Table . Samples of hairy roots and shoots were dried at 105°C for 60 min and then dried at 65°C for 72 h in a forage dryer . Then , 50–100 mg of each dry sample was weighed , and 2 mL of HNO3 and 8 mL of deionized water were added . The mixtures were digested at 200°C for 10 min using an Ethos Microwave Digestion Labstation ( Milestone Rrl . , Sorisole , Italy ) . After cooling , the digested samples were diluted to 50 mL with distilled water . The contents of Na+ , Ca2+ , K+ Zn2+ and Mn2+ were then detected using an Optima 8000 DV Inductively Coupled Plasma Optical Emission Spectrometer ( ICP-OES ) system ( PerkinElmer Inc . , Waltham , MA , USA ) . The genome sequence of GmCDF1 from 31 soybean genotypes ( S5 Table ) was amplified using the specific primers described in S2 Table . The sequences were assembled and aligned using ContigExpress in Vector NTI Advance 10 ( Invitrogen , Carlsbad , CA , USA ) and MEGA version 6 [69] , respectively . Polymorphisms with MAF>0 . 05 , including SNPs and indels , were identified among these genotypes , and their association with salt tolerance indices was calculated with Tassel 5 . 0 [70] . The analysis of GmCDF1 haplotypes and the pairwise LD analysis were performed with Haploview 4 . 2 [71] . Markers were defined as being significantly associated with the phenotype based on the significant association thresholds of -log10 ( P ) >1 . 30 and P<0 . 05 .
|
Soil salinity can seriously threaten soybean growth and development and seed germination is a key phase in the soybean growth cycle . Thus , understanding the genetic mechanism of salt tolerance at the germination stage is very important for improving the salt tolerance of soybean at the germination stage . An analysis combining linkage mapping , GWAS and whole-genome sequencing of two soybean accessions revealed that the GmCDF1 was associated with soybean salt tolerance . The transformation of soybean hairy roots and an inductively coupled plasma optical emission spectrometry ( ICP-OES ) analysis confirmed that GmCDF1 negatively regulates salt tolerance by maintaining ion homeostasis in soybean . A haplotype analysis found ten haplotypes of GmCDF1 in 31 soybean accessions , and a candidate gene association analysis identified two probable causative GmCDF1 polymorphisms . Moreover , higher GmCDF1 expression resulted in salt sensitivity . Our results not only reveal the role of GmCDF1 in soybean adaptation to salt stress but also help elucidate the genetic molecular mechanisms of salt tolerance and facilitate the implementation of marker-assisted selection for salt tolerance in soybean breeding programs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"genome-wide",
"association",
"studies",
"plant",
"anatomy",
"ecology",
"and",
"environmental",
"sciences",
"medicine",
"and",
"health",
"sciences",
"integumentary",
"system",
"quantitative",
"trait",
"loci",
"plant",
"physiology",
"genetic",
"mapping",
"plant",
"science",
"genome",
"analysis",
"crops",
"plant",
"pathology",
"plant",
"ecology",
"hair",
"genomics",
"crop",
"science",
"gene",
"expression",
"seeds",
"genetic",
"loci",
"agriculture",
"plant",
"defenses",
"haplotypes",
"soybean",
"plant",
"resistance",
"to",
"abiotic",
"stress",
"anatomy",
"ecology",
"heredity",
"genetics",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"plant-environment",
"interactions",
"human",
"genetics"
] |
2019
|
A cation diffusion facilitator, GmCDF1, negatively regulates salt tolerance in soybean
|
Scandinavia was one of the last geographic areas in Europe to become habitable for humans after the Last Glacial Maximum ( LGM ) . However , the routes and genetic composition of these postglacial migrants remain unclear . We sequenced the genomes , up to 57× coverage , of seven hunter-gatherers excavated across Scandinavia and dated from 9 , 500–6 , 000 years before present ( BP ) . Surprisingly , among the Scandinavian Mesolithic individuals , the genetic data display an east–west genetic gradient that opposes the pattern seen in other parts of Mesolithic Europe . Our results suggest two different early postglacial migrations into Scandinavia: initially from the south , and later , from the northeast . The latter followed the ice-free Norwegian north Atlantic coast , along which novel and advanced pressure-blade stone-tool techniques may have spread . These two groups met and mixed in Scandinavia , creating a genetically diverse population , which shows patterns of genetic adaptation to high latitude environments . These potential adaptations include high frequencies of low pigmentation variants and a gene region associated with physical performance , which shows strong continuity into modern-day northern Europeans .
As the ice sheet retracted from northern Europe after the Last Glacial Maximum ( LGM ) , around 23 , 000 years ago , new habitable areas emerged [1] , allowing plants [2 , 3] and animals [4 , 5] to recolonize the Scandinavian peninsula ( hereafter referred to as Scandinavia ) . There is consistent evidence of human presence in the archaeological record from approximately 11 , 700 years before present ( BP ) both in southern and northern Scandinavia [6–9] . At this time , the ice sheet was still dominating the interior of Scandinavia [9 , 10] ( Fig 1A , S1 Text ) , but recent climate modeling shows that the Arctic coast of ( modern-day ) northern Norway was ice free [10] . Similarities in late-glacial lithic technology ( direct blade percussion technique ) of Western Europe and the oldest counterparts of Scandinavia appearing around 11 , 000 calibrated ( cal ) BP [11] ( S1 Text ) have been used to argue for an early postglacial migration from southwestern Europe into Scandinavia , including areas of northern Norway . However , studies of another lithic technology , the “pressure blade” technique , which first occurred in the northern parts of Scandinavia around 10 , 200 cal BP , indicates contact with groups in the east and possibly an eastern origin of the early settlers [7 , 12–15] ( S1 Text ) . The first genetic studies of Mesolithic human remains from central and eastern Scandinavian hunter-gatherers ( SHGs ) revealed similarities to two different Mesolithic European populations , the “western hunter-gatherers” ( WHGs ) from western , central , and southern Europe and the “eastern hunter-gatherers” ( EHGs ) from northeastern and eastern Europe [16–24] . Archaeology , climate modeling , and genetics suggest several possibilities for the early postglacial migrations into Scandinavia , including migrations from the south , southeast , northeast , and combinations of these; however , the early postglacial peopling of Scandinavia remains elusive [1 , 4 , 6–19 , 25 , 26] . In this study , we contrast genome sequence data and stable isotopes from Mesolithic human remains from western , northern , and eastern Scandinavia to infer the early postglacial migration routes into Scandinavia—from where people came , what routes they followed , how they were related to other Mesolithic Europeans [17–21 , 27]—and to investigate human adaptation to high-latitude environments .
In order to compare the genomic sequence data of the seven SHGs to genetic information from other ancient individuals and modern-day groups , data were merged with shotgun sequence data and SNP capture data from six published Mesolithic individuals from Motala in central Scandinavia , and 47 published Stone Age ( Upper Paleolithic , Mesolithic , and Early Neolithic ) individuals from other parts of Eurasia ( S6 Text ) [17–22 , 26 , 27 , 29 , 35–38] , as well as with a world-wide set of 203 modern-day populations [18 , 32 , 39] . All 13 SHGs—regardless of geographic sampling location and age—display genetic affinities to both WHGs and EHGs ( Fig 1A and 1B , S6 Text ) . One individual , SF11 , seems to be a slight genetic outlier in the principal component analysis ( PCA ) , which could be due to the lower coverage or driven by nuclear contamination ( Table 1 , S6 Text ) . Generally , the pattern of dual ancestry is consistent with a scenario in which SHGs represent a mixed group tracing parts of their ancestry to both the WHGs and the EHGs [17–19 , 22 , 24 , 40] . The SHGs from northern and western Scandinavia show a distinct and significantly stronger affinity to the EHGs compared to the central and eastern SHGs ( Fig 1 ) . Conversely , the SHGs from eastern and central Scandinavia were genetically more similar to WHGs compared to the northern and western SHGs ( Fig 1 ) . Using qpAdm [19] , the EHG genetic component of northern and western SHGs was estimated to 48 . 9% ( ± 5% ) and differs from the 37 . 8% ( ± 3 . 2% ) observed in eastern and south-central SHGs . The latter estimate is similar to ancestry estimates obtained for eastern Baltic hunter-gatherers from Latvia [29] ( 33 . 7% ± 4 . 7% , Fig 1A ) . Although the difference in ancestry estimates between northern and western SHG , and eastern and south-central SHG is only marginally significant ( Z = 1 . 87 , p = 0 . 062 ) , this pattern is in agreement with other analyses such as ADMIXTURE and TreeMix ( S6 Text ) . Furthermore , the direct comparison using D statistics with Chimpanzee ( Chimp ) as an outgroup ( D ( Chimp , WHG; eastern or south-central SHG , northern or western SHG ) < 0 , Z = −5 . 14 and D ( Chimp , EHG; eastern or south-central SHG , northern or western SHG ) > 0 , Z = 1 . 72 ) show that WHG are genetically closer to eastern and south-central SHG , whereas EHG tend to share more alleles with northern and western SHGs ( S2 Fig ) . These patterns of genetic affinity within SHGs are in direct contrast to the expectation based on geographic proximity with EHGs and WHGs . From about 11 , 700 cal BP , consistent archaeological evidence of human presence exists in southern Scandinavia following the retreat of the ice sheet [6 , 41 , 42] ( S1 Text ) . Artifacts and tools found at these sites show similarities with the Ahrensburgian tradition of northern central Europe [15 , 43] , suggesting that these hunter-gatherers likely had a southern origin from a WHG-like gene pool as no EHG ancestry has been found in central and western Europe [18 , 21 , 24 , 27] . Although this genetic component would have entered from today’s northern Germany and Denmark ( Fig 2 , Scenario a ) , it remains unclear how and where the EHG component entered Scandinavia ( Fig 2 , Scenarios b , c and/or d ) . The EHG-related migration likely took place after the migration of WHGs from the south as the earliest eastern-associated pressure blade finds postdate the southwestern-associated direct blade finds in Scandinavia ( S1 Text ) . Two migrations with admixture at different time-periods would generate a genetic gradient with the highest contribution of a source close to its geographic region of entry . The observed genetic pattern is consistent with a migration of the EHGs from the northeast moving southwards along the ice-free Norwegian Atlantic coast where the two groups started mixing ( Fig 2 , Scenarios a and b ) , which would cause more EHG ancestry in western SHGs . If the EHG migration had crossed the Baltic Sea into Scandinavia , where it would meet and mix with a WHG-like population ( Fig 2 , combination of Scenarios a and c ) , a gradient with most EHG ancestry in eastern SHGs would have been created—exactly opposite to the observed pattern . A similar pattern would be expected if the EHG migration went around the Baltic Sea along current day’s Finnish west coast and down via today’s Swedish east coast ( scenario not depicted in Fig 2 ) . An EHG migration along the southern Baltic coast ( Fig 2 , Scenarios a and d ) should cause a related pattern to a crossing of the Baltic Sea with more EHG ancestry in central and eastern SHGs . Furthermore , such a scenario would likely also make the Latvian Mesolithic hunter-gatherers the group with most EHG ancestry , which is in stark contrast to the empirical data in which the Latvian group shows the lowest proportion of EHG ancestry ( 33 . 7% ± 4 . 7% ) , and also not consistent with chronology , as the dated settlements east of the Baltic Sea are younger than the early settlements in Scandinavia ( S1 Text ) . Thus , the only scenario consistent with both genetic and archaeological data is a migration of a WHG-related group migrating into Scandinavia from the south , followed by an EHG-related group migrating to Scandinavia from the northeast along the Norwegian Atlantic coast . Notably , such a migration along the Norwegian coast could have been facilitated by the use of the more specialized pressure blade technique ( S1 Text ) [12 , 14] . The individuals sequenced here postdate these migrations , but a genetic east-west gradient would be maintained over time in Scandinavia and only additional large-scale migrations from different sources would alter this pattern . This observation is important as the geographic pattern still holds without the chronologically much younger Steigen individual , which might represent local continuity or later migrations into north-western Scandinavia from the east . Interestingly , stable nitrogen and carbon isotope analysis of northern and western SHGs revealed an extreme marine diet , suggesting a pronounced maritime subsistence , in contrast to the more mixed terrestrial and aquatic diet of eastern and central SHGs ( S1 Text ) . Mobility is difficult to trace based solely on carbon and nitrogen isotope data; however , the patterns are consistent with a migration along the Norwegian Atlantic coast relying on local resources . By sequencing complete ancient genomes , we can compute unbiased estimates of genetic diversity , which are informative of past population sizes and population history . Here , we restrict the analysis to WHGs and SHGs because only SNP capture data are available for EHGs ( S7 Text ) . In modern-day Europe , there is greater genetic diversity in the south compared to the north . During the Mesolithic period , by contrast , we find lower levels of runs of homozygosity ( RoH ) ( Fig 3A ) and linkage disequilibrium ( LD ) ( Fig 3B ) in SHGs compared to WHGs ( represented by Loschbour and Bichon [18 , 35] ) . By using a multiple sequentially Markovian coalescent ( MSMC ) approach [44] for the high-coverage , high-quality genome of SF12 , we find that right before the SF12 individual lived , the effective population size of SHGs was similar to that of WHGs ( Fig 3C ) . At the time of the LGM and back to approximately 50 , 000 years ago , both the WHGs and SHGs go through a bottleneck , but the ancestors of SHGs retained a greater effective population size in contrast to the ancestors of WHGs who went through a more severe bottleneck ( Fig 2C ) , which is consistent across 100 bootstrap replicates ( S2 Fig ) . These differences in effective population size estimates may be attributed to the admixture in SHGs as migration events can have delayed effects on estimates of effective population size over time [45] . Around 50 , 000–70 , 000 years ago , the effective population sizes of the ancestors of SHGs , WHGs , Neolithic groups ( represented by Stuttgart [18] ) , and Paleolithic Eurasians ( represented by Ust-Ishim [38] ) align , suggesting that these diverse groups all trace their ancestry back to a common ancestral group , which likely represents the early migrants out of Africa . With the aim of detecting signs of adaptation to high-latitude environments and selection during and after the Mesolithic period , we employed two different approaches that utilize the Mesolithic genomic data . In the first approach , we assumed that SHGs adapted to high-latitude environments of low temperatures and seasonally low levels of light , and searched for gene variants that carried over to modern-day people in northern Europe . Modern-day northern Europeans trace limited amounts of genetic material back to the SHGs ( due to the many additional migrations during later periods ) , and any genomic region that displays extraordinary genetic continuity would be a strong candidate for adaptation in people living in northern Europe across time . We designed a statistic , Dsel ( S9 Text ) , that captures this specific signal and scanned the whole genome for gene variants that show strong continuity ( little differentiation ) between SHGs and modern-day northern Europeans while exhibiting large differentiation to modern-day southern European populations [46] ( Fig 4A; S9 Text ) . Six of the top 10 SNPs with greatest Dsel values were located in the TMEM131 gene that has been found to be associated with physical performance [47] , which could make it part of the physiological adaptation to cold [48] . This genomic region was more than 200 kbp ( kilo base pairs ) long and showed the strongest haplotypic differentiation between modern-day Tuscan individuals ( TSIs ) and modern-day Finnish individuals ( FINs ) across the genome ( S9 Text ) . The particular haplotype was relatively common in SHGs , it is even more common among today’s Finnish population ( S9 Text ) and showed a strong signal of local adaptation ( S9 Text ) . Other top hits included genes associated with a wide range of metabolic , cardiovascular , and developmental and psychological traits ( S9 Text ) potentially linked to physiological adaptation to cold environments [48] . In addition to performing this genome-wide scan , we studied the allele frequencies in three pigmentation genes ( SLC24A5 , SLC45A2 , which have a strong effect on skin pigmentation , and OCA2/HERC2 , which has a strong effect on eye pigmentation ) in which the derived alleles are virtually fixed in northern Europeans today . The differences in allele frequencies of those three loci are among the highest between human populations , suggesting that selection was driving the differences in eye color , skin , and hair pigmentation as part of the adaptation to different environments [50–53] . All of the depigmentation variants at these three genes are in high frequency in SHGs in contrast to both WHGs and EHGs ( Fig 4B ) . We conduct neutral simulations of the allele frequencies in an admixed SHG population to estimate p-values for observing these allele frequencies without selection ( S9 Text ) . The p-values for all three SNPs are lower than 0 . 2; the combined p-value [54] for all three pigmentation SNPs is 0 . 028 . Therefore , the unique configuration of the SHGs is not fully explained by the fact that SHGs are a mixture of EHGs and WHGs , but could rather be explained by a continued increase of the allele frequencies after the admixture event , likely caused by adaptation to high-latitude environments [50 , 52] . By combining information from climate modeling , archaeology , and Mesolithic human genomes , we were able to reveal the complexity of the early migration patterns into Scandinavia and human adaptation to high-latitude environments . We disentangled two migration routes and linked them to particular archaeological patterns . We also demonstrated greater genetic diversity in Mesolithic northern Europe compared to southern and central Europe—in contrast to modern-day patterns—and showed that many genetic variants that were common during the Mesolithic period have been lost today . These findings reiterate the importance of human migration for dispersal of novel technology in human prehistory [14–20 , 27 , 40 , 55–58] .
Genomic sequence data were generated from teeth and bone samples belonging to seven ( eight , including SF13 ) Mesolithic SHGs ( S1 Text ) . A detailed description on the archaeological background of the samples as well as post-LGM Scandinavia can be found in S1 Text . Additional libraries were sequenced for two previously published Neolithic hunter-gatherers , Ajvide58 and Ajvide70 [17] ( S2 Text ) . All samples were prepared in the dedicated aDNA facilities at Uppsala University ( SF9 , SF11 , SF12 , SF13 , SBj , Hum1 , Hum2 , Ajvide58 , Ajvide70 ) and at Stockholm University ( Steigen ) . Bones and teeth were decontaminated prior to analysis by wiping them with a 1% Sodiumhypoclorite solution and DNA-free water . Furthermore , all surfaces were UV irradiated ( 6 J/cm2 at 254 nm ) . After removing 1 millimeter of the surface , approximately 30–300 mg of bone was powderized and DNA was extracted following silica-based methods as in [59] with modifications as in [57 , 60] or as in [61] and eluted in 25–110 μl of EB buffer . Between one and 16 extractions were made from each sample and one extraction blank with water instead of bone powder was included per six to 10 extracts . Blanks were carried along the whole process until quantitative PCR ( qPCR ) and/or PCR and subsequent quantification . DNA libraries were prepared using 20 μl of extract , with blunt-end ligation coupled with P5 and P7 adapters and indexes as described in [57 , 62] . From each extract one to five double stranded libraries were built . Because aDNA is already fragmented , the shearing step was omitted from the protocol . Library blank controls , including water as well as extraction blanks , were carried along during every step of library preparation . In order to determine the optimal number of PCR cycles for library amplification , qPCR was performed . Each reaction was prepared in a total volume of 25 μl , containing 1 μl of DNA library , 1X MaximaSYBRGreen mastermix , and 200 nM each of IS7 and IS8 [62] reactions were set up in duplicates . Each blunt-end library was amplified in four to 12 replicates with one negative PCR control per index-PCR . The amplification reactions had a total volume of 25 μl , with 3 μl DNA library , and the following in final concentrations: 1 X AmpliTaq Gold Buffer , 2 . 5 mM MgCl2 , 250 μM of each dNTP , 2 . 5 U AmpliTaq Gold ( Thermo Fisher Scientific , Waltham , MA ) , and 200 nM each of the IS4 primer and index primer [62] . PCR was done with the following conditions: an activation step at 94°C for 10 min followed by 10–16 cycles of 94°C for 30 s , 60°C for 30 s , and 72°C for 30 s , and a final elongation step of 72°C for 10 min . For each library , four amplifications with the same indexing primer were pooled and purified with AMPure XP beads ( Agencourt; Beckman Coulter , Brea , CA ) . The quality and quantity of libraries was checked using Tapestation or BioAnalyzer using the High Sensitivity Kit ( Agilent Technologies , Cary , NC ) . None of the blanks showed any presence of DNA comparable to that of a sample and were therefore not further analyzed . For initial screening , 10–20 libraries were pooled at equimolar concentrations for sequencing on an Illumina HiSeq 2500 using v . 4 chemistry , and 125 bp paired-end reads or HiSeqX , 150 bp paired-end length using v2 . 5 chemistry at the SNP & SEQ Technology Platforms at Uppsala University and Stockholm University . After evaluation of factors such as clonality , proportion of human DNA , and genomic coverage samples were selected for resequencing , aiming to yield as high coverage as possible for each library . Based on the results of the non-damage-repair sequencing , the SF12 individual was selected for large-scale sequencing in order to generate a high-coverage genome of high quality where damages had been repaired using UDG . In addition to the 15 extracts previously prepared and used for non-damage repair libraries , another 111 extracts were made based on a variety of silica-based methods [27 , 57 , 59 , 60] . From these 126 extracts , a total of 258 damage-repaired double-stranded libraries were built for Illumina sequencing platforms . Libraries were built as above , except a DNA repair step in which UDG and endonuclease VIII or USER enzyme ( NEB ) treatment was included in order to remove deaminated cytosines [63] . qPCR was performed in order to quantify the number of molecules and the optimal number of PCR cycles prior to amplification for each DNA library . Furthermore , this step included extraction blanks , library blanks , and amplification blanks to monitor potential contamination . All of these negative controls showed an optimal cycle of amplification significantly higher to those of our aDNA libraries ( >10 cycles ) and they were thus deemed as negative . Our experimental results show minimal levels of contamination , which is in concordance with mt DNA and X chromosome estimates of contamination ( see S4 Text and Table 1 ) . Each reaction was done in a total volume of 25 μl , containing 1 μl of DNA library , 1 X MaximaSYBRGreen mastermix ( Thermo Fisher Scientific ) , and 200 nM each of IS7 and IS8 [62] , reactions were set up in duplicate . The PCRs were set up using a similar system as for the nondamage repair samples ( in quadruplicates that were pooled prior to cleanup of the PCR products ) , except for using AccuPrime DNA polymerase ( Thermo Fisher Scientific ) instead of AmpliTaqGold ( Thermo Fisher Scientific ) and the following PCR conditions: an activation step at 95°C for 2 min followed by 10–16 cycles of 95°C for 15 s , 60°C for 30 s , and 68°C for 1 min , and a final elongation step of 68°C for 5 min . Blank controls , including water as well as extraction blanks , were carried out during every step of library preparation . Amplified libraries were pooled , cleaned , quantified , and sequenced in the same manner as non-damage repaired libraries . A small proportion of the libraries ( n = 14 ) were also subjected to whole genome capture ( WGC ) using European MYbaits from MYcroarray , and following the manufacturers protocol as done in [64] . In order to sequence libraries to depletion , two to eight libraries were pooled together and sequenced until reaching a clonality of >50%; if sequencing was halted before reaching that clonality level , it was either because the library was classified as unproductive based on the genome coverage generated , or that the sequencing goal ( >55 × coverage ) was already reached and further sequencing was deemed unnecessary . Sequencing was performed as above . Paired-end reads were merged using MergeReadsFastQ_cc . py [65]; if an overlap of at least 11 base pairs was found , the base qualities were added together and any remaining adapters were trimmed . Merged reads were then mapped single-ended with bwa aln 0 . 7 . 13 [66] to the human reference genome ( build 36 and 37 ) using the following nondefault parameters: seeds disabled -l 16500 -n 0 . 01 -o 2 [17 , 18] . To remove PCR duplicates , reads with identical start and end positions were collapsed using a modified version , to ensure random choice of bases , of FilterUniqSAMCons_cc . py [65] . Reads with less than 10% mismatches to the human reference genome , reads longer than 35 base pairs , and reads with mapping quality higher than 30 were used to estimate contamination . The genetic data obtained from the two bone elements SF9 and SF13 showed extremely high similarities , which suggested that the two individuals were related . Using READ [67] , a tool to estimate kin-relationship from aDNA , SF9 and SF13 were classified as either identical twins or the same individual . Therefore , we merged the genetic data for both individuals and refer to the merged individual as SF9 throughout the genetic analysis . All data show damage patterns indicative of authentic aDNA ( S3 Text ) . Contamination was estimated using three different sources of data: ( 1 ) the mt genome [68] , ( 2 ) the X chromosome if the individual was male [69 , 70] , and ( 3 ) the autosomes [71] . The data mapping to the human genome can be considered largely endogenous , as the contamination estimates were low across all three methods ( S4 Text ) . Most population genomic analyses require a set of reference data for comparison . We compiled three different data sets from the literature and merged them with the data from ancient individuals ( S6 Text ) . The three reference SNP panels were as follows: These data sets were merged with ancient individuals of less than 15× genome coverage using the following approach: for each SNP site , a random read covering that site with minimum mapping quality 30 was drawn ( using samtools 0 . 1 . 19 mpileup [73] ) and its allele was assumed to be homozygous in the ancient individual . Transition sites were coded as missing data for individuals that were not UDG-treated , and SNPs showing additional alleles or indels in the ancient individuals were excluded from the data . The six high-coverage ancient individuals ( SF12 , NE1 [27] , Kotias [35] , Loschbour [18] , Stuttgart [18] , and Ust-Ishim [38] ) used in this study were treated differently , as we generated diploid genotype calls for them . First , the base qualities of all Ts in the first five base pairs of each read as well as all As in the last five base pairs were set to 2 . We then used Picard [74] to add read groups to the files . Indel realignment was conducted with GATK 3 . 5 . 0 [66] using indels identified in phase 1 of the 1000 Genomes Project as reference [32] . Finally , GATK’s UnifiedGenotyper was used to call diploid genotypes with the parameters -stand_call_conf 50 . 0 , -stand_emit_conf 50 . 0 , -mbq 30 , -contamination 0 . 02 , and—output_mode EMIT_ALL_SITES using dbSNP version 142 as known SNPs . SNP sites from the reference data sets were extracted from the VCF files using vcftools [72] if they were not marked as low-quality calls . Plink 1 . 9 [75 , 76] was used to merge the different data sets . We performed PCA to characterize the genetic affinities of the ancient Scandinavian genomes to previously published ancient and modern genetic data . PCA was conducted on 42 present-day west Eurasian populations from the Human Origins data set [18 , 39] using smartpca [77] with numoutlieriter: 0 and lsqproject: YES options . A total of 59 ancient genomes ( 52 previously published and 7 reported here ) ( S6 Text ) were projected into the reference PCA space and computed from the genotypes of modern individuals . For all individuals , a single allele was selected randomly—making the data set fully homozygous . The result was plotted using the ploteig program of EIGENSOFT [77] with the–x and–k options . popstats [78] was used to calculate D statistics to test deviations from a tree-like population topology of the shape ( ( A , B ) ; ( X , Y ) ) [39] . Standard errors were calculated using a weighted block jackknife of 0 . 5 Mbp . The tree topologies are balanced at zero , indicating no recent interactions between the test populations . Significant deviations from zero indicate a deviation from the proposed tree topology depending on the value . Positive values indicate an excess of shared alleles between A and X or B and Y , whereas negative values indicate more shared alleles between B and X or A and Y . Using an outgroup as population A limits the test results to depend on the recent relationships between B and Y ( if positive ) or B and X ( if negative ) . Here , we used high-coverage Mota [37] , Yoruba [32] , and Chimp genome as ( A ) outgroups . popstats [78] was used to calculate f4 statistics in order to estimate shared drift between groups . Standard errors and Z scores for f4 statistics were estimated using a weighted block jackknife ( Fig 1C ) . A model-based clustering algorithm , implemented in the ADMIXTURE software [79] , was used to estimate ancestry components and to cluster individuals . ADMIXTURE was conducted on the Human Origins data set [18 , 39] , which was merged with the ancient individuals as described above . Data was pseudo-haploidized by randomly selecting one allele at each heterozygous site of present-day individuals . Finally , the data set was filtered for LD using PLINK [75 , 76] with parameters ( --indep-pairwise 200 25 0 . 4 ) , this retained 289 , 504 SNPs . ADMIXTURE was run in 50 replicates with different random seeds for ancestral clusters from K = 2 to K = 20 . Common signals between independent runs for each K were identified using the LargeKGreedy algorithm of CLUMPP [80] . Clustering was visualized using rworldmap , ggplot2 , SDMTools , and RColorBrewer packages of GNU R version 3 . 3 . 0 . Starting from K = 3 , when the modern samples split up into an African and eastern and western Eurasian clusters , the Mesolithic Scandinavians from Norway show slightly higher proportions of the Eastern cluster than Swedish Mesolithic individuals . This pattern continues to develop across higher values of K and it is consistent with the higher Eastern affinities of the Norwegian samples seen in the PCA and D- and f4 statistics . The results for all Ks are shown in S1 Fig . In addition to ADMIXTURE , we assessed the admixture patterns in Mesolithic Scandinavians using a set of methods implemented in ADMIXTOOLS [39] , qpWave [81] , and qpAdm [19] . Both methods are based on f4 statistics , which relate a set of test populations to a set of outgroups in different distances from the potential source populations . We used the following set of outgroup populations from the Human Origins data set: Ami_Coriell , Biaka , Bougainville , Chukchi , Eskimo_Naukan , Han , Karitiana , Kharia , and Onge . We first used qpWave to test the number of source populations for Mesolithic west Eurasians ( WHG ) . qpWave calculates a set of statistics X ( u , v ) = f4 ( u0 , u; v0 , v ) where u0 and v0 are populations from the sets of test populations L and outgroups R , respectively . To avoid having more test populations than outgroups , we built four groups consisting of ( 1 ) genetically western and central hunter-gatherers ( Bichon , Loschbour , KO1 , LaBrana ) , ( 2 ) EHGs ( UzOO74/I0061 , SVP44/I0124 , UzOO40/I0211 ) , ( 3 ) Norwegian hunter-gathers ( Hum1 , Hum2 , Steigen ) , and ( 4 ) Swedish hunter-gatherers ( individuals from Motala and Mesolithic Gotland ) . qpWave tests the rank of the matrix of all X ( u , v ) statistics . If the matrix has rank m , the test populations can be assumed to be related to at least m + 1 “waves” of ancestry , which are differently related to the outgroups . A rank of 0 is rejected in our case ( p = 3 . 13e-81 ) , whereas a rank of 1 is consistent with the data ( p = 0 . 699 ) . Haak et al . [19] already showed , using the same approach , that WHG and EHG descend from at least two sources ( confirmed with our data as rank 0 is rejected with p = 1 . 66e-86 , whereas rank 1 is consistent with the data ) , and adding individuals from Motala does not change these observations . Therefore , we conclude that European Mesolithic populations , including Swedish and Norwegian Mesolithic individuals , have at least two source populations . We then used qpAdm to model Mesolithic Scandinavian individuals as a 2-way admixture of WHG and EHG . qpAdm was run separately for each Scandinavian individual x , setting T = x as target and S = ( EHG , WHG ) as sources . The general approach of qpAdm is related to qpWave: target and source are used as L ( with T being the base population ) , and f4 statistics with outgroups from R ( same as above ) are calculated . The rank of the resulting matrix is then set to the number of sources minus one , which allows to estimate the admixture contributions from each population in S to T . The results are shown in Fig 1 . We calculate a Z score for the difference between Norwegian and Swedish SHG as Z=anor-asweSEnor2+SEswe2 where a are the ancestry estimates and SE are the respective block-jackknife estimates of the standard errors . Heterozygosity is a measurement for general population diversity and its effective population size . Analyzing the extent of homozygous segments across the genome can also give us a temporal perspective on the effective population sizes . Many short segments of homozygous SNPs can be connected to historically small population sizes , whereas an excess of long RoH suggests recent inbreeding . We restricted this analysis to the six high-coverage individuals ( SF12 , NE1 , Kotias , Loschbour , Stuttgart , Ust-Ishim ) for which we obtained diploid genotype calls and we compared them to modern individuals from the 1000 Genomes Project . The length and number of RoH were estimated using Plink 1 . 9 [75 , 76] and the parameters--homozyg-density 50 , --homozyg-gap 100 , --homozyg-kb 500 , --homozyg-snp 100 , --homozyg-window-het 1 , --homozyg-window-snp 100 , --homozyg-window-threshold 0 . 05 , and--homozyg-window-missing 20 . The results are shown in Fig 3A . Similar to RoH , the decay of LD harbors information on the demographic history of a population . Long-distance LD can be caused by a low effective population size and past bottlenecks . Calculating LD for aDNA data is challenging , as the low amounts of authentic DNA usually just yields haploid allele calls with unknown phase . In order to estimate LD decay for ancient populations , we first combine two haploid ancient individuals to a pseudo-diploid individual ( similar to the approach chosen for conditional nucleotide diversity , S7 Text ) . Next , we bin SNP pairs by distance ( bin size 5 kb ) and then calculated the covariance of derived allele frequencies ( 0 , 0 . 5 , or 1 . 0 ) for each bin . This way , we do not need phase information to calculate LD decay because we do not consider multilocus haplotypes , which is similar to the approach taken by ROLLOFF [39 , 82] and ALDER [83] to date admixture events based on admixture LD decay . For Fig 3B , we used two modern 1000 Genomes Project populations to scale the LD per bin . The LD between two randomly chosen PELs ( modern-day Peruvian individuals ) was set to 1 and the LD between two randomly chosen TSIs was set to 0 . This approach is used to obtain a relative scale for the ancient populations , and we caution against a direct interpretation of the differences to modern populations because technical differences in the modern data ( e . g . , SNP calling or imputation ) may have substantial effects . We are using MSMC’s implementation of PSMC’ [44] to infer effective population sizes over time from single high-coverage genomes . We restrict this analysis to UDG-treated individuals ( SF12 , Loschbour , Stuttgart , Ust-Ishim ) as postmortem damage would cause an excess of false heterozygous transition sites . Input files were prepared using scripts provided with the release of MSMC ( https://github . com/stschiff/msmc-tools ) and MSMC was run with the nondefault parameters--fixedRecombination and -r 0 . 88 in order to set the ratio of recombination to mutation rate to a realistic level for humans . We also estimate effective population size for six high-coverage modern genomes [84] ( Fig 3C ) . We plot the effective population size assuming a mutation rate of 1 . 25x10e-8 and a generation time of 30 years . The curves for ancient individuals were shifted based on their average C14 date . Additionally , we used multihetsep_bootstrap . py to generate 100 bootstraps per individual . The results are shown in S4 Fig . We scanned the genomes for SNPs with similar allele frequencies in Mesolithic and modern-day northern Europeans and contrasted it to a modern-day population from southern latitudes . Pooling all Mesolithic Scandinavians together , we obtain an allele frequency estimate for SHGs , which is compared to FINs and TSIs from the 1000 Genomes Project [32] . We use the Finnish population as representatives of modern-day northern Europeans ( this sample contains the largest number of sequenced genomes from a northern European population ) . Tuscans are used as an alternative population , who also trace some ancestry to Mesolithic populations , but who do not trace their ancestry to groups that lived at northern latitudes in the last 7 , 000–9 , 000 years . Our approach is similar to PBS [85] and inspired by DAnc [46] . For each SNP , we calculated the statistic Dsel , comparing the allele frequencies between one ancestral and two modern populations: Dsel=|DAFSHG−DAFTSI|−|DAFSHG−DAFFIN| This scan was performed on all transversion SNPs extracted from the 1000 Genomes Project data . Only sites with a high-confidence ancestral allele in the human ancestor ( as used by the 1000 Genomes Project [32] ) and with coverage for at least six ancient Scandinavians were included in the computation . More information can be found in S9 Text .
|
The Scandinavian peninsula was the last part of Europe to be colonized after the Last Glacial Maximum . The migration routes , cultural networks , and the genetic makeup of the first Scandinavians remain elusive and several hypotheses exist based on archaeology , climate modeling , and genetics . By analyzing the genomes of early Scandinavian hunter-gatherers , we show that their migrations followed two routes: one from the south and another from the northeast along the ice-free Norwegian Atlantic coast . These groups met and mixed in Scandinavia , creating a population more diverse than contemporaneous central and western European hunter-gatherers . As northern Europe is associated with cold and low light conditions , we investigated genomic patterns of adaptation to these conditions and genes known to be involved in skin pigmentation . We demonstrate that Mesolithic Scandinavians had higher levels of light pigmentation variants compared to the respective source populations of the migrations , suggesting adaptation to low light levels and a surprising signal of genetic continuity in TMEM131 , a gene that may be involved in long-term adaptation to the cold .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Materials",
"and",
"methods"
] |
[
"geologic",
"time",
"population",
"genetics",
"geographical",
"locations",
"social",
"sciences",
"ethnicities",
"archaeology",
"effective",
"population",
"size",
"geology",
"genome",
"analysis",
"mesolithic",
"period",
"paleontology",
"paleogenetics",
"population",
"biology",
"europe",
"genomic",
"libraries",
"genomics",
"european",
"people",
"people",
"and",
"places",
"population",
"metrics",
"population",
"size",
"earth",
"sciences",
"genetics",
"norwegian",
"people",
"biology",
"and",
"life",
"sciences",
"population",
"groupings",
"computational",
"biology",
"evolutionary",
"biology",
"stone",
"age"
] |
2018
|
Population genomics of Mesolithic Scandinavia: Investigating early postglacial migration routes and high-latitude adaptation
|
Hepatitis C virus ( HCV ) is a major cause of end-stage liver diseases . With 3–4 million new HCV infections yearly , a vaccine is urgently needed . A better understanding of virus escape from neutralizing antibodies and their corresponding epitopes are important for this effort . However , for viral isolates with high antibody resistance , or antibodies with moderate potency , it remains challenging to induce escape mutations in vitro . Here , as proof-of-concept , we used antibody-sensitive HVR1-deleted ( ΔHVR1 ) viruses to generate escape mutants for a human monoclonal antibody , AR5A , targeting a rare cross-genotype conserved epitope . By analyzing the genotype 1a envelope proteins ( E1/E2 ) of recovered Core-NS2 recombinant H77/JFH1ΔHVR1 and performing reverse genetic studies we found that resistance to AR5A was caused by substitution L665W , also conferring resistance to the parental H77/JFH1 . The mutation did not induce viral fitness loss , but abrogated AR5A binding to HCV particles and intracellular E1/E2 complexes . Culturing J6/JFH1ΔHVR1 ( genotype 2a ) , for which fitness was decreased by L665W , with AR5A generated AR5A-resistant viruses with the substitutions I345V , L665S , and S680T , which we introduced into J6/JFH1 and J6/JFH1ΔHVR1 . I345V increased fitness but had no effect on AR5A resistance . L665S impaired fitness and decreased AR5A sensitivity , while S680T combined with L665S compensated for fitness loss and decreased AR5A sensitivity even further . Interestingly , S680T alone had no fitness effect but sensitized the virus to AR5A . Of note , H77/JFH1L665S was non-viable . The resistance mutations did not affect cell-to-cell spread or E1/E2 interactions . Finally , introducing L665W , identified in genotype 1 , into genotypes 2–6 parental and HVR1-deleted variants ( not available for genotype 4a ) we observed diverse effects on viral fitness and a universally pronounced reduction in AR5A sensitivity . Thus , we were able to take advantage of the neutralization-sensitive HVR1-deleted viruses to rapidly generate escape viruses aiding our understanding of the divergent escape pathways used by HCV to evade AR5A .
About 150 million people are chronically infected with hepatitis C virus ( HCV ) with an increased risk of developing end-stage liver diseases , including cirrhosis and hepatocellular carcinoma [1–3] . Until recently , the treatment against HCV consisted of interferon and ribavirin but its efficacy was limited by side effects and a low rate of sustained virological response [4] . Progress in understanding HCV virology and the development of in vitro experimental systems to study antivirals has resulted in new interferon-free therapies [5] . These new treatment regimens consist of combinations of direct-acting antivirals ( DAA ) with or without ribavirin with greatly improved response rates . However , the high number of occult infections and the high cost of DAAs limit access , and the treatment does not provide protection against viral re-infection [2 , 6] . Thus , the need for a prophylactic HCV vaccine remains high . The HCV genome encodes a single polyprotein that is processed into 3 structural proteins ( Core , and envelope proteins E1 and E2 ) , p7 and 6 nonstructural proteins ( NS2-NS5B ) . HCV is an enveloped single positive-strand RNA virus belonging to the Flaviviridae family , and it is divided into seven major genotypes based on sequence homology [7 , 8] . The HCV envelope glycoprotein complex E1/E2 , present on the surface of virus particles , plays a critical role in viral entry through interactions with cellular receptors such as CD81 [9] , scavenger receptor class B type I ( SR-BI ) [10] , the low-density lipoprotein receptor ( LDLr ) [11] and several late-stage host entry factors [12] . The E1/E2 complex is the target of neutralizing antibodies ( NAbs ) [13–17] . The presence of NAbs in the early phase of acute HCV infection has been associated with viral clearance [18 , 19] . Moreover , passive transfer of monoclonal and polyclonal antibodies has conferred protection against HCV infection in experimental animal models [20–25] . Efforts by several research groups have identified promising NAbs against HCV [26–33] . Using phage display libraries , human monoclonal antibodies ( HMAbs ) against conserved epitopes on E1 and E2 were isolated [27 , 33] . Epitope-masking allowed for isolation of antibodies against less frequently targeted epitopes [27] , and one of these HMAbs , AR5A , which recognizes antigenic region 5 , showed cross-genotype neutralization activity against cell cultured HCV [27] . Although the high mutation rate of HCV has been shown to enable the virus to escape neutralization [34–38] , the barrier of escape for HMAb AR5A has not been assessed . The optimal vaccine candidate should induce NAbs against cross-genotype conserved epitopes with a universally high barrier to resistance . Previously , antibody-specific resistance mutations have been induced by co-culturing HCV infected cells with NAbs or by direct neutralization of HCV virus followed by virus amplification of the non-neutralized quasispecies [39–43] . However , the variants studied has been limited to neutralization-sensitive isolates , such as JFH1 [40 , 41] or the recombinant virus HJ3-5 [43] , introducing a potential bias . Hypervariable regions of the structural glycoprotein E2 can influence neutralization sensitivity [44–46] . We have previously described increased susceptibility of HCV to antibody neutralization when the hypervariable region 1 ( HVR1 ) of E2 is deleted [44 , 47] . The susceptibility to antibody neutralization for viruses without HVR1 has been shown to be between 10–1000 fold higher than for the parental virus [44 , 48] . HVR1 is not essential for infectivity given that viruses with this region deleted remained infectious in vitro [44;46] and in vivo [44 , 49 , 50] . Thus , we anticipated that the HVR1-deleted viruses could be an important tool for the study of antibody resistance , particularly for mapping conserved residues independent of this highly variable region frequently used by HCV to escape antibodies . Here , we used highly sensitive HVR1-deleted variants to study AR5A resistance pathways for the two isolates H77 ( genotype 1a ) and J6 ( genotype 2a ) of medium or high antibody resistance , respectively; we also attempted escape studies for the original H77 and J6 viruses retaining HVR1 . We induced AR5A-specific escape by culturing the JFH1-based Core-NS2 recombinants H77/JFH1ΔHVR1 and J6/JFH1ΔHVR1 in the presence of AR5A and identified mutations responsible for escape . Importantly , the mutations were also found to confer resistance in parental viruses . In addition , we were able to induce low-level non-AR5A specific resistance in H77/JFH1 , but not J6/JFH1 , by culturing these original viruses with high doses of AR5A . Finally , we used the JFH1-based Core-NS2 recombinant viruses of genotypes 1–6 to study resistance against AR5A . The information generated in these studies will be important to understanding HCV antibody escape , and the novel methodology using HVR1-deleted viruses could prove useful for designing future studies with other important NAbs against highly conserved epitopes . Furthermore , the present study provides novel insights into different pathways of AR5A virus escape for divergent isolates or even genotypes of this heterogeneous virus .
On the assumption that the greatly increased sensitivity of HVR1-deleted HCV variants would make it easier to prevent spread when culturing the viruses in the presence of antibody , we attempted to generate neutralization escape mutants of HCV by culturing HVR1-deleted H77/JFH1 ( H77/JFH1ΔHVR1 ) in Huh7 . 5 cells treated with relatively low doses of AR5A . Huh7 . 5 cells were infected with this genotype 1a recombinant at a multiplicity of infection ( MOI ) of 0 . 001 . When the number of infected cells reached 10% , monitored by immuno-staining , cells were split into 4 wells and cultured either without antibody ( untreated ) or with 5 μg/ml of AR5A , which equals ~500 times the IC50 for H77/JFH1ΔHVR1 ( treatments I , II , and III , Fig 1A ) . Seven days post-treatment the number of infected cells in the untreated well reached >80% , however , the number of infected cells in the treated wells decreased until it reached less than 1% . The percent of infected cells in treatment I increased from day 19 and reached >90% at day 30 post-initiation of antibody treatment . Treatment II and III were followed until day 58 when the cultures were closed due to infection suppresion as evidenced by the absence of antigen-positive cells for more than two weeks . Culture supernatant of AR5A treated ( day 30 ) and untreated virus ( day 7 ) was used to generate 1st passage virus stocks by infecting naive Huh7 . 5 cells . These stock viruses were tested in a dose-response focus forming units ( FFU ) reduction assay for AR5A antibody sensitivity ( Fig 1B ) . The AR5A-treated virus was >10 , 000-fold more resistant against AR5A ( Untreated virus IC50 = 0 . 0035 μg/ml; Treated virus IC50 > 50 μg/ml ) , indicating that the virus had escaped the HMAb . By direct sanger sequencing of the structural genes E1 and E2 , we identified substitutions N417D ( nucleotide change A1590G ) and L665W ( T2335G ) in E2 ( Table 1; all positions in the manuscript are given relative to H77 reference strain , GenBank accession number AF009606 ) in the AR5A treated virus; the untreated virus did not have any envelope mutations . The amino acid position N417 corresponds to glycosylation site 1 of E2 . Thus , the resistant phenotype of the escape virus was likely due to one of these E2 substitutions . To determine which mutation conferred resistance against AR5A , mutations encoding N417D and L665W were introduced into H77/JFH1ΔHVR1 . Viruses H77/JFH1ΔHVR1 , H77/JFH1ΔHVR1/N417D and H77/JFH1ΔHVR1/L665W reached the peak of infection at day 4 post-transfection . Additionally , viruses harboring either N417D or L665W had titers similar to or slightly higher than the control virus H77/JFH1ΔHVR1 ( Fig 2A ) . We generated a 1st passage stock of all three viruses and their envelope gene sequences were analyzed . While we found no change for H77/JFH1ΔHVR1 and H77/JFH1HVR1/N417D , we observed the dominant E2 substitution T534A ( A1941G , Table 1 ) for H77/JFH1ΔHVR1/L665W . The neutralization sensitivity of these three viruses was compared using a dose-response FFU reduction assay for AR5A antibody sensitivity ( Fig 2B ) . H77/JFH1ΔHVR1/L665W ( IC50 = 49 μg/ml ) was highly resistant to antibody neutralization compared to H77/JFH1ΔHVR1/N417D ( IC50 = 0 . 012 μg/ml ) and H77/JFH1ΔHVR1 ( IC50 = 0 . 0035 μg/ml ) . To rule out that the substitution T534A , which appeared in the H77/JFH1ΔHVR1/L665W virus , was involved in resistance against AR5A , we used a highly adapted HVR1-deleted virus , H77/JFH1ΔHVR1/N417D/N532D for the investigation . We generated stocks of this virus with and without L665W . These virus stocks did not have any additional changes in the envelope genes and the virus harboring L665W was fully resistant to antibody neutralization using AR5A ( S1 Fig ) . Thus , we showed that E2 substitution L665W alone conferred resistance to the HVR1-deleted 1a virus H77/JFH1 against AR5A . To test if L665W conferred resistance against AR5A for the parental virus retaining HVR1 we constructed the recombinant H77/JFH1L665W . Following transfection of Huh7 . 5 cells with in vitro transcribed RNA , H77/JFH1 and H77/JFH1L665W both reached titers of about 3 . 5 Log10 FFU/ml at day 4 post-transfection ( Fig 3A ) . Virus supernatants were used to generate 1st passage stocks and their envelope sequences were confirmed . Using these virus stocks we compared the susceptibility to AR5A using a dose-response FFU reduction assay ( Fig 3B ) . H77/JFH1 was completely neutralized at the highest doses of antibody with an IC50 value of 1 . 9 μg/ml , whereas H77/JFH1L665W was neutralized less than 50% even at the highest dose used in the assay ( 250 μg/ml ) . Next , to test if the effects of L665W on H77/JFH1 sensitivity were specific for AR5A , we tested neutralization sensitivity against HMAbs AR3A and AR4A . These antibodies target conformational envelope epitopes that are distinct from the AR5A epitope [27 , 33] . The antibody susceptibility against AR4A was similar between the parental virus and H77/JFH1L665W ( Fig 3C ) . However , the susceptibility to AR3A was slightly higher for H77/JFH1L665W ( IC50 = 1 . 5 μg/ml ) as compared with the parental virus ( IC50 = 10 μg/ml ) ( Fig 3D ) , although this effect was not observed when we performed the analogous neutralization experiments for the HVR1-deleted virus H77/JFH1ΔHVR1/L665W ( S2 Fig ) . Thus , the mutation L665W conferred specific resistance against AR5A . To determine whether the decrease in neutralization sensitivity of H77/JFH1 harboring L665W was due to impaired binding of the antibody to the particle we performed immunoprecipitation of H77/JFH1 and H77/JFH1L665W using HMAbs AR3A , AR4A , and AR5A . We quantified bead-associated HCV RNA by RT-qPCR and found greatly decreased immunoprecipitation of H77/JFH1L665W by AR5A , whereas AR3A and AR4A immunoprecipitation was similar for the two viruses ( Fig 4A ) . To determine whether AR5A binding was generally decreased by L665W , even outside the context of a mature virion , we carried out immunostaining of cells infected with viruses H77/JFH1 and H77/JFH1L665W using AR5A and AR4A . We observed a strong AR5A-specific fluorescence signal for H77/JFH1 infected cells , but not for H77/JFH1L665W infected cells ( Fig 4B ) . In contrast , the antibody AR4A was able to recognize E2 for both viruses ( S3 Fig ) . Escape mutations in E2 have been shown to change protein structure and reduce binding to CD81 [51] . To investigate this for virus with L665W , we carried out dose-response entry blocking using antibodies against the receptors CD81 ( Fig 4C ) , SR-BI ( Fig 4D ) , and LDLr ( Fig 4E ) . H77/JFH1L665W was as susceptible to CD81 , SR-BI , and LDLr blocking as the unmodified virus . Furthermore , H77/JFH1 and H77/JFH1L665W virus entry were similarly reduced by incubation with soluble CD81 receptor prior to infection ( S4 Fig ) . As AR5A is cross-genotype-reactive we attempted to generate escape mutants for another HVR1-deleted virus , J6/JFH1ΔHVR1 ( genotype 2a ) . Similar to the approach for H77/JFH1ΔHVR1 we infected Huh7 . 5 cells with J6/JFH1ΔHVR1 at an MOI of 0 . 001 . After the infection reached 10% , cells were split into 4 wells and cultured without antibody or with antibody at concentrations of 5 , 10 , or 20 μg/ml of AR5A antibody ( Fig 5A ) . While the control infection reached >90% infected cells at day 5 post-treatment , cells treated with 5 , 10 , and 20 μg/ml of AR5A reached >90% of infection only at day 19 , 21 , and 23 , respectively . Sequences obtained from AR5A cultures treated with 5 and 10 μg/ml did not show any changes in the envelope proteins . However , the viral sequence from the supernatant harvested from the culture treated with 20 μg/ml displayed substitutions I345V ( A1374G ) in E1 , and L665S ( T2335C ) and S680T ( T2379A ) in E2 ( Table 1 ) . The antibody susceptibility of the 1st passage virus generated from the 20 μg/ml treatment was evaluated in a dose-response assay ( Fig 5B ) . Similarly to what we observed for treated H77/JFH1ΔHVR1 , the treated J6/JFH1ΔHVR1 virus ( IC50 >50 μg/ml ) was found to be >8000-fold less sensitive to AR5A compared to the parental J6/JFH1ΔHVR1 virus ( IC50 = 0 . 006 μg/ml ) . Next , we introduced mutations encoding I345V , L665S , and S680T alone or in combination into J6/JFH1ΔHVR1 . Viruses reached >90% of infection 4 days post-transfection . J6/JFH1ΔHVR1 harboring L665S or the combination L665S/I345V had greatly reduced fitness compared to the parental virus ( Fig 6A and Fig 6B ) . In contrast , viruses containing S680T displayed titers similar to the parental virus even in combination with L665S and I345V , suggesting S680T was compensating for the deleterious effects of L665S . We attempted to generate a first passage stock for every mutant virus however J6/JFH1ΔHVR1/L665S/I345V lost the L665S substitution . J6/JFH1ΔHVR1/L665S ( IC50 = 8 . 4 μg/ml ) was >800-fold more resistant to AR5A than the parental virus ( IC50 = 0 . 01 μg/ml ) ( Fig 6C ) , however it was still >5-fold more susceptible to neutralization than J6/JFH1ΔHVR1/L665S/S680T ( IC50 = 43 μg/ml ) and J6/JFH1ΔHVR1/L665S/I345V/S680T ( IC50 >50 μg/ml ) ( Fig 6D ) , which were both similarly sensitive as the treated polyclonal escape virus ( Fig 5B ) . J6/JFH1ΔHVR1/I345V had a neutralization profile similar to the parental virus ( Fig 6C ) . Remarkably , J6/JFH1ΔHVR1/S680T and J6/JFH1ΔHVR1/I345V/S680T were more susceptible to neutralization with AR5A than J6/JFH1ΔHVR1 . Substitutions I345V , L665S , and S680T affected viability of parental J6/JFH1 in a similar way as described above for J6/JFH1ΔHVR1 ( Fig 7A and Fig 7B ) . 1st passage virus stocks were prepared for all the viruses , and again J6/JFH1L665S/I345V lost the L665S mutation while J6/JFH1L665S had a partial reversion of L665S . All other mutant viruses were generated successfully with no additional envelope mutations . The virus J6/JFH1L665S did not appear resistant to AR5A , however , this could be due to the partial loss of the mutation . Viruses J6/JFH1L665S/S680T ( IC50 > 250μg/ml ) and J6/JFH1L665S/I345V/S680T ( IC50 > 250 μg/ml ) were >7-fold more resistant to antibody neutralization for AR5A than the parental virus ( IC50 = 36 μg/ml ) ( Fig 7D ) . Similar to what we observed for the HVR1-deleted virus , the antibody susceptibility increased when viruses had the mutation S680T without L665S ( Fig 7C and Fig 7D ) . Our results confirmed that the mutation L665S together with S680T conferred resistance to J6/JFH1 against AR5A . To test whether L665W ( for H77/JFH1ΔHVR1 ) or L665S and S680T ( for J6/JFH1ΔHVR1 ) were influencing cell-to-cell spread we transfected Huh7 . 5 cells with HVR1-deleted HCV recombinants with and without these substitutions and mixed the transfected cells with naive cells either in the absence or in the presence of high levels of NAb , AR3A ( to prevent cell-free spread ) . L665S decreased both cell-free and cell-to-cell spread ( in line with our observations on the negative effect of this mutation on virus fitness , Fig 6A ) whereas none of the other mutations had any effect on cell-to-cell spread ( S5 Fig ) , indicating that changes in cell-to-cell spread were not the reason for the resistance of viruses harboring substitutions at residues 665 and 680 . Substitutions L665W and L665S decreased susceptibility against AR5A neutralization for viruses H77/JFH1 and J6/JFH1 , respectively . To analyze the effects on viability and AR5A neutralization sensitivity of other amino acid substitutions we generated H77/JFH1 recombinants with serine ( S ) , alanine ( A ) , valine ( V ) , tyrosine ( Y ) , and threonine ( T ) at position 665 . Viruses with substitutions L665Y and L665V were less infectious and the remaining substitutions rendered the virus non-viable resulting in the absence of infectious particles at the four tested time points post-transfection ( Fig 8A ) . We attempted to generate 1st passage virus stocks of H77/JFH1 harboring either L665Y or L665V , however infection with these viruses did not result in efficient viral spread and infected cells were not detected 3 weeks post-infection . Our data suggests that a very limited number of isolate-specific substitutions are allowed at this conserved E2 position . To address whether the decreased viability of the L665 mutants was due to perturbation of the E1/E2 interaction we performed immunoprecipitation ( using the E2-conformational antibody , AR3A ) of native E1/E2 complexes from 293T cells transiently expressing E1E2 in cis . Western blots of the E2-enriched fractions showed comparable levels of both E2 and the co-immunoprecipitated E1 protein ( Fig 8B ) , strongly indicating that E1/E2 interaction was conserved for all L665 mutants . To evaluate whether substitutions at position 665 or 680 could be induced by culturing H77/JFH1 or J6/JFH1 in the presence of AR5A similar treatments as the ones described above were performed . Here , H77/JFH1 and J6/JFH1 were used to infect naive Huh7 . 5 cells at an MOI of 0 . 001 . Once the virus spread to 1% of the cells both were cultured either without antibody or at the AR5A dose for which escape mutations were identified for the HVR1-deleted counterparts ( 5 μg/ml and 20 μg/ml , respectively ) , as well as a high dose of 200 μg/ml . Not surprisingly , only the high dose resulted in delayed viral spread ( Fig 9A and 9B ) . Sequence analysis of the envelope protein encoding sequences from the treated cultures identified the unique coding substitution A349D ( C1387A ) for H77/JFH1 and R408K ( G1564A ) for J6/JFH1 in the samples treated at 200 μg/ml ( Table 1 ) . These viruses were passaged to generate virus stocks with sequence-confirmed envelope proteins harboring these substitutions and without additional coding changes in the envelope proteins . In dose-response neutralization experiments against AR5A it was clear that the mutation A349D conferred low level resistance to H77/JFH1 ( ~8-fold , Fig 9C ) , whereas the mutation R408K had no effect on neutralization sensitivity of J6/JFH1 ( Fig 9D ) . Interestingly , the mutation A349D also conferred low level resistance against AR3A ( ~12-fold , Fig 9E ) , but not AR4A ( Fig 9F ) . To test the effect of the mutations on virus viability we constructed the recombinants H77/JFH1A349D and J6/JFH1R408K . Following transfection of Huh7 . 5 cells with in vitro transcribed RNA and infectivity titration of transfection supernatants it was evident that the mutation A349D increased the infectivity titer of H77/JFH1 , whereas the mutation R408K had no effect on the infectivity of J6/JFH1 ( S6A and S6B Fig ) . Thus , our data suggests that it is very difficult to induce specific escape mutations for resistant or even moderately resistant HCV recombinants as the virus is able to spread even at high antibody doses without acquiring specific resistance mutations . Alignment of the E2 protein of the genotype 1a ( H77 ) , 2a ( J6 ) , 3a ( S52 ) , 4a ( ED43 ) , 5a ( SA13 ) , and 6a ( HK6a ) [52–54] revealed that the position L665 was conserved across these genotype 1–6 isolates , and indeed also universally conserved in the Los Alamos sequence database [55] ( Table 1 ) and the European HCV database [56] . Since AR5A has demonstrated cross-genotype neutralizing potential [27 , 57] , it was interesting to test whether changes at position L665 could confer AR5A-resistance in other HCV genotypes . Using the JFH1-based Core-NS2 HCV recombinants harboring envelope protein sequence from genotype 2a ( J6 ) , 3a ( S52 ) , 4a ( ED43 ) , 5a ( SA13 ) , and 6a ( HK6a ) [54;58–61] we introduced the mutations encoding L665W . The parental and the HVR1-deleted viruses with or without L665W were tested for all the genotypes except genotype 4 HVR1-deleted virus which was non-viable [44] . The effect of L665W on viral fitness varied greatly ( Fig 10 ) . While the parental and HVR1-deleted SA13/JFH1 viruses were virtually unaffected ( Fig 10D ) , L665W seemed to attenuate both parental and HVR1-deleted J6/JFH1 ( Fig 10A ) and HK6a/JFH1 ( Fig 10E ) , as well as the parental ED43/JFH1 ( Fig 10C ) . On the other hand , L665W only affected the viability of the parental S52/JFH1 but not the HVR1-deleted counterpart ( Fig 10B ) . We used supernatants from each transfection culture to infect Huh7 . 5 cells and generated 1st passage stocks of each virus . The virus J6/JFH1ΔHVR1/L665W developed the substitution I262L ( A1125C , Table 1 ) , the virus S52/JFH1L665W developed I355F ( A1404T , Table 1 ) and the virus S52/JFH1ΔHVR1/L665W developed A372V ( C1456T , Table 1 ) , Q454H/q ( A1703C/a , Table 1 ) and F580V ( T2079G , Table 1 ) . The remaining virus stocks did not acquire coding mutations in the envelope genes . Using these virus stocks we tested if the substitution L665W affected AR5A susceptibility across genotypes and found that it broadly decreased sensitivity to AR5A neutralization ( Fig 11 ) . The change in antibody susceptibility caused by L665W was higher for the HVR1-deleted viruses than for the parental viruses , most likely due to the generally increased neutralization sensitivity of the HVR1-deleted variants . Finally , we compared neutralization sensitivity against the HMAb AR3A ( S7 Fig ) and AR4A ( S8 Fig ) for the different genotype viruses harboring L665W . With a few exceptions the neutralization susceptibility was unaffected by the introduction of L665W . J6/JFH1ΔHVR1/L665W was 20-fold less susceptible to neutralization with AR3A; this was not the case for the parental virus J6/JFH1L665W ( S7A Fig ) . Virus ED43/JFH1L665W was more susceptible to neutralization with both AR3A ( S7C Fig ) and AR4A ( S8C Fig ) , in contrast both parental and HVR1-deleted SA13/JFH1 viruses were more resistant to antibody neutralization with AR3A ( S7D Fig ) and AR4A ( S8D Fig ) . The other viruses had unaltered neutralization susceptibility against AR3A and AR4A . Thus , we concluded that overall L665W in E2 specifically reduced sensitivity against AR5A across HCV genotypes .
We identified different pathways of neutralization escape from the HMAb AR5A using the highly antibody susceptible HVR1-deleted viruses H77/JFH1ΔHVR1 and J6/JFH1ΔHVR1 and the original viruses H77/JFH1 and J6/JFH1 . The mutation A349D , found by culturing H77/JFH1 with a high dose of AR5A , conferred broader low-level neutralization resistance for H77/JFH1 . Mutations encoding amino acid changes at the highly conserved position L665 , found by culturing neutralization sensitive HVR1-deleted H77/JFH1 and J6/JFH1 with at least 10-fold lower doses of AR5A , conferred specific high-level resistance against AR5A for both viruses , with and without HVR1 , whereas mutation S680T was required to compensate for fitness loss as well as further increasing AR5A-specific resistance for J6/JFH1 and J6/JFH1ΔHVR1 . In addition , we also observed several cases of adaptive mutations in the envelope proteins that likely also contributed to the ability of the virus to overcome AR5A neutralization . Finally , while we observed divergent , isolate-dependent , pathways to AR5A resistance we showed that substitutions at L665 could confer resistance across HCV genotypes 1–6 , highlighting the potential importance of this position in AR5A resistance . The development of an effective vaccine will be a cost-effective way to control HCV-associated liver disease worldwide [38;62;63] . In vitro studies of antibody resistance against HMAbs is important for HCV vaccine development to understand potential challenges with resistant viruses . Previous studies demonstrated an overlap between mutations found in resistant variants isolated in vitro and escape mutations identified from in vivo infection [39;40;42;43;64] . The relevance of in vitro escape studies has been shown for HMAb CBH-2 , for which resistant variants had developed mutations at position D431 [40] , also identified in a naturally occurring CBH-2 resistant variant [42] . Moreover , mutations at position N415 and N417 have been identified during in vitro escape studies against the antibody AP33 [39;43] and mutations at these same positions were also observed in resistant viruses found in liver transplant patients treated with HCV1 , a HMAb recognizing the same E2 antigenic site [64] . Thus , in vitro escape studies can provide useful information on resistance pathways and barriers to resistance for virus NAb escape . Our previous findings that combinations of antibodies can have synergistic effects in neutralizing HCV [57] , would be well complemented by detailed analyses concerning the barrier to resistance for these interesting epitopes . The inherent broad neutralization antibody-resistance of some HCV isolates hinders the study of escape mutations for these viruses , partly due to the excessive amounts of antibody needed to suppress viral spread [65] . One way to minimize the use of HMAbs is to treat sensitive HCV recombinants such as the original JFH1 , which is also not highly efficient in cell culture making it difficult for the virus to outgrow the antibody . Escape variants of JFH1 were isolated by culturing in the presence of around 10- or 50-times the IC50 concentrations of HMAbs CBH-2 or HC-11 , respectively [40] . A distinct approach that reduced the amount of antibody relied on several iterations of a single neutralization step of the highly sensitive recombinant virus HJ3-5 with AP33 prior to inoculation of naive cells with the non-neutralized viruses [43] . The escaped variants were isolated by 3 rounds of antibody neutralization with >100 times the IC50 ( 100 μg/ml ) of the antibody followed by amplification of the non-neutralized virus . However , the resistant viruses showed a considerable reduction of viral fitness indicating that the escaped virus needed additional mutations to compensate for the fitness loss . Another approach has been to passage the virus first in low doses of antibody followed by increasingly higher doses in subsequent passages , in the hope that escape will develop gradually under the low selective pressure . This approach has been shown to generate multiple JFH1 escape variants against HMAb HC33 . 1 [41] . While these are innovative solutions , they demand large amounts of antibody to generate escape variants for resistant HCV recombinants as shown in the case of selecting Jc1 ( genotype 2a recombinant ) escape mutants to AP33 [39] . Jc1 was serially passaged in the presence of AP33 at increasing concentrations up to ~40 times of the IC90 concentration ( i . e . 200 μg/ml ) [39] . There is currently no good way to induce escape mutations in naturally resistant HCV isolates ( such as J6/JFH1 ) leading to a potential bias of the studied viruses and concomitantly the escape barriers . Here , we induced AR5A-specific escape in H77/JFH1ΔHVR1 and J6/JFH1ΔHVR1 at antibody doses ( 5 μg/ml and 20 μg/ml , respectively ) that would be easily overcome by the parental viruses ( Fig 9A and 9B ) . In fact , we were unable to induce escape mutations in J6/JFH1 even at 200 μg/ml ( the highest dose we have observed others use to generate NAb escape [39] ) and only induced low level ( ~8-fold ) resistance at this high dose for H77/JFH1 , which was not specific to AR5A . This highlights another advantage of using HVR1-deleted viruses to study specific NAb escape as it eliminates the acquisition of mutations in the envelope proteins that could contribute to a general increase in neutralization resistance thus impeding direct identification of antibody-specific escape mutations [66] . Here , we studied the moderately neutralization sensitive H77/JFH1 and the generally resistant J6/JFH1 ( IC50 values against AR5A for H77/JFH1 and J6/JFH1 of 1 . 9 μg/ml and 36 μg/ml , respectively ) . The HVR1-deleted viruses had drastically reduced , and similar , IC50 values ( IC50 was 0 . 0035 μg/ml for H77/JFH1ΔHVR1 and 0 . 0065 μg/ml for J6/JFH1ΔHVR1 ) . Thus the deletion of HVR1 allowed us to induce escape at 5 μg/ml and 20 μg/ml of HMAb AR5A for H77/JFH1ΔHVR1 and J6/JFH1ΔHVR1 , respectively . Interestingly , while H77/JFH1ΔHVR1 and J6/JFH1ΔHVR1 were similarly sensitive to neutralization with AR5A the latter was able to outgrow the antibody at 5 and 10 μg/ml . Although viral spread was significantly delayed , we did not observe escape variants following treatment of J6/JFH1ΔHVR1 with AR5A concentrations lower than 20 μg/ml , thus indicating that the highly infectious J6/JFH1ΔHVR1 was able to simply outgrow the antibody . Thus , it seems that viability of the virus used in escape studies is an important factor to consider when selecting the appropriate concentration of antibody . However , it should be noted that the viability of H77/JFH1 is similar to J6/JFH1ΔHVR1 indicating that decreased viability of HVR1-deleted viruses does not , by itself , explain why it is relatively easy to inhibit viral spread at low doses of antibody . In addition , using low multiplicities of the IC50 values of antibody others have isolated mutations that did not specifically affect virus susceptibility [39;41] . Thus , a high multiplicity of the IC50 values could be essential to the rapid isolation of resistant variants . We propose that the use of the highly neutralization susceptible HVR1-deleted viruses allows for the isolation of antibody-specific resistant variants for otherwise neutralization-resistant HCV recombinants and eases the generation of escaped viruses for multiple isolates of different HCV genotypes . However , since the deletion of HVR1 has been shown to affect receptor interactions and the lipid composition of the viral particles [44;67] , more studies using other NAbs are needed to evaluate if escape results obtained using the HVR1-deleted variants will have general relevance . We identified escape mutations that arose in viruses cultured in the presence of AR5A . These mutations can be grouped according to their effects on virus susceptibility to AR5A and on virus viability . The only resistance mutation we identified through culturing viruses retaining HVR1 with AR5A was the substitution A349D , which was found by culturing H77/JFH1 at a high dose of AR5A . The mutation also conferred low level resistance against AR3A and could be similar in this respect to envelope mutations that broadly modulate NAb resistance such as those described by others [68–70] . However , it should be noted that aspartic acid ( D ) was not observed at this position in the Los Alamos database ( Table 1 ) , indicating that it is not a natural polymorphism . We identified two different mutations at position L665 ( L665W for H77/JFH1 and L665S for J6/JFH1 ) . Previously , we mapped the AR5A binding site to position R635 with an alanine substitution resulting in a reduction of relative binding of AR5A by 50–75% [27] . In the same study the alanine substitution at position L665 decreased the relative binding level of AR5A by 60% . However , since alanine is smaller than leucine and both amino acids are non-polar , the alanine scanning could be underestimating the importance of position L665 in AR5A binding . In contrast , the bigger amino acid group of tryptophan and the polarity of serine could induce the conformational change necessary to impair antibody binding . We showed that L665W abrogated AR5A binding to HCV particles and to intracellular E1/E2 protein , while also greatly increasing resistance to AR5A neutralization for genotype 1–6 viruses . It should be noted that while we are not able to provide evidence for whether the different pathways to resistance against AR5A are isolate or genotype dependent the high envelope protein heterogeneity between the genotype isolates we have tested for AR5A resistance does represent the high diversity of HCV . Recently , we showed that polymorphisms at positions 636 and 708 could be involved in the higher resistance against HMAb AR5A of the HCV strain S52 [48] , however the effect in resistance was lower than the one induced by the mutation L665W in the same genotype . Thus , our data strongly indicates that L665 is a part of the AR5A epitope . L665 was universally conserved in E2 sequences deposited in the Los Alamos HCV database [55] and the European HCV database [56] ( Table 1 ) , and the decrease in virus fitness for isolates from genotypes 2 , 3 , 4 , and 6 with L665W suggests that L665 is important for virus infectivity . It has previously been shown that highly conserved epitopes tend to have a higher barrier to resistance , probably due to their importance in critical interactions [28;40] . However , we found that while L665 was highly conserved the barrier to resistance was isolate-dependent . The observed differences in the escape mutations for viruses with and without HVR1 suggests that it might be of interest in future studies to attempt to induce escape for viruses with and without HVR1 and with non-viral neutralization epitopes in their envelope proteins , such as the flag-epitope with neutralizing potential at the N-terminus of E2 [71] , to investigate how escape mutations for such epitopes would cluster . L665 is localized within the stem region that connects the E2 ectodomain to the C-terminal transmembrane region [72] . We previously proposed that the stem region could play a role in virus entry [73] , however , E2 protein with the stem and transmembrane regions deleted could still interact with host entry factors [74] . Here , we did not observe any changes in receptor dependency for viruses harboring the mutation L665W . Furthermore , in a few cases L665W altered the sensitivity to other conformational antibodies suggesting it might have some effects on the structure of the E1/E2 complex . However , the substitutions at this position did not seem to affect E1/E2 interactions or cell-to-cell spread . It is conceivable that the mutations affect some other critical entry process , such as viral fusion . It will be of interest in future studies to further elucidate the role of this conserved site . During the treatment with AR5A we observed the mutation N417D for H77/JFH1ΔHVR1 and I345V for J6/JFH1ΔHVR1 . These mutations did not significantly alter AR5A susceptibility but increased viral fitness . Higher fitness of a treated virus seems to be an important predictor of whether the virus can outgrow a NAb as we previously proposed upon treatment in culture of the highly infectious HCV recombinant SA13/JFH1 [58] . The mutation N417D abolished the first N-linked glycan of the E2 protein and has been implicated both in increasing virus infectivity and in protecting the virus from NAbs [75] . Since we did not observe any change in antibody susceptibility , it is possible that the glycosylation at position N417 does not protect AR5 or even that the glycan shield is less effective in shielding HVR1-deleted viruses . The substitution N417T shifts the glycan at N417 to N415 and has been implicated in resistance against AP33 , which targets the epitope at 412–423 [39] . In a separate study with another 412–423 specific antibody , HC33 . 1 , the mutation N417T appeared at low concentrations of antibody , but was probably increasing fitness of the virus rather than conferring specific escape [41] . The mutation I345V is novel , but the near-adjacent substitution , I347L , has been identified as a compensatory mutation adapting the HVR1-deleted Jc1 virus to cell culture [46] . Thus , the two adaptive envelope mutations N417D and I345V are probably decreasing the effectiveness of AR5A in culture through increased viral fitness . The substitution S680T was of interest as it compensated for the fitness loss of L665S in J6/JFH1 and also increased AR5A-specific resistance . However , S680T had the opposite effect on neutralization sensitivity in the absence of L665S , hinting at a complex relationship between positions 665 and 680 while also suggesting that S680 is not part of the AR5A epitope . Several studies have described mutations that can indirectly alter HCV antibody susceptibility without themselves being a part of the antibody epitope [41;51;68;69;76] . Moreover , the phenotype of some of these mutations has been shown to depend on the envelope protein context [41;68] . However , S680T appears to be unique , as it modulates AR5A sensitivity in a position-665-dependent manner and simultaneously compensates for the lower viability of viruses harboring the L665S escape mutation . Since H77/JFH1 has threonine at position 680 ( as does most genotype 1 , 3 , 4 and 5 isolates of HCV , whereas genotype 2 and 6 typically have serine at this position ) and the virus H77/JFH1L665S was not viable , our results indicate that the compensatory effect of S680T for mutation L665S was specific for J6/JFH1 , indicating that the two pathways to resistance ( L665W for H77/JFH1 and L665S/S680T for J6/JFH1 ) are likely , at least partly , isolate-dependent . It is important that we observed isolate-dependent pathways requiring either one or two mutations to confer complete resistance of a viable virus as the former would be considered to have a low barrier to resistance whereas the other is relatively high . Using HVR1-deleted viruses enable researchers to perform escape studies of multiple isolates from different genotypes , providing relevant information even for otherwise resistant isolates , such as J6/JFH1 . In conclusion , using neutralization-sensitive HVR1-deleted viruses we were able to generate escape viruses in vitro against the HMAb AR5A . We showed that the selective pressure can result in different escape mutations depending on the envelope protein sequence . We identified novel escape mutations at positions either involved directly in binding , such as L665S and L665W or indirectly , such as S680T . Our data emphasizes the complexity of viral resistance against conformational antibodies , in this case HMAb AR5A , and highlights the importance of studying resistance using multiple isolates , preferably both of high and low inherent neutralization sensitivity . These important findings are only attainable through escape studies and are highly relevant in the context of viral infectivity and fitness . The knowledge gained concerning pathways and barriers to antibody resistance will aid the development of a broadly effective HCV vaccine .
Human monoclonal antibodies AR3A , AR4A , and AR5A against HCV structural proteins were produced as described previously [27;33] . Antibodies against HCV receptors were anti-CD81 ( BD Pharmingen Cat . JS81 ) , anti-SR-BI C16-17 [67;77] and polyclonal anti-LDLr ( R&D Systems Cat . AF2148 ) . Control antibodies for receptor blocking assays were the isotype antibody 553447 for CD81 , the antibody D for SR-BI [77] , and a goat IgG ( R&D Systems Cat . AB-108C ) for LDLr . Soluble CD81 protein was a kind gift from Steven Foung [73] . Adapted recombinants with the core-NS2 sequence from genotypes 1a ( H77 ) , 2a ( J6 ) , 3a ( S52 ) , 4a ( ED43 ) , 5a ( SA13 ) and 6a ( HK6a ) and UTR´s as well NS3-NS5B region from genotype 2a ( JFH1 ) with or without HVR1 were described previously [44;54;58–61] . H77 E1E2 expression plasmids , previously validated for functionality in HCVpp assays [67] , were used for E1/E2 interaction studies . Plasmids with point mutations were generated by conventional cloning techniques ( Fusion PCR and Quikchange ) . The HCV sequence of the final plasmid DNA preparations were confirmed by direct sequencing ( Macrogen ) . Antibody against NS5A , 9E10 [61] , was provided by Charles Rice . Huh7 . 5 cells [78] , provided by Charles Rice , were cultured in Dulbecco’s modified eagle medium DMEM ( Gibco/Invitrogen Corporation , Carlsbad , CA ) supplemented with 10% of heat-inactivated fetal bovine serum ( FBS ) , penicillin 100 U/mL and streptomycin 100 μg/mL ( Gibco/Invitrogen Corporation ) at 5% of CO2 at 37°C . Cells were split every 48 to 72 hours . Virus stocks were generated by low MOI infection of naive Huh7 . 5 cells . Supernatants were collected at the peak of infection , then filtered and stored at -80°C . HCV envelope ORF sequencing from culture supernatants was done by long RT-nested PCR procedures , as described [44;54;58–60] . Huh7 . 5 cells were plated at 4x105 per well in 6-well plates 24 hours prior to transfection at which point they were ~70% confluent . Plasmids were linearized by XbaI treatment ( New England BioLabs ) . RNA was generated by T7-mediated in-vitro transcription and was transfected into Huh7 . 5 cells using lipofectamine 2000 ( Invitrogen ) . 6 hours post transfection , Huh7 . 5 cells were trypsinized and reseeded into four wells of 24-well plates at a cell density of 8x104 ( for the 4 time-points of the assay ) along with plating in 6-well chamber slides for assessing percent infected cells at the four time-points , using the primary mouse anti-HCV NS5A 9E10 antibody and the secondary antibody Alexa488 goat anti-mouse IgG ( H+L ) ( Invitrogen ) as described [59] and Hoechst 33342 ( Molecular Probes ) counterstain for nuclei . Viral spread was monitored every 24 hours along with harvesting of virus supernatant up to 96 hours post transfection . Supernatants collected during experiments were sterile filtered and stored at −80°C . The virus titers were determined as described previously [79 , 80] . Huh7 . 5 cells ( ~70% confluent ) were infected with HCV virus H77/JFH1 or J6/JFH1 with or without HVR1 at an MOI of 0 . 001 and incubated in 5% of CO2 at 37°C . Virus infection was monitored by immuno-staining every 2–3 days as described above . After the virus infection spread to 1–10% , the indicated concentrations of antibody AR5A were added in each well . Supernatants were collected and filtered when cell infection reached 80–90% of the cells . 6x103 Huh7 . 5 cell per well was plated in poly-D-lysine 96-well plates and incubated for 24 hours . Next day , a volume of virus stock corresponding to a read-out of 50–300 focus forming units per well ( FFUs/well ) were incubated in quadruplicates with a dilution series of monoclonal antibody and relevant control antibody . Virus-antibody mixes along with eight replicates of virus only were incubated for 1 hour at 37°C and were added to Huh7 . 5 cells ( ~70% confluent ) and incubated for 4 hours at 37°C and 5% CO2 . Subsequently , the cells were washed and fresh medium was added prior to incubation for a total infection time of 48 hours . Cells were fixed and stained with 9E10 antibody as described previously [59] . The data was normalized to 8 replicates of virus only and analyzed using three or four parameters curve fitting in GraphPad Prism [67] . In all cases , the neutralization data were confirmed in two independent experiments . 6x103 Huh7 . 5 cells per well was plated in poly-D-lysine 96–well plates and incubated for 24 hours . The next day the Huh7 . 5 cells were incubated in four replicates with a 1 to 5 dilution series of monoclonal antibody against CD81 or SR-BI receptors or the polyclonal antibody against LDLr and four replicates of the respective control antibodies at the highest concentrations used in the assay . A volume of virus stock corresponding to a read-out of 50–300 FFUs/well of HCV were added to the cell-antibody mix and incubated for 4 hours at 37°C ( cells were ~70% confluent at time of infection ) . Cells were washed and fresh medium was added prior to incubation for a total infection time of 48h . Cells were fixed and stained with 9E10 antibody as described [59] . The data was normalized to 8 replicates of virus only and analyzed using four parameters curve fitting in GraphPad Prism . Blocking data were confirmed in two independent experiments . Huh7 . 5 cells were plated at 400 . 000 cells/well in 6-well plates and the following day the cells were transfected with HCV RNA transcripts ( cells were ~70% confluent ) as previously described [59] . The following day cells were trypsinized and mixed with naive trypsinized Huh7 . 5 cells prior to plating at 12 , 000 Huh7 . 5 cells per well in poly-D-lysine 96–well plates ( NUNC ) . 6 wells per plate for each virus condition were seeded along with 12 wells per plate with only naive Huh7 . 5 cells to estimate background staining on each plate . A ratio of transfected/naive cells of 1:150 was used for cells plated in standard medium and a ratio of 1:30 was used for cells plated in standard medium in the presence of 10 μg/ml of the cross-genotype reactive HCV neutralizing antibody , AR3A [33] . This dose represented at least 500-fold the IC50 value for the tested HVR1-deleted viruses [48] . 6 replicates of cell mixes from each virus construct were plated in the absence or presence of AR3A for fixation at the two time-points; 24 hours and 48 hours post-plating . At 24 hours and 48 hours a plate was fixed in methanol and all plates were subsequently stained for infection using the NS5A-specific antibody , 9E10 . Number of single infected cells , number of focus forming units ( clusters of single cells ) and size of FFUs were counted and calculated using adapted BioSpot software ( Cellular Technology Lmtd . ) . Data was analyzed using GraphPad Prism . 293T cells ( American Type Culture Collection , ID: CRL-1573 ) were plated at 400 . 000 cells/well in a 6-well plate and the following day the cells were transfected using Lipofectamine 2000 ( Invitrogen ) with 5 μg of E1E2 expression plasmids . The following day the cells were replated back into the 6-well plate except for the plating of two slides . The next day the slides were fixed with 4% paraformaldehyde at room temperature for 15 minutes and the cells were permeabilized with 0 . 1% Tween20 . Subsequently the slides were stained for E2 or E1/E2 using AR3A or AR4A , respectively , followed by incubation with anti-human Alexa488 coupled secondary antibody and Hoechst 33342 ( Molecular Probes ) counterstain for nuclei . Transfection efficiencies were comparable . Cells in 6-wells were lysed using a 1% nDDM detergent in NativePAGE sample buffer ( Novex life technologies/Thermo Scientific ) supplemented with 1x Halt protease inhibitor cocktail ( Thermo Scientific ) by pipetting the solution up and down several times . The lysates were cleared by centrifugation at 4°C at 20 . 000xrcf for 30 minutes . The samples were transferred to fresh tubes , MgCl2 was added in a final concentration of 2mM and samples were treated with Benzonase endonuclease ( Sigma ) for 30 min prior to immunoprecipitation with the conformationally-reactive E2 antibody , AR3A , using Protein G coupled Dynabeads ( Thermo Scientific ) as per the manufacturer´s instructions . The E1/E2 protein complexes were eluted directly into LDS buffer by heating the beads for 10 minutes at 70°C . The E1/E2 protein complexes were loaded onto SDS-PAGE gels and run for 1 hour at 200 V under reducing conditions . This was followed by transfer onto a PVDF membrane at 35 V for 1 hour . E2 protein was visualized using mouse H52 antibody [81] and E1 protein was visualized on separate membranes using mouse A4 antibody [82] incubated overnight at 4°C on a shaker followed by incubation with anti-mouse coupled with horse radish peroxidase for 1 hour at room temperature on a shaker . The Supersignal West Femto Chemi-luminescence ( Thermo Scientific ) kit was used . The size of proteins was estimated by using the Precision Plus Protein WesternC standard ( Bio-Rad ) size marker as per the manufacturer´s instructions . We plated 6x103 Huh 7 . 5 cell per well in poly-D-lysine 96-well plates and incubated for 24 hours . Next day , a volume of virus stock corresponding to a read-out of 50–300 FFUs/well of HCV was incubated in four replicates with a dilution series of soluble CD81 receptor protein . Virus-CD81 mixes along with eight replicates of virus only were incubated for 1 hour at 37°C and were added to Huh7 . 5 cells ( ~70% confluent ) and incubated for 4 hours at 37°C and 5% CO2 . Subsequently , the cells were washed and incubated for a total infection time of 48h with fresh medium . Cells were fixed and stained with the 9E10 antibody as described [59] . The data was normalized to 8 replicates of virus only and analyzed using four parameters curve fitting in GraphPad Prism . Experiments data were confirmed in two independent experiments . Huh7 . 5 cells were infected with viruses H77/JFH1 or H77/JFH1L665W with a multiplicity of infection ( MOI ) equal to 0 . 01 and incubated for 24 hours . We plated 2 . 5x104 infected Huh7 . 5 cells per well in an 8 wells slide and incubated them for an additional 24 hours . Cells were fixed with para-formaldehyde at room temperature for 15 minutes and the cells were permeabilized with 0 . 1% Tween20 . Antibody anti-NS5A in combination with either AR4A or AR5A was used as primary antibodies . We used a combination of Alexa594 goat anti-mouse IgG ( H+L ) ( Invitrogen ) and Alexa488 goat anti-human ( Invitrogen ) as secondary antibodies with a Hoechst 33342 ( Molecular Probes ) counterstain for nuclei . Immunoprecipitation was carried out using the immunoprecipitation kit Dynabeads Protein G ( Thermo Scientific ) as previously described [67] . We used 5 μg of antibodies AR3A , AR4A or AR5A along with a relevant isotype antibody control b6 . Magnetic bead-associated HCV RNA was eluted with lysis buffer from the QIAmp MinElute Vacuum kit ( Qiagen ) and the HCV RNA was extracted using the kit protocol together with a dilution series of a standard sample with a known HCV RNA concentration and a blank solution . HCV RNA was eluted in 30μl of elution buffer and 8μl of each sample was used for reverse transcription-quantitative PCR ( RT-qPCR ) using a Light Cycler as described before [59;67] .
|
Worldwide hepatitis C virus ( HCV ) is one of the leading causes of chronic liver diseases , including cirrhosis and cancer . Treatment accessibility is limited and development of a preventive vaccine has proven difficult , partly due to the high mutation rate of the virus . Recent studies of HCV antibody neutralization resistance have revealed important information about escape pathways and barriers to escape for several clinically promising human monoclonal antibodies . However , due to the varying levels of antibody shielding between HCV isolates these studies have been mostly limited to a few neutralization-sensitive HCV isolates . Here , we took advantage of the fact that deletion of the hypervariable region 1 ( HVR1 ) increased antibody sensitivity of HCV isolates by increasing the exposure of important epitopes , thus facilitating studies of antibody escape for neutralization resistant isolates . We identified escape mutations in the envelope glycoprotein E2 , at amino acid position L665 , which conferred antibody resistance in parental HCV viruses from genotypes 1–6 . We found that antibody escape was associated with loss of binding to HCV particles and intracellular envelope protein complexes . We also identified escape substitutions at L665 that were isolate-specific . Thus , our data sheds new light on antibody resistance mechanisms across diverse HCV isolates .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"microbial",
"mutation",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"hepacivirus",
"pathogens",
"immunology",
"microbiology",
"viral",
"structure",
"viruses",
"mutation",
"substitution",
"mutation",
"rna",
"viruses",
"immunoprecipitation",
"mutation",
"databases",
"antibodies",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"biological",
"databases",
"hepatitis",
"c",
"virus",
"hepatitis",
"viruses",
"viral",
"replication",
"precipitation",
"techniques",
"biochemistry",
"viral",
"envelope",
"flaviviruses",
"virology",
"physiology",
"viral",
"pathogens",
"genetics",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2017
|
Applying antibody-sensitive hypervariable region 1-deleted hepatitis C virus to the study of escape pathways of neutralizing human monoclonal antibody AR5A
|
Microscopy , being relatively easy to perform at low cost , is the universal diagnostic method for detection of most globally important parasitic infections . As quality control is hard to maintain , misdiagnosis is common , which affects both estimates of parasite burdens and patient care . Novel techniques for high-resolution imaging and image transfer over data networks may offer solutions to these problems through provision of education , quality assurance and diagnostics . Imaging can be done directly on image sensor chips , a technique possible to exploit commercially for the development of inexpensive “mini-microscopes” . Images can be transferred for analysis both visually and by computer vision both at point-of-care and at remote locations . Here we describe imaging of helminth eggs using mini-microscopes constructed from webcams and mobile phone cameras . The results show that an inexpensive webcam , stripped off its optics to allow direct application of the test sample on the exposed surface of the sensor , yields images of Schistosoma haematobium eggs , which can be identified visually . Using a highly specific image pattern recognition algorithm , 4 out of 5 eggs observed visually could be identified . As proof of concept we show that an inexpensive imaging device , such as a webcam , may be easily modified into a microscope , for the detection of helminth eggs based on on-chip imaging . Furthermore , algorithms for helminth egg detection by machine vision can be generated for automated diagnostics . The results can be exploited for constructing simple imaging devices for low-cost diagnostics of urogenital schistosomiasis and other neglected tropical infectious diseases .
Microbiological diagnostics at the basic levels of the health care system has to meet the challenges of harsh environmental conditions , inadequately trained personnel and difficulties to maintain routines for quality assurance . There is a widespread failure to understand that diagnosis is essential to the prevention and treatment of disease [1] A major problem is that methods developed in well-equipped laboratories are difficult to maintain due to lack of resources [2] , [3] . The methodological requirements differ depending on the purpose , diagnostics , epidemiology , effect of intervention etc . [4] , [5] . Effective control measures put the focus on diagnostics , which needs to be adapted to the stage of control . The number of countries reporting on schistosomiasis treatment increased from 17 in 2008 to 21 in 2009 but the number of people treated for schistosomiasis in 2009 is only 8 . 2% of the estimated number of people infected [6] . Even more disturbing is the calculation that fewer than 5% of the infected population is receiving antischistosomal treatment and the conclusion that we may be facing “one of the first great failures of the global health decade” that began in 2000 [7] . The focus is increasingly on diagnostics of chronic multiple parasitic infections affecting populations in poor rural areas [8] , [9] . There is a definite shift from morbidity control to transmission control after long-term treatment campaigns [10] , [11] . This underlines the need for monitoring tools . With decreasing transmission rates , factors related to chronic infection - not parasite load - become important . Diagnostics of urogenital schistosomiasis based on the presence of blood in urine has been successfully used as a low-cost method in a high prevalence situation [12] , but such indirect methods become less useful in low prevalence and low intensity situations [13] . With decreasing endemicity the requirement for high specificity becomes more and more important – and increasingly difficult to maintain . Flawed information becomes a growing problem , which may have an unexpected impact on the fundamentals of a control effort . A large-scale commitment to eliminate major neglected infectious diseases by the end of the decade , the “London Declaration on Neglected Tropical Diseases” [14] will require reliable tools for diagnostics and monitoring . It is obvious that recent developments are going to have an impact on our possibilities to perform diagnostics even under difficult conditions in poor endemic regions . These developments affect information transfer and analysis but also novel tools for obtaining basic information . Digital web-based microscopy over the Internet offers the possibility not only for education and quality assurance [15] , a central computer may also serve as a diagnostic unit as mobile imaging devices such as mobile phones can bring microscopy in contact with diagnostics at a distance . The interpretation of an image can be performed by an expert or even by crowd-sourcing diagnostics to non-experts [16] located essentially anywhere . It is the basis for much of telemedicine and used in radiology , cardiology etc . The use of a mobile phone to transfer a microscope image as described by Frean [17] , has an unexploited potential - as shown e . g . by Zimic [18] - considering the proliferation of mobile networks and the possibility of integrating various types of health care related information [19] . Novel tools are emerging for producing high magnification images . By placing a small ball lens over the mobile phone camera lens [20] or by placing objects in close proximity to the surface of a sensor chip , “on-chip imaging” [21] . These methods have the potential to revolutionize diagnostic imaging , which today can be achieved only using a microscope . A current limitation of the methods is that either the sharp image area is very small , as in case a ball lens is used , or the resolution is limited as in case of on-chip imaging . On-chip imaging using rather elaborate computational holographic techniques such as diffraction analysis and super-resolution by the pixel shift technique [22] , [23] , [24] has shown that high-resolution imaging can be performed . In this study we wanted to test if the basic technique of imaging an object directly on the surface of an image sensor chip of a webcam or a mobile phone camera - which is so simple that it has been presented as a hobby project ( http://makeprojects . com/Project/Lensless-Microscope/220/1 ) - can be used for the diagnostics of helminth infections . As “proof-of-principle” , analysis of webcam images was done using computer algorithms for the identification of Schistosoma haematobium eggs in the urine . We show that it is possible to use pattern recognition to “duplicate the abilities of human vision by electronically perceiving and understanding an image approach to image analysis” [25] in the diagnostics of urogenital schistosomiasis .
Patient urine and stool materials containing excreted helminth eggs were handled in accordance with the Swedish “Biobanks in Medical Care Act” ( 2002:297 ) and the “New Biobanks Act” ( Swedish government Report , SOU 2010:81 ) stating that “… . samples may be collected , stored and used for certain purposes ( including research and cross-border exchanges of samples and data ) , with respect for the individual integrity and privacy . ” ( http://www . hsern . eu/index . php/news/show/sw-swedish-government-published-a-report-sou-2010-81-entitled-a-new-biobank-act ) Samples were not collected specifically for this study . All human samples obtained under oral consent and anonymized were from an already-existing sample collection for education and quality assurance ( “Panel för Cystor & Maskägg” at SMI ) . Anonymized stool samples provided by Jürg Utzinger at the Swiss Tropical and Public Health Institute , Basel , Switzerland were collected for collaborative evaluation of diagnostics and quality assurance and approved by the ethics committees of Basel ( EKBB , 377/09 ) and Côte d'Ivoire ( reference no . 1993 MSHP/CNER; date 2010-05-10 ) . Schistosoma mansoni eggs were obtained from mice kept according to national guidelines ( Swedish Board of Agriculture SJVFS 2012:26 ) . Mice were experimentally infected to provide materials for diagnostics of human infections . The protocol was evaluated and approved by the regional ethical committee Stockholm North , Dpt . 1 , Stockholm District Court ( reference no . N527/11; date 2011-01-26 ) . Experiments reported here were performed on a urine sediment obtained by pooling urines from individuals shown to excrete S . haematobium eggs . The formalin fixed sediment was stored at +4°C . For on-chip experiments , aliquots of the sediment were diluted in saline to give a concentration of about 250 eggs per ml . The concentration corresponds to a 10-fold concentration of 250 eggs in 10 ml of urine allowed to sediment and then re-suspended into 1 ml . The concentration of more than 50 eggs per 10 ml is considered to reflect an infection of high intensity [26] , [27] . The S . haematobium sample was obtained from the diagnostic parasitology laboratory of the Swedish institute for communicable disease control ( SMI ) , Solna , Sweden . Samples containing intestinal helminth eggs ( S . mansoni , Trichuris trichiura and Diphyllobothrium latum ) and Strongyloides larvae were from standard formalin or SAF-fixed ( fixative containing sodium acetate , acetic acid and formalin ) human stool samples . Pooled isolated S . mansoni eggs used for some on-chip experiments were obtained from experimentally infected mice as previously described [28] , [29] . Images of helminth eggs were captured for reference purposes using established techniques: For part of the samples we used a microscope ( Leica DMRB , Leica , Leitz ) equipped with a digital camera ( AxioCam; Carl Zeiss; Oberkochen , Germany ) and using image capture . Imaging software ( Openlab , Improvision; Coventry , United Kingdom ) on a desktop computer ( Apple Macintosh G4 with McOS 9; Cupertino , CA ) was used for image capture . Specimens containing S . haematobium eggs and Strongyloides larvae were digitized for web-based virtual microscopy as described before [15] . Some samples were also digitized with an automated whole slide scanner ( Pannoramic P250 , 3DHistech Ltd , Budapest , Hungary ) , using a 20×objective ( numerical aperture 0 . 8 ) equipped with a three-CCD ( charge-coupled device ) digital camera ( CIS 3CCD , 2 megapixel , CIS Corporation , Tokyo , Japan ) . The pixel resolution was 0 . 22 µm . The images were compressed with a conservative compression ratio of 1∶5 to a wavelet file format ( Enhanced Compressed Wavelet , ECW , ER Mapper , Erdas Inc , Atlanta , Georgia ) and made available for web-based virtual microscopy [15] . To enable fixation of liquid samples for the whole-slide imaging , specimens of helminth eggs were mounted and immobilized under coverslips on microscope slides with a drop of glycerin-gelatin ( Sigma-Aldrich product GG1 aqueous slide mounting medium ) at 55–60°C . Samples were scanned not only in the x and y planes , but also in different focal planes in order to generate z stacks to enable focusing in the web-based viewer . On-chip imaging was performed essentially by placing the specimen in contact with an image sensor , which was then illuminated to produce a shadow of objects present in the specimen . The CMOS ( Complementary Metal Oxide Semiconductor ) sensor chip of an imaging device was made available for imaging experiments by removing the optics ( see supporting information S1 and S2 ) . The main results reported here were obtained with the exposed sensor of a low cost webcam ( Live ! Cam Sync; ( VFO520 , 640×480 pixel , Creative Technology Ltd . Singapore , sold by Clas Ohlson Co; Insjön , Sweden as Webbkamera , product 38-3612 for 99 , 00 SEK ( i . e . approximately 11€ ) with a calculated pixel size of 3 . 658 µm , in which the sensor was covered with a protective glass at a level allowing direct contact with microscope slides and resolution slide ( see below ) . In reality , the sensor area is smaller than the cover glass and thus pixel size is smaller if calculated based on images obtained ( see results ) . We could not test the effect of bringing objects closer to the surface of this particular actual sensor as we were unable to remove the protective covering glass without causing damage to the sensor surface . Depending on the physical appearance of the exposed sensor , additional modified imaging devices were used in experiments related to specific issues , such as the effect of image sensor pixel size on image resolution and construction of a chamber on top of the image sensor for the analysis of fluid samples . The image sensors were surrounded by protruding components of the camera circuit board , which prevented positioning of a flat microscope glass slide directly on top of the sensor chip surface . In such cases a drop of the sample was placed directly in contact with the surface of the sensor . This was done after protecting the components of the circuit board from exposure to fluid using silicone or acrylate polymer . Such on-chip experiments were performed with the exposed sensor of another webcam ( “Venus” USB 2 . 0 PC Camera , Vimicro Corporation; Beijing , China , sold by Clas Ohlson Co; Insjön , Sweden as Webbkamera , product 38-4068 ) . Some experiments were performed using the exposed 8 megapixel ( 3624×2448 pixels; pixel size 1 . 75 µm ) sensor of mobile phone after removal of the thick protective glass and replacing it with a piece of coverslip with a thickness of 0 . 1 mm ( Sony Ericsson C905 , Sony , Japan ) . Superior resolution was obtained using the exposed sensor of a mobile phone ( Nokia E71 , 3 . 2 megapixel; 2048×1536 pixels , pixel size 1 . 75 µm , Nokia , Finland ) camera . For the experiments , replacement camera modules ( n = 60 ) were acquired for the mobile phone ( E71 Camera w/Flex Ribbon; eBay , Unclemartin; China ) . The sensor surface was hard to access and it was difficult to protect the surrounding circuit board components from damage caused by fluid samples . As the image sensor of this particular camera module was not protected by a cover glass it was easily damaged by the drying urine sample , which became attached to the microlens polymers on the sensor surface . Thus a new camera had to be installed for each experiment . In some modifications , a chamber consisting of a plastic test tube was fitted above the image sensor: A rectangular hole corresponding to the size of the sensor was cut in the lid of the tube and fixed to the circuit board with silicone as described above . A sedimentation chamber was obtained by attaching a test-tube to the lid . Sedimentation was allowed to take place by inverting the test tube to allow particles to sediment onto the sensor surface . The test tube with supernatant was then removed and replaced with a light source . In all on-chip imaging experiments the resolution of images depended on the intensity , the size of the light source and the degree of collimation of light hitting the sensor . Near collimated light was obtained placing a small LED light at a distance of about 20 cm from the sensor . The distance from the light source could be decreased to about 25 mm using a plano-convex lens ( radius curvature 12 . 7 mm , diameter 25 . 4 mm BK 7 KPX043 , Newport Corporation; Irvine , CA , United States of America ) . As an alternative to LED light , we used indirect daylight from a 1 mm plastic monofilament core 2 . 2 mm diameter fibre optic cable ( HARTING; Sibiu , România , purchased from Elfa Distrelec AB; Solna , Sweden ) . Resolution was measured using an optical resolution slide ( NBS USAF 1951 Test chart – R70 TN8 6HA . Pyser-SGI; Fircroft Way , Edenbridge , United Kingdom ) . The square glass slide was cut to a width of 25 mm in order to fit in close approximation to the exposed sensor chip of the webcam Live ! Cam Sync . Calibration beads of similar size as helminth eggs were polydisperse glass particle standards ( refractive Index 1 . 51–1 . 52 ) for image analysis calibration with a range of 50–350 µm ( WhitehouseScientific Co . ; Waverton , Chester , CH3 7PB , United Kingdom , http://www . whitehousescientific . com/ ) Image-capture and transfer was done using the camera imaging software provided by the manufacturer . Still images 640×480 pixels ( VGA resolution ) of S . haematobium eggs were used to create algorithms for computer vision ( see supporting information S3 and below ) . The purpose was to identify , by computer vision , S . haematobium eggs in on-chip images of urine sediment obtained with the modified lensless webcam ( Live ! Cam Sync ) . Image analysis algorithms were used to detect S . haematobium eggs in images . Images with eggs detected by the observer were classified as positive samples . ( see Supporting Information S4 , Algorithms for the detection of S . haematobium eggs by computer vision ) . To develop the detection method ( Algorithm 1 ) , 243 images were used for training a parametric model . Images were preprocessed to normalize brightness differences and to enhance contrast [30] . The preprocessed images were thresholded based on grey values [31] . On the thresholded images , regions of interest ( ROIs ) were generated based on morphological methods [32]: the images were morphologically opened in order to remove small structures . Too large and too small blobs ( binary large objects , [31] ) were eliminated . The ROIs were classified into positive ( eggs ) and negative ( no eggs ) based on the area , shape and contrast of regions in the original image . As shape features , eccentricity and major and minor axis were used . Also , pairs of blobs within a certain distance from each other were combined to one egg hypothesis . The parameters for the grey value threshold and for the features to classify the ROIs were derived from 660 manually labeled eggs in 243 training images . A second set of 545 labeled eggs in 119 images with was then used for testing the detection method . The initial results of image analysis based on the Algorithm 1 described above gave results ( see below ) which warranted further image analysis studies using a more advanced algorithm ( Algorithm 2 ) where images are processed by a sequence of classifiers each stage rejecting false positive samples passed through the previous stages . ( See S3 ) The Haar-feature based cascade classifier [33] with 45 stages containing a total of 454 weak classifiers was trained using 500 cropped egg images . 400 of these were the confirmed detections from the first classification method . The training set was extended by 100 images , which were generated by applying small distortions of randomly selected cropped images . Ten thousand negative ROIs were obtained from images of urine sediment with no eggs present . Due to the limited number of samples , the classifier was tested using seventy-five synthetically generated images where egg images were rotated and added to background . The size variation of detections was limited between ±15% of the expected egg size . ( S . Varjo and J . Hannuksela: A Mobile Imaging System for Medical Diagnostics , Proc . Advanced Concepts for Intelligent Vision Systems ( ACIVS 2013 ) , Poznan , Poland , due to appear in volume 8192 of the Lecture Notes in Computer Science series . ) Sensitivity ( recall/completeness ) was calculated as the percentage of true positive ( TP ) divided by true positives and false negatives ( FN ) . Positive predictive value ( precision/correctness ) was calculated as the percentage of true positives divided by true positives and false positives ( FP ) .
S . haematobium eggs could be recognized in on-chip images obtained using a simple modified webcam ( Figure 1; Supporting information S1 , S2 and S3 ) . Samples on microscope slides or in liquid form could be placed directly or pipetted on the exposed sensor ( Figure 1A ) after protecting surrounding components with acrylic resin or silicone ( Figure 1B ) . A chamber could easily be fitted on top of the sensor , e . g . , using a pierced test tube lid ( Figure 1C ) and the inverted test tube could function as a sedimentation chamber . After removing the test tube , it could be replaced with an appropriate light source ( see below ) . After trimming the edges of the glass supported USAF 1951 resolution chart glass slide to 2 cm width with a glass cutter it could be positioned onto the exposed webcam image sensor . The sensors of the other devices tested were inaccessible to the resolution test slide due to components of the circuit board protruding above the sensor level . The maximum resolution of on-chip images seen on this webcam was about 40 lines per mm ( group number 5 , element 3 or 4 ( Figure 2A ) . The length of schistosome eggs equaled roughly that of the line length of the first element of group 4 in the USAF 1951 resolution chart , which is 0 . 15625 mm . The sample field of view dimensions correspond to the dimensions of the sensor size , which is 2 . 341×1 . 756 mm , i . e . approximately 4 . 11 mm2 . As the sensor has 640×480 pixels , the calculated pixel size is 3 . 658 µm . Figure 2B shows on-chip image of 50–350 µm calibration beads . The images of S . haematobium eggs obtained by on-chip imaging ( Figure 2C and supportive information S3 ) were of low resolution as compared to microscope images . For reference , see scanned slide “Schistosoma haematobium” , in the virtual webmicroscope ( fimm . webmicroscope . net/Research/Momic/helmintex ) . However , it was possible to use such images obtained by on-chip imaging for the development of a computer algorithm ( see below ) . It was necessary to adjust the amount of light directed towards the sensor . In fact adjusting the amount of light was necessary for obtaining an image . When a LED light source at a distance of about 10 cm from the sensor chip surface was used together with a pin-hole aperture of 2 mm diameter , sufficient light was obtained . For the parameter tuning of the computer vision classifier , 660 eggs in the 243 training images were annotated as certain eggs ( egg identified without doubt; n = 564 ) uncertain eggs ( may be an egg; n = 96 ) or negative ( not an egg ) ( Figure 2D ) . Five hundred sixty four were labeled “positive” and 96 were labeled “uncertain” . After training , the approach was tested on a second image set consisting of 119 test images , which were taken from a new sample at a later date . A total of 545 parasite eggs were manually labeled; 414 certain and 131 uncertain eggs and the object co-ordinates in the image used as reference in the evaluation of the computer vision algorithm ( Figure 3 and 4 ) . When the performance of the algorithm ( Algorithm 1 ) in detecting individual eggs ( detection at the object level ) was calculated , we obtained a sensitivity ( recall/completeness ) of 26% and positive predictive value ( precision/correctness ) of 91% . Using the more complex cascade classifier involving 45 stages ( Algorithm 2 ) , Using the more complex cascade classifier , the sensitivity was 71% and positive predictive value 79% calculated from detections in 75 test image pairs , each image pair containing an image with a single egg and its companion image from which the egg had been replaced by background . ( Sensitivity = TP/ ( TP+FN ) = 53/ ( 53+22 ) = 0 . 7066→70 , 7% ) , positive predictive value = TP/ ( TP+FP ) = 53/ ( 53+14 ) = 0 . 7910→79 . 1% . ) ( TP = true positive; FP = false positive; FN = false negative; TN = true negative ) . The calculated high specificity was based on the large number of true negative images used for generating the cascade classifier ( with 10000 true negative samples the specificity - TN/ ( FP+TN ) - approached 100% . When the performance of Algorithms 1 and 2 was compared using the same set of 75 images the difference in precision and recall was evident , but less pronounced ( see supporting information S4 and S5 ) . On-chip imaging experiments performed using stool samples containing various helminth eggs showed that eggs from helminths of different species could be distinguished from each other . Morphological features such as the lateral spine of S . mansoni eggs could be identified with certainty in some eggs ( Figure 5A ) . The shape of S . mansoni eggs was clearly distinct from that of , e . g . , T . trichiura eggs . ( Figure 5C ) , and Strongyloides larvae were visible both in still images ( Figure 5E ) and in video recordings ( see supporting information S6 ) . On-chip images obtained with the webcam and Ericsson mobile phone sensors had a relatively poor resolution in comparison with images obtained by conventional microscopy ( Figure 5B ) . However , despite technical problems due to the inaccessibility of the exposed image sensor ( see above ) , we were able to obtain the highest resolution on-chip images with one of the mobile phone cameras in which the sensor surface was not covered by protective glass ( Nokia E71; Figure 6 ) . This setup minimized the distance between the sample and the sensor surface . On-chip imaging provides further diagnostic possibilities based on motion detection . On-chip video recordings of stool samples showed that moving Strongyloides larvae in fluid specimens ( Figure 5E ) may be identified ( see supporting information S6 ) .
In the present study we show that helminth eggs placed directly on the image sensor of inexpensive imaging devices can be visualized in sufficient detail to be identified on standard displays . ( se presentation; supporting information S7 ) If a mobile phone image sensor is used , the image can be directly sent for analysis at a distance . In this study we had access only to pooled samples and no attempt was made to establish the sensitivity of egg detection on a sample level in comparison to the standard reference method , which is expressed as egg counts per 10 ml of urine [17] . Image analysis could be performed either locally or centrally . Centralized diagnostics after image transfer may be performed based not only on visual inspection , but a central computer may also perform diagnostics more or less independently based on a computer vision algorithm . Image analysis may be performed by a remote computer [15] accessible through a network of servers [34] in much the same way as pattern recognition is applied for diverse purposes such as identification of immunoelectrophoretic patterns [35] and automated facial recognition for checking the passport at border controls [36] . Field studies are necessary in order to establish the sensitivity and specificity of both visual image interpretation and computer vision . The methods need to be assessed with ordinary microscopy as reference . Visual identification of helminth eggs present in on-chip images obtained with the simple webcam was successful , but the high specificity of image analysis was linked to a sensitivity , which needs to be improved before the method can be established in real-world situations . The performance of the computer algorithms in comparison to visual identification suggested , that we need to consider variables such as variation in egg size and shape , presence of inflammatory cells , casts , bacteria etc . in the background . To identify S . haematobium eggs in urine , we developed two computer algorithms . The original algorithm based on morphological features , was capable of identifying eggs correctly ( precision/positive predictive value/true positive rate ) only in 26% of on-chip images with a specificity of 90% . The precision of the algorithm was 71% . The second algorithm , based on cascading classifiers had a specificity approaching 100% and an improved sensitivity of 79% as compared to visual identification . In the present study we compared computer vision to visual image interpretation ( as gold standard ) and generated computer algorithms with high specificity for the evaluation of algorithm sensitivities ( recall rates , which depend on the rate of false negatives ) and precision ( which is affected by the number of false positives ) . The computer algorithms generated both false positive and false negative results ( about one out of four eggs ) , which reflect the superiority of visual interpretation of images . For a real-world test the 79% sensitivity should probably be improved to above 90% . On chip imaging as a potential field assay for automated diagnostics depends on two parameters , first , the capacity of imaging objects ( helminth eggs ) with sufficient resolution to permit identification . Second , diagnostics depends on correct identification of objects in the test sample . The image quality can be determined in terms of resolution – lines per mm -as in the present study . Clearly resolution depends on several parameters in addition to pixel size of the sensor; distance from the sample to the sensor surface and a well-controlled illumination are critical for the acquisition of on-chip images suitable for computer vision . A higher resolution will improve the accuracy of image analysis . There is a theoretical limit to the resolution , which can be achieved , set by pixel size as stated by the Nyquist-Shannon sampling theorem , stating that the maximum achievable resolution is twice the sampling frequency . Our experimental setup using the inexpensive webcam , Live ! Cam Sync , allowed us to detect objects with a resolution of 12 . 41 µm . The pixel size of the webcam image sensor was calculated to be 3 . 7 µm and the observed resolution therefore somewhat poorer than the theoretical resolution limit of 7 . 4 µm . In practice the resolution is less than the theoretical limit depending on the distance of an object to the sensor surface and due to imperfect collimation of light . The observed image resolution was much better with the Nokia E71 3 . 2 megapixel image sensor ( 2048×1536 pixels , pixel size 1 . 75 µm ) without a protective cover glass . The 3 . 5 µm theoretical resolution-limit reflects the small pixel size . Further improvements of on-chip mini-microscopes are therefore within reach , since recently introduced camera sensors have a pixel size close to 1 µm . We envision that an on-chip imaging device can be incorporated as an add-on to mobile phones capable of image transfer . On-chip imaging can benefit from the proliferation of mobile phones and the expanding data communication networks may provide the necessary infrastructure for functioning communication with a central server . Thus on-chip imaging may become an integrated part of telemedicine platform based on image capture , -transfer , -analysis and feedback . To meet the Millennium Development Goals complex political and social re-thinking is needed at different levels [2] , [37] . Health care is not isolated from the social and economic life of humans and we need to understand in detail how novel tools can be integrated in point-of-care diagnostics in much the same way as novel tools for microeconomics have revolutionized the life of “the bottom billion” [38] , [39] . Limitations posed by current microscopy-based diagnostics - and national surveillance systems depending upon them- need to be resolved . Especially among poor populations of the world , microscopy needs to be adequate [40] . However complex these issues are , there is a widespread opinion that telemedicine will play an increasingly important role in managing health care in affluent and resource-poor societies alike , and tele-medical solutions will without doubt contribute to democratization of the relationship between patient-physician and family [41] , [42] . Numerous compact and cost-effective optical imaging platforms , “mini-microscopes” have been developed in recent years to improve access to effective and affordable healthcare [23] . In a recent study it was shown that soil transmitted helminths can be detected in images obtained with a mobile phone [43] equipped with a small ball lens , a technique shown to generate high magnification high resolution images [21] . Like other described “mini-microscopes” on-chip imaging as described in the present paper does not require any new procedures since microscopy is the ‘gold’ standard for identification of parasitic infections . Standardized methods exist for sample collection , handling and preparation [44] , [45] . The on-chip image quality is subject to diffraction artifacts caused by the absence of optical components . Proposed solutions to this problem include computational reconstruction methods ( partially coherent in-line holography approach ) [reviewed in 23] to obtain microscope-like images . However , in the case of imaging helminth eggs , diffraction artifacts or distortions do not seem to undermine visual identification , as seen in e . g . Fig . 6A . One big advantage of the on-chip method described here is the large field of view - a simple webcam sensor has an area of over 10 mm2 , e . g . more than 6 times the visual field of a conventional microscope using a 20×objective and more than 10 times the field of view using a typical ball lens [23] . On-chip microscopy , even without a computer algorithm , involves examining fewer visual fields . It can alter the tiresome routine microscopy for finding and correctly identifying parasites present in low numbers - one of the major reasons for perceived low status of the method and its failure [46] , [47] . Our results suggest that automated diagnostics of helminth infections for field use using a simple imaging device and appropriate algorithms are within reach . A decisive advantage of a mini-microscope such as the one we describe , may prove to be the potential of providing diagnostic support by computer vision at a distance . Furthermore , our results suggest that diagnostics based image analysis has a potential to compete with laborious conventional microscopy e . g . by providing automated motion recognition for the detection of live nematode larvae .
|
There is a need to develop diagnostic methods for parasitic infections specifically designed for use in resource-deficient situations . Worm infections are common in many poor countries and even if repeated treatment can be arranged at low cost , diagnostics and identification of treatment failures demand resources not easily available . With the proliferation of mobile phones , data transfer networks and digital microscopy applications the stage is set for alternatives to conventional microscopy in endemic areas . Our aim was to show , as proof of concept , that it is possible to achieve point-of-care diagnostics by an inexpensive mini-microscope for direct visualization on a display and remote diagnostics by computer vision . The results show that parasitic worm eggs can be recognized by on-chip imaging using a webcam stripped off the optics . Images of eggs from the blood fluke S . haematobium present in urine of an infected patient could be interpreted visually and by computer vision . The method offers both an inexpensive alternative to conventional microscopy and diagnostic assistance by computer vision .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2013
|
On-Chip Imaging of Schistosoma haematobium Eggs in Urine for Diagnosis by Computer Vision
|
Domains are modules within proteins that can fold and function independently and are evolutionarily conserved . Here we compared the usage and distribution of protein domain families in the free-living proteomes of Archaea , Bacteria and Eukarya and reconstructed species phylogenies while tracing the history of domain emergence and loss in proteomes . We show that both gains and losses of domains occurred frequently during proteome evolution . The rate of domain discovery increased approximately linearly in evolutionary time . Remarkably , gains generally outnumbered losses and the gain-to-loss ratios were much higher in akaryotes compared to eukaryotes . Functional annotations of domain families revealed that both Archaea and Bacteria gained and lost metabolic capabilities during the course of evolution while Eukarya acquired a number of diverse molecular functions including those involved in extracellular processes , immunological mechanisms , and cell regulation . Results also highlighted significant contemporary sharing of informational enzymes between Archaea and Eukarya and metabolic enzymes between Bacteria and Eukarya . Finally , the analysis provided useful insights into the evolution of species . The archaeal superkingdom appeared first in evolution by gradual loss of ancestral domains , bacterial lineages were the first to gain superkingdom-specific domains , and eukaryotes ( likely ) originated when an expanding proto-eukaryotic stem lineage gained organelles through endosymbiosis of already diversified bacterial lineages . The evolutionary dynamics of domain families in proteomes and the increasing number of domain gains is predicted to redefine the persistence strategies of organisms in superkingdoms , influence the make up of molecular functions , and enhance organismal complexity by the generation of new domain architectures . This dynamics highlights ongoing secondary evolutionary adaptations in akaryotic microbes , especially Archaea .
Proteins are biologically active molecules that perform a wide variety of functions in cells . They are involved in catalytic activities ( e . g . enzymes ) , cell-to-cell signaling ( hormones ) , immune response initiation against invading pathogens ( antibodies ) , decoding genetic information ( transcription and translation machinery ) , and many other vital cellular processes ( receptors , transporters , transcription factors ) . Proteins carry out these functions with the help of well-packed structural units referred to as domains . Domains are modules within proteins that can fold and function independently and are evolutionarily conserved [1]–[4] . It is the domain make up of the cell that defines its molecular activities and leads to interesting evolutionary dynamics [5] . Different mechanisms have been described to explain the evolution of domain repertoires in cells [3] . These include the reuse of existing domains [2] , [6] , interplay between gains and losses [7]–[9] , de novo domain generation [1] , and horizontal gene transfer ( HGT ) [10] . Domains that appeared early in evolution are generally more abundant than recently emerged domains and can be reused in different combinations in proteins . This recruitment of ancient domains is an ongoing evolutionary process that leads to the generation of novel domain architectures ( i . e . ordering of domains in proteins ) by gene fusion , exon recombination and retrotransposition [2]–[4] , [11] . For example , aminoacyl-tRNA synthetases are enzymes that charge tRNAs with ‘correct’ amino acids during translation [12] , [13] . These crucial enzymes are multidomain proteins that encode a catalytic domain , an anticodon-binding domain , and in some cases , accessory domains involved in RNA binding and editing [13] . Evolutionary analysis suggests that these domains were recruited gradually over time [14] . In fact , recruitment of ancient domains to perform new functions is a recurrent phenomenon in metabolism [15] . In addition to the frequent reuse of domains , the dynamics between gains and losses also impacts the evolution of proteome repertoires [7]–[9] . Previous studies identified high rates of gene gains and losses in 12 closely related strains of Drosophila [7] , Prochlorococcus ( a genus of cyanobacteria ) [16] , and 60 isolates of Burkholderia ( a genus of proteobacteria ) [17] . A recent analysis of Pfam domains [18] revealed that ∼3% of the domain sequences were unique to primates and had emerged quite recently [19] . This implies that emergence of novel domains is an incessant evolutionary process [1] . In contrast , different selective pressures can lead to loss of domains in certain lineages and trigger major evolutionary transitions . For example , the increased rate of domain loss has been linked to reductive evolution of the proteomes of the archaeal superkingdom [20] , adaptation to parasitism in cells [21] ( e . g . transition from the free-living lifestyle to obligate parasitism in Rickettsia [22] ) , and ‘de-evolution’ of animals [23] , [24] from their common ancestor . In these studies , gain and loss inferences were restricted to only particular groups of phyla or organisms . A global analysis involving proteomes from the three superkingdoms remained a challenge . Finally , changes to domain repertoires are also possible by HGT that is believed to occur with high frequency in microbial species , especially Bacteria [25] , [26] . Here , we describe the evolutionary dynamics of protein domains grouped into fold families ( FFs ) and model the effects of domain gain and loss in the proteomes of 420 free-living organisms that have been fully sequenced and were carefully sampled from Archaea , Bacteria , and Eukarya ( Dataset S1 ) . The 420-proteome dataset was previously used by our group to reconstruct the evolutionary history of free-living organisms ( see [27] ) and was updated here to account for recent changes in protein classification and functional annotation . The dataset is very well annotated , especially regarding organism lifestyles that are otherwise problematic to assign , has already produced patterns of protein and proteome evolution that are very useful ( including those described in [27] ) , and has produced timelines of FF evolution that are being actively mined . We conducted phylogenomic analyses using the abundance ( total redundant number of each FF in every proteome ) [28] , [29] and occurrence ( presence or absence ) [30] , [31] counts of FFs as phylogenetic characters to distinguish the 420 sampled taxa ( i . e . proteomes ) . FF information was retrieved from the Structural Classification of Proteins ( SCOP ) database , which is considered a ‘gold standard’ for the classification of protein domains into different hierarchical levels [32] . Current SCOP definitions group protein domains with high pair-wise sequence identity ( >30% ) into a common FF , FFs that are evolutionarily related into fold superfamilies ( FSFs ) , FSFs with similar secondary structure arrangement into folds ( Fs ) , and Fs with common secondary structure elements into a handful of protein classes [33] , [34] . A total of 110 , 800 SCOP domains ( ver . 1 . 75 ) are classified into a finite set of only 1 , 195 Fs , 1 , 962 FSFs and 3 , 902 FFs . The lower number of distinct FSFs and FFs suggests that domain structure is far more conserved than molecular sequence ( e . g . see [35] ) and is reliable for phylogenetic studies involving the systematic comparison of proteomes [27] . Another advantage of using SCOP domains is the consideration of known structural and inferred evolutionary relationships in classifying domains into FFs and FSFs [36] . In comparison , evolutionary relationships for the majority of the Pfam domains are unknown . We further restricted the analysis to include only FF domains as they are conserved enough to explore both the very deep and derived branches of the tree of life ( ToL ) and are functionally orthologous [37] . In contrast , FSF domains represent a higher level in SCOP hierarchy and are more conserved than FFs but may or may not be functionally orthologous . Moreover , high conservation of FSF domains is useful for exploring the deep branches of the ToL but may not be very informative for the more derived relationships . The analysis of retracing the history of changes in the occurrence and abundance of FF domains on each branch of the reconstructed ToLs revealed that FFs were subject to high rates of gains and losses . Domain gains generally outnumbered losses but both occurred with high frequencies throughout the evolutionary timeline and in all superkingdoms . Remarkably , the gains-to-loss ratios increased with evolutionary time and were relatively higher in the late evolutionary periods . Finally , functional annotations of FFs illustrated significant differences between superkingdoms and described modern tendencies in proteomes .
The 420-proteome dataset used in this study included proteomes from 48 Archaea , 239 Bacteria , and 133 Eukarya . The dataset did not include any parasitic organisms as they harbor reduced proteomes and bias the global phylogenomic analyses ( e . g . [38] ) . FFs were assigned to proteomes using SUPERFAMILY ver . 1 . 73 [39] hidden Markov models [40] , [41] at an E-value cutoff of 10−4 [42] . A total of 2 , 397 significant FF domains were detected in the sampled proteomes . The definitions of eight FFs in the 420-proteome dataset were updated in SCOP ver . 1 . 75 and were therefore renamed in our dataset . FFs were referenced using SCOP concise classification strings ( css ) ( e . g . ‘Ferredoxin reductase FAD-binding domain-like’ FF is b . 43 . 4 . 2 , where b represents the class [all-beta proteins] , 43 the fold , 4 the FSF and 2 the FF ) . We considered the genomic abundance [28] , [29] and occurrence [30] , [31] of 2 , 397 FFs as phylogenetic characters to reconstruct phylogenies describing the evolution of 420 free-living organisms ( i . e . taxa ) using maximum parsimony . The raw abundance values of each FF in every proteome ( gab ) were log-transformed and divided by the logarithm of maximum value in the matrix ( gmax ) to account for unequal proteome sizes and variances ( see formula below ) [29] , [43] . The transformed abundance values were then rescaled from 0 to 23 ( scaling constant ) in an alphanumeric format ( 0–9 and A-N ) to allow compatibility with the phylogenetic reconstruction software . The transformed abundance matrix with 24 possible character states was imported into PAUP* 4 . 0b10 [44] for the reconstruction of abundance trees . For occurrence trees , we simply used 0 and 1 ( indicating absence and presence ) as the valid character state symbols . We polarized both abundance and occurrence trees using the ANCSTATES command in PAUP* and designated character state 0 as the ancestral state , since the most ancient proteome is closer to a simple progenote organism that harbors only a handful of domains [20] , [38] . The stem lineage of this organism gradually increased its domain repertoire , supporting the polarization from 0 to N and Weston's generality criterion , in which the taxic distribution of a set of character states is a subset of the distribution of another [45] , [46] . Phylogenetic trees are adequately interpreted when rooted . This provides direction to the flow of evolutionary information and is useful to study species adaptations . In this study , we choose to root trees using the Lundberg method [47] . This scheme first determines the most parsimonious unrooted tree , which is then attached to a hypothetical ancestor . The hypothetical ancestor may be attached to any of the branches in the tree . However , only the branch that gives the minimum increase in overall tree length is selected [48] . This branch , which exhibits the largest numbers of ancestral ( plesiomorphic ) character states was specified using the ANCSTATES command in PAUP* . Thus , Lundberg rooting automatically roots the trees by preserving the principle of maximum parsimony . This method is simple and free from artificial biases introduced by alternative rooting methods ( e . g . the outgroup method ) . While selection of an appropriate outgroup to root the ToL is virtually impossible , Lundberg rooting provides a parsimonious estimate of the overall phylogeny and should be considered robust as long as the assumptions used to root the trees are not proven false . To evaluate support for the deep branches of ToLs , we ran bootstrap ( BS ) analysis with 1 , 000 replicates . Character state changes were recorded by specifying the ‘chglist’ option in PAUP* . Trees were visualized using Dendroscope ver . 3 . 0 . 14b [49] . To determine congruence between abundance and occurrence trees , we used the nodal module implemented in the TOPD/FMTS package ver . 3 . 3 [50] . The module takes as input a set of trees in Newick format and calculates a root mean squared deviation ( RMSD ) value for each pairwise comparison . The RMSD value is 0 for identical trees and increases with incongruence . To evaluate the significance of calculated RMSD values , we implemented the ‘Guided randomization test’ with 100 replications to determine whether the calculated RMSD value was smaller than the chance expectation . The randomization test randomly changes the positions of taxa in trees , while maintaining original tree topology , and calculates an RMSD value for each random comparison [50] . The result is a random distribution of RMSD values with a mean and standard deviation . The calculated RMSD value was compared with the mean of the random distribution to determine whether the observed differences were better than what would be expected merely by chance . The spread of each FF was given by its distribution index ( f-value ) , defined by the total number of proteomes encoding a particular FF divided by the total number of proteomes . The f-value ranges from 0 ( absence from all proteomes ) to 1 ( complete presence ) . To determine the relative age of FF domains in our dataset , we reconstructed trees of domains ( ToDs ) from the abundance and occurrence matrices used in the reconstruction of ToLs . The matrices were transposed , treating FFs as taxa and proteomes as characters . The reconstructed ToDs described the evolution of domains grouped into FFs and identified the most ancient and derived FFs ( refer to [27] for an elaborate description and discussion on ToDs ) . To root the trees , we declared character state ‘N’ as the most ancestral state . This axiom of polarization considers that history of change for the most part obeys the ‘principle of spatiotemporal continuity’ ( sensu Leibnitz ) that supports the existence of Darwinian evolution . Specifically , it considers that abundance and diversity of individual FFs increases progressively in nature by gene duplication ( and associated processes of subfunctionalization and neofunctionalization ) and de novo gene creation , even in the presence of loss , lateral transfer or evolutionary constraints in individual lineages . Consequently , ancient domains have more time to accumulate and increase their abundance in proteomes . In comparison , domains originating recently are less popular and are specific to fewer lineages . We note that the N to 0 polarization is supported by the observation that FFs that appear at the base of the ToDs are structures that are widespread in metabolism and are considered to be of very ancient origin ( e . g . [27] ) . The age of each FF was drawn directly from the ToDs using a PERL script that calculates the distance of each node from the root . This node distance ( nd ) is given on a relative scale and portrays the origin of FFs from 0 ( most ancient ) to 1 ( most recent ) . The geological ages of FFs were derived from a molecular clock of protein folds [51] , [52] that was used to calibrate important events in proteome evolution . We have previously shown that nd correlates with geological time , following a molecular clock that can be used as a reliable approximation to date the appearance of protein domains [51] , [52] . We used the SUPERFAMILY functional annotation scheme ( based on SCOP 1 . 73 ) to study the functional roles of FF domains in our dataset [53]–[55] . The SUPERFAMILY annotation assigns a single molecular function to FSF domains ( and by extension to its descendant FFs ) . The annotation scheme gives a simplified view of the functional repertoire of proteomes using seven major functional categories including , i ) metabolism , ii ) information , iii ) intracellular processes , iv ) extracellular processes , v ) general , vi ) regulation and vii ) other ( includes domains with either unknown or viral functions ) . We assumed that FFs grouped into an FSF performed the same function that was assigned to their parent FSF . While this simplistic representation does not demonstrate the complete functional capabilities of a cell , it is sufficient to illustrate the major functional preferences in proteomes ( refer to [21] for further description and use of the functional annotation scheme in large-scale proteomic studies ) . We conducted a GO enrichment analysis [56] , [57] on FF domains to identify biological processes [58] , [59] that were significantly enriched . For this purpose , the list of FF domains was given as input to domain-centric Gene Ontology ( dcGO; http://supfam . org/SUPERFAMILY/dcGO ) resource and the most specific and significant associations to GO terms corresponding to different biological processes were retrieved . The statistical significance was evaluated by P-value computed under the hypergeometric distribution [56] , while the false discovery rate ( FDR ) was set to default at <0 . 01 [60] .
A Venn diagram describes the sharing patterns of 2 , 397 FFs in seven Venn distribution groups ( Figure 1A ) . For simplicity , we name these sets ‘taxonomic groups’ with the understanding that their taxonomic status is endowed by patterns of distribution of FFs in superkingdoms . The number of FFs decreased in the order Eukarya ( total FFs = 1 , 696 ) , Bacteria ( 1 , 510 ) and Archaea ( 703 ) . Eukarya also had the highest number of superkingdom specific FFs ( 758 ) , followed by Bacteria ( 522 ) , and Archaea ( 89 ) . ABE FFs were universal ( i . e . present in all three superkingdoms ) and made the third largest group with 484 FFs , while BE was the fourth largest taxonomic group with 414 FFs ( Figure 1A ) . The lowest number of FFs was in AE with only 40 FFs that were unique to both Archaea and Eukarya . The number of Archaea-specific FFs was also low ( 89 ) but comparable to the number of akaryotic FFs ( i . e . AB = 90 ) . We observed that Archaea was mostly about sharing ( or not innovating new FFs ) . This was evident by the fact that only 13% of the total archaeal FFs were Archaea-specific . This was in striking contrast with Bacteria and Eukarya where superkingdom-specific FFs made large proportions of the FF repertoires with 35% and 45% FFs , respectively ( Figure 1A ) . We plotted the distribution of domain ages ( nd ) for FFs in each taxonomic group to determine the order of their evolutionary appearance ( Figure 1B ) ( see Methods ) . The first FF to appear in evolution was the ‘ABC transporter ATPase domain-like’ ( c . 37 . 1 . 12 ) FF at nd = 0 in the ABE taxonomic group ( Figure 1B ) . ABC transporters are multifunctional proteins that are primarily involved in the transport of various substrates across membranes [61] , [62] . These domains are ubiquitous and highly abundant in extant species and considered to be very ancient . In our timeline , c . 37 . 1 . 12 appeared first , supporting its widespread presence and significance in cells . ABE was the most ancient taxonomic group spanning the entire time axis with a median nd of 0 . 24 ( Figure 1B ) . This suggested that the majority of the FFs that were common across all superkingdoms appeared very early in evolution . ABE was followed by the appearances of BE ( at nd = 0 . 15 ) , AB ( 0 . 26 ) , B ( 0 . 26 ) , E ( 0 . 551 ) , A ( 0 . 555 ) , and AE ( 0 . 57 ) taxonomic groups , in that order ( Figure 1B ) . The first complete loss event for any FF in the primordial world likely triggered the appearance of the BE taxonomic group . Our data indicates that this occurred at nd = 0 . 15 ( roughly >3 . 2 billion [Gyrs] years ago ) with the complete loss of the ‘Heat shock protein 90 , HSP90 , N-terminal domain’ ( d . 122 . 1 . 1 ) FF in Archaea ( Figure 1B ) . Heat-shock proteins are molecular chaperones that assist in protein folding and clearing of cell debris [63] . These are highly conserved in bacterial and eukaryal species , but relatively less abundant in Archaea . In fact , homologs of Hsp90 or Hsp100 are completely absent in archaeal species [63] . This knowledge is compatible with our finding of loss of d . 122 . 1 . 1 FF in Archaea that occurred very early in evolution . We propose that this event exemplifies reductive evolutionary processes that were at play early in evolution in nascent archaeal lineages as emergent diversified cells were unfolding different mechanisms of protein folding . In light of our results , Archaea was the first superkingdom to follow reductive trends . The first superkingdom-specific FF appeared in B at nd = 0 . 26 ( ∼2 . 8 Gyrs ago ) , while both Archaea and Eukarya acquired unique FF domains concurrently at around nd = 0 . 55 ( ∼1 . 6 Gyrs ago ) ( Figure 1B ) . Emergence of taxonomic groups in evolution described three important evolutionary epochs: ( i ) early ( 0≤nd<0 . 15 ) , a period before the start of reductive evolution in the archaeal superkingdom , ( ii ) intermediate ( 0 . 15≤nd<0 . 55 ) , a period marked by early domain discovery in Bacteria , and ( iii ) late ( 0 . 55≤nd≤1 ) , a period during which simultaneous diversification of Archaea and Eukarya occurred ( Figure 1B ) . To determine the popularity of FFs across organisms , we computed an f-value representing the fraction of proteomes encoding an FF . The median f-value decreased in the order , ABE>AE>E>BE>AB>A>B ( Figure 1C ) . We observed that universal FFs of the ABE taxonomic group were most popular and shared by the majority of the proteomes ( median f = 0 . 58 ) . The FFs in AE and E were also distributed with higher f-values ( median f = 0 . 54 and 0 . 27 ) . In contrast , most of the bacterial taxonomic groups ( e . g . BE , AB and B ) had lower median f-values ( 0 . 22 , 0 . 10 , and 0 . 02 , respectively ) . The Venn diagram indicated that ∼22% of the total FFs were bacteria-specific ( Figure 1A ) but the median f-value of those FFs was extremely low ( 0 . 02 ) ( Figure 1C ) . This implied that FFs unique to Bacteria were very unevenly distributed among bacterial species . This also suggested that the rate of FF discovery in Bacteria was very high but their spread was quite limited . A recent study proposed concepts of economy ( i . e . organism budget in terms of number of unique genes and domain structures ) , flexibility ( potential of an organism to adapt to environmental change ) and robustness ( ability to resist damage and change ) to help explain the persistence strategies utilized by organisms in the three superkingdoms [64] . To determine how persistence strategies distributed in our dataset , we redefined economy ( i . e . total number of unique FFs in a proteome ) , flexibility ( total number of redundant FFs in a proteome ) and robustness ( ratio of flexibility to economy ) . When plotted together on a 3D plot , interesting patterns were revealed ( Figure 1D ) . As expected , the proteomes of the akaryotic microbes in Archaea and Bacteria were most economical but least flexible and robust ( Figure 1D ) . Within these superkingdoms , archaeal proteomes ( red circles ) exhibited greatest economy but lowest flexibility and robustness . In contrast , Bacteria exhibited intermediate levels of economy , flexibility and robustness . Finally , eukaryal proteomes were least economical but highly flexible and robust ( Figure 1D ) . Table 1 lists the lower and upper bounds for economy , flexibility , and robustness for the three superkingdoms . The median values for the three parameters always increased in the order , Archaea , Bacteria , and Eukarya ( Table 1 ) . The analysis revealed that the survival strategy of microbial species lies in encoding smaller domain repertoires while the eukaryal species trade-off economy with more flexibility and robustness and harbor richer proteomes [64] . The number of both unique ( economy ) and redundant FFs ( flexibility and robustness ) was considerably higher in eukaryotes . We compared the distributions of molecular functions in taxonomic groups ( Figure 2A ) and dated their appearance in evolutionary time ( nd ) ( Figure 2B–H ) . Metabolism was the most abundant and widely distributed molecular function in organisms , especially in the ABE , BE , and AB taxonomic groups . However , significant deviations were observed in the AE and A taxonomic groups , where informational FFs ( e . g . those belonging to the replication machinery ) outnumbered FFs in other functional categories ( Figure 2A ) . These results are consistent with previous knowledge regarding high sharing of informational proteins between Archaea and Eukarya and a common metabolic apparatus between Bacteria and Eukarya . This observation has often led to proposals relating the origin of eukaryotes to a confluence between akaryotic cells ( reviewed in [65]; see also [66]–[69] ) . However , our data show that the presence of bacterial metabolic enzymes in Eukarya is better explained by primordial endosymbiotic events leading to mitochondria and plastids in a proto-eukaryote stem cell-line ( read below ) . In comparison , sharing of informational enzymes between Archaea and Eukarya occurred relatively late in evolution and could actually reflect late domain losses in Bacteria . Intracellular processes and general were distributed similarly while regulation and extracellular processes appeared to be preferential only in Eukarya ( Figure 2A ) . The distribution of molecular functions in taxonomic groups was largely in agreement with the distribution previously explained for individual species [21] . We explored the order of evolutionary appearance of molecular functions by generating nd vs . f plots for the seven taxonomic groups ( Figure 2B–H ) . The ABE FFs were present with largest f-values and as expected spanned the entire nd-axis ( Figure 2B ) . In fact , 13 FFs had an f-value of 1 . 0 indicating universal presence in organisms , while 62 near-universal FFs were present in >95% of the proteomes . ABE FFs were generally enriched in metabolic functions ( Figure 2B ) . This suggested that the last common ancestor of diversified life was structurally and metabolically versatile ( e . g . [38] ) . However , the f-value distribution of ABE FFs followed a bimodal pattern with a significant drop in f during the intermediate evolutionary epoch . Most of the FFs of intermediate age were classified as metabolic ( grey circles ) , informational ( red circles ) , or with intracellular roles ( light blue circles ) ( Figure 2A , 2B ) . BE followed a distribution similar to ABE but the first FF appeared during the intermediate evolutionary epoch at nd = 0 . 15 ( Figure 2C ) . This also marked the first loss of an FF in Archaea ( boxplot for BE in Figure 1B ) . This observation implies that Archaea was the first superkingdom to escape from the ancestral community and evolved by streamlining genomes . Perhaps , genome reduction was better suited for harsher environments . Other selective pressures that may have triggered early domain loss in Archaea could include escape from RNA viruses ( because RNA is unstable at extreme temperatures ) and phagotrophs [70] . The majority of the BE FFs served metabolic , informational and intracellular roles ( Figure 2A , 2C ) , just like ABE . The akaryotic-specific ( AB ) FFs appeared during the intermediate and late evolutionary epochs and were largely dominated by metabolic and other FFs ( Figure 2A , 2D ) . Most of these FFs had very low f-values ( Figure 2D ) indicating that this taxonomic group exhibited low popularity levels . In contrast , all of the 40 AE FFs appeared in the late epoch and were dominated by domains involved in informational ( red ) ( Table 2 ) and regulatory processes ( green ) ( Figure 2A , 2E ) . This validated the hypothesis that informational enzymes in eukaryotes are very similar to their archaeal counterparts rather than bacterial enzymes [71]–[73] . This argument has been used to propose a sister relationship between Archaea and Eukarya and an ancient origin of Bacteria . However , our analysis revealed that sharing of informational domains between archaeal and eukaryal species was only a recent event ( i . e . was evident in the late evolutionary epoch; nd≥0 . 55 ) and that the sister relationship between Archaea and Eukarya inferred from the 16S rRNA trees [74] was influenced by the high rates of modern sharing between Archaea and Eukarya ( see Discussion ) [75] . AE FFs were generally distributed with higher f-values ( Figure 2E ) . FFs unique to Archaea ( A ) appeared in the late epoch at nd≥0 . 55 and were generally distributed with lower f-values ( Figure 2F ) . The discoveries of these FFs were biased towards informational and other domains ( Figure 2A , 2F ) . A large number of bacteria-specific FFs ( B ) also appeared during the intermediate and late evolutionary epochs ( Figure 2G ) . We note that , in general , bacterial FFs appearing in the intermediate epoch were biased towards informational roles while those that appeared later served metabolic and general roles ( Figure 2A , 2G ) . Lastly , all of the Eukarya-specific ( E ) FFs appeared in the late epoch ( Figure 2H ) , just like Archaea ( Figure 2F ) . Eukarya discovered a large number of recent FF domains ( nd≥0 . 55 ) that were involved in regulation ( green circles ) and extracellular processes ( blue circles ) and were distributed with relatively high f-values in the eukaryal proteomes ( Figure 2A , 2H ) . Superkingdom-specific FFs appeared in both Archaea and Eukarya at around the same time , and both showed a tendency to become widespread in species ( Figure 2F , 2H ) . In contrast , the discovery of Bacteria-specific ( B ) FFs started much earlier but with limited spread ( Figure 2G ) . This suggested that while Archaea was the first superkingdom to follow reductive trends , it was Bacteria that diversified first and was capable of unfolding superkingdom-specific domain structures . The primordial stem-line ( that was structurally and functionally complex ) later evolved into eukaryotes , possibly after engulfment of already diversified microbes ( Discussion ) . In this regard , we identified a set of mitochondrial FFs , all of which appeared at nd≥0 . 55 , during and after the rise of the E taxonomic group , including the ‘Mitochondrial resolvase ydc2 catalytic domain’ ( c . 55 . 3 . 7; nd = 0 . 55 ) and the ‘Mitochondrial cytochrome c oxidase subunit VIIb’ ( f . 23 . 5 . 1; nd = 0 . 59 ) FFs ( Table 3 ) . Thus , our timelines do not support fusion hypotheses for the origin of eukaryotes linked to a confluence between akaryotes . The fusion scenarios have been discussed elsewhere [65] , [70] , [76]–[79] and it is beyond the scope of this study to evaluate what model is better . In light of our data that is based on the genomic census of conserved FF domains in hundreds of free-living organisms , we support a phagotrophic and eukaryote-like nature of the host ( anticipated in [78] , [79] ) that acquired the primordial alpha-proteobacterium as an endosymbiont , which later became mitochondria and triggered the diversification of eukaryotes ( at nd = 0 . 55; roughly ∼1 . 6 billion years ago ) . A formal test of this hypothesis is warranted and will be explored in a future study . The exercise also revealed that the lower median f-values observed earlier ( Figure 1C ) were due to the significant drop in f in the intermediate evolutionary epoch . We note that the majority of the bacterial FFs ( belonging to the ABE , BE , B and AB taxonomic groups ) also appeared during this period and thus affected the overall medians . We generated rooted ToLs from abundance ( Figure 3A ) and occurrence ( Figure 3B ) counts of 2 , 397 FF domains in the 420 free-living proteomes ( see Dataset S1 for taxon names ) using maximum parsimony as the optimality criterion in PAUP* 4 . 0b10 [44] . Both reconstructions recovered a previously established tripartite world of cellular organisms [20] , [27] , [74] , [80] . The archaeal superkingdom always formed a paraphyletic group at the base of the ToLs . The deep branches of the ToLs were occupied by thermophilic and hyperthermophilic archaeal species ( Thermofilum pendens and Cand . Korarchaeum ) ( Figure 3 ) . The archaeal rooting of the ToL is supported by a number of previous studies ( e . g . [14] , [20] , [27] , [81]–[83] ) and is in conflict with the traditional Archaea-Eukarya sister relationship ( Discussion ) . Bacteria and Eukarya formed strong monophyletic clades that were supported by high BS values ( ≥99% ) and were separated from Archaea with 53% ( Figure 3A ) and 78% ( Figure 3B ) BS support . Both ToLs had strong phylogenetic signal ( g1 = −0 . 33 and −0 . 28 ) . Overall , phylogenomic patterns resembled traditional groupings and supported previous analyses of similar kind [20] , [27] . Moreover , the dissimilarity between two reconstructions was 5 . 37 , which was smaller than the mean RMSD calculated from 100 random comparisons ( Figure 3 ) ( Methods ) . Because the ToLs were supported with high confidence and resembled previous analyses [20] , [27] , they made useful tools for the study of domain gain and loss events on the many branches ( read below ) . To quantify the relative contributions of domain gains and losses impacting the evolution of superkingdoms , we retraced the history of character state changes ( i . e . changes in the abundance or occurrence of FFs ) on each branch of the reconstructed ToLs . For each FF domain , we counted the number of times it was gained and lost in different branches of the phylogenetic tree . Gains were recorded when the abundance/occurrence of a particular FF at a node was higher than the corresponding value at the immediate ancestral node . In contrast , losses were incremented when the abundance/occurrence of a particular FF at a node was lower . Because we allowed character changes in both forward and backward directions ( Wagner parsimony ) , each FF character could be both gained and lost a number of times across the many branches of the ToL . This assumption is reasonable as different lineages of organisms utilize domain repertoires differently . Because abundance counts are expected to be higher in the eukaryotic species ( especially in metazoa ) due to increased gene duplication events and a persistence strategy that favors flexibility and robustness ( Figure 1D ) [64] , we also considered gains and loss statistics from the occurrence trees . To evaluate the performance of both models , we first compared the number of FFs that were gained ( i . e . net sum above zero ) and lost ( net sum below zero ) in both reconstructions . Out of the total 2 , 397 ( 2 , 262 parsimony informative ) FF domains in the abundance model , 1 , 955 ( 86% ) were gained , while only 236 ( 10% ) were lost ( Dataset S2 ) . In contrast , occurrence identified 60 . 1% FFs as gained ( 1 , 353/2 , 249 ) and 30 . 5% ( 686/2 , 249 ) as lost ( Dataset S3 ) . Nearly 96% ( 1300/1 , 353 ) of the occurrence gains were also gained in abundance while only 26% ( 178/686 ) losses were common to both models . This suggested that abundance included nearly all the occurrence gains and likely overestimated the number of gains ( due to gene duplications and domain reuse ) . In contrast , occurrence led to more balanced distributions and likely overestimated losses ( read below ) . To provide additional support to the gain/loss model , we pruned taxa from the original ToLs leaving only one superkingdom and recalculated character state changes on the pruned trees . This eliminated any biases resulting from the differences in the persistence strategies of the three superkingdoms and yielded four phylogenetic trees , Total ( taxa = 420 , total FF characters = 2 , 397 ) , Archaea ( 48 , 703 ) , Bacteria ( 239 , 1 , 510 ) and Eukarya ( 133 , 1 , 696 ) . For each of the four trees , we calculated the sum of gain and loss events for all parsimony informative FF characters and represented the values in boxplots ( Figure 4A ) . In all distributions , medians were above 0 indicating that the sum of net gains and losses was a non-negative number for both abundance ( Figure 4A:abundance ) and occurrence ( Figure 4A:occurrence ) models . The exception was the eukaryal tree pruned from the occurrence model , for which the median was exactly zero . The result revealed that while both gains and losses occurred quite frequently , the former was more prevalent in proteome evolution . The histograms in Figure 4B describe the distributions of gain and loss counts for all parsimony informative FF characters in the Total dataset . When plotted against evolutionary time ( nd ) , results highlighted remarkable patterns in the evolution of domain repertoires . Domain gains outnumbered losses in both abundance ( 80 , 904 gains vs . 47 , 848 losses ) and occurrence ( 17 , 319 vs . 13 , 280 ) tree reconstructions ( Figure 4B ) . The gain-to-loss ratios were 1 . 69 and 1 . 30 , respectively , indicating an increase of 69% and 30% in gains relative to losses . Relative differences in the numbers of gains ( red ) versus losses ( blue ) suggested that gains increased with the progression of evolutionary time in both reconstructions ( read below ) . We note that different evolutionary processes may be responsible for shaping the proteomes in individual superkingdoms . For example , the origin of Archaea has been linked to genome reduction events [20] , [84] , while HGT is believed to have played an important role in the evolution of bacterial species [25] . In contrast , eukaryal proteomes harbor an increased number of novel domain architectures that are a result of gene duplication and rearrangement events [6] , [43] . Therefore , to eliminate any biases resulting from the effects of superkingdoms in the global analysis ( Figure 4B ) , we recalculated the history of character changes on the pruned superkingdom tress recovered earlier ( Figure 4C ) . For abundance reconstructions , the exercise supported earlier results where the number of gains was significantly higher than the corresponding number of losses for Archaea ( 4 , 616 vs . 2 , 009 ) , Bacteria ( 36 , 606 vs . 20 , 196 ) , and Eukarya ( 40 , 515 vs . 25 , 036 ) ( Figure 4C: abundance ) . The overall gain to loss ratios decreased from 2 . 30 in Archaea to 1 . 81 in Bacteria and 1 . 62 in Eukarya ( Figure 4C: abundance ) . The increased gain-to-loss ratios in akaryotic microbial species are remarkable; it implies that the rate of gene discovery in akaryotic microbes ( by de novo creation , gene duplication , acquisition by HGT and/or recruitment ) is higher than the rate in eukaryotes . This tendency in microbial species could be a novel ‘collective’ persistence strategy to compensate for their economical proteomes . For histograms representing occurrence models , global gain-to-loss ratios decreased in the order , Archaea>Bacteria>Eukarya ( Figure 4C: occurrence ) . Remarkably , the ratio in Eukarya dropped below 1 indicating prevalence of domain loss events relative to gains . This result supports recent studies that have proposed the evolution of newly emerging eukaryal phyla via genome reduction [85] . When partitioned into the early , intermediate , and late evolutionary epochs , the gain-to-loss ratios exhibited an approximately linear trend towards increasing gains ( Figure 5 ) . For abundance , the ratios increased from 1 . 32 in the early epoch to 1 . 45 in the intermediate and 1 . 96 in the late evolutionary epochs . Similar trends were also observed for occurrence , with calculated ratios of 0 . 61 , 0 . 97 , and 1 . 68 , respectively ( Figure 5A ) . In fact , both gains and losses increased linearly with evolutionary time in all reconstructions . However , accumulation of gains overshadowed the number of losses ( Figure 5 ) . Remarkably , the occurrence model suggested predominant losses in the first two phases of evolution ( 0 . 61 and 0 . 97 ) that were compensated by significantly higher amounts of gains ( 1 . 68 ) in the late epoch . In contrast , abundance failed to illustrate this effect and indicated overwhelming gains in all evolutionary epochs . When looking at the individual epochs for pruned trees ( Figure 5B ) , we noticed that the rate of domain gain increased with time ( as before ) ( Figure 5A ) . However , the ratios in the initial two evolutionary epochs were considerably higher in Archaea for both the abundance and occurrence models . For example , Archaea exhibited gain-to-loss ratios of 2 . 06 and 2 . 14 , in comparison to 1 . 26 and 1 . 39 in Bacteria , and 1 . 55 and 1 . 67 in Eukarya for early and intermediate evolutionary epochs ( Figure 5B:abundance ) . In contrast , Bacteria exhibited an overwhelming gain-to-loss ratio of 2 . 88 in comparison to 2 . 67 in Archaea and 1 . 61 in Eukarya , in the late evolutionary epoch . Overall , the gain-to-loss ratios increased with evolutionary time in all superkingdoms with the sole exception of Eukarya that had a lower ratio in the late ( 1 . 61 ) compared to the intermediate ( 1 . 67 ) epoch ( Figure 5B:abundance ) . Results based on occurrence indicated similar trends but with relatively more balanced gain-to-loss ratios and still highlighted the abundance of domain gains in evolution . The individual ratios were 1 . 42 , 1 . 66 , and 2 . 44 in Archaea , 0 . 60 , 0 . 91 , and 2 . 61 in Bacteria , and 0 . 51 , 0 . 95 , and 0 . 95 in Eukarya ( Figure 5B:occurrence ) . Both Bacteria and Eukarya showed increased levels of ancient domain loss . However , Bacteria compensated this decrease by engaging in massive gain events during the late evolutionary epoch ( ratio of 2 . 61 ) . In contrast , Eukarya exhibited an even exchange between FF gain and loss events ( ratio = 0 . 95 ) in both the intermediate and late epochs . Occurrence results also supported the evolution of Eukarya by gene loss , which is in line with recently published analyses [23] , [85] . Abundance also indicated this drop in gene discovery rate for recent domains in Eukarya . However , the drop appears to be compensated by increased duplications of other domains that lead to an increase in the overall number of domains that are gained ( Figure 5B: abundance ) . This apparent discrepancy can be explained by the power of both models in depicting true evolutionary relationships between organisms . Abundance accounts for a number of evolutionary processes such as HGT , gene duplication , and gene rearrangements while occurrence merely describes presence and absence of FFs and because of its more ‘global’ nature fails to illustrate a complete evolutionary picture ( Discussion ) . To test whether unequal sampling of proteomes per superkingdom was contributing any bias to the calculations of domain gains and losses , we extracted 100 random samples of 34 proteomes each from the three superkingdoms and generated 100 random trees . From each of the random trees , we recalculated the gain-to-loss ratios using both abundance and occurrence models ( Figure 6 ) . Random and equal sampling supported the overall conclusion that gains were overwhelming during the evolution of domain repertoires ( Figure 6 ) . The median ratios for random trees were 2 . 47 in Archaea , 2 . 35 in Eukarya , and 2 . 34 in Bacteria for abundance reconstructions ( Figure 6A ) . In comparison , the ratios decreased from 2 . 11 in Archaea to 1 . 93 in Bacteria and 1 . 11 in Eukarya for occurrence reconstructions ( Figure 6B ) . Based on the results of random and equal sampling , we safely conclude that the gain of domains in proteomes is a universal process that occurs in all three superkingdoms of life . Moreover , the gain-to-loss ratios increase with time ( Figure 5 ) and their effects are directly responsible for evolutionary adaptations in superkingdoms ( Discussion ) . We also propose that using abundance increases the reliability of the phylogenomic model and accounts for many important evolutionary events , a feat that is not possible when studying occurrence . We identified FFs that were gained ( i . e . net sum of gains and losses was above 0 ) and lost ( net sum below 0 ) directly from the pruned superkingdom trees . To eliminate any redundancy , we only kept FFs that were gained ( or lost ) in both abundance and occurrence reconstructions and excluded those where both methods disagreed . Using this stringent criterion , we classified a total of 368 archaeal FFs as being gained and 40 as being lost . In comparison , Bacteria and Eukarya gained 892 and 633 FFs , respectively , while they lost only 148 and 164 FFs . Both gained and lost FFs for each superkingdom were provided as input to the online dcGO resource [56] , [57] to retrieve the highly specific and significantly enriched biological process GO terms ( Methods ) . For FFs that were gained , a total of six GO terms were significantly enriched in archaeal proteomes representing biological processes involved in the biosynthesis of nucleotides and metabolism , such as ‘tricarboxylic acid cycle [GO:0006099]’ , ‘pyruvate metabolic process [GO:0006090]’ , ‘acyl-CoA metabolic process [GO:0006637]’ , ‘thioester biosynthetic process [GO:0035384]’ , ‘purine nucleobase metabolic process [GO:0006144]’ , and ‘pyrimidine nucleoside metabolic process [GO:0006213]’ ( Table 4 ) . In comparison , only one biological process in Bacteria ( ‘polysaccharide catabolic process [GO:0000272]’ ) and 37 in Eukarya were significantly enriched ( Table 4 ) . While , the bacterial GO term corresponded to metabolic roles ( similar to Archaea ) , eukaryal functions encompassed a diverse range of processes including ‘sex determination [GO:0007530]’ , regulatory [GO:0044089] and immunological roles [GO:0046634] , functions related to the development of mammary glands [GO:0061180] , and others ( Table 4 ) . Finally , none of the archaeal or eukaryal lost FFs was significantly associated with any of the highly-specific biological process GO terms , indicating that loss of FFs in these two superkingdoms occurred without any functional constraint . In contrast , two biological processes were predicted to be lost from Bacteria including , ‘cellular modified amino acid biosynthetic process [GO:0042398]’ , and ‘pyrimidine-containing compound biosynthetic process [GO:0072528]’ ( Table 5 ) .
We report the evolutionary dynamics of gain and loss events of protein domain FFs in hundreds of free-living organisms belonging to the three cellular superkingdoms . Structural phylogenomic methods were used to reconstruct ToLs from genomic abundance and occurrence of FF domains in proteomes . Standard character reconstruction techniques were then used to trace domain gain and loss events along the branches of the universal trees . Finally , molecular functions and biological processes of FFs were studied using traditional resources . The exercise revealed remarkable patterns: How reliable is our study ? Both abundance and occurrence were congruent with respect to the overall tree topologies and general conclusions drawn from the analyses . Both supported the existence of overwhelming gains in evolution . However , discrepancies also existed especially in the numerical differences for the gain-to-loss ratios among superkingdoms . In general , abundance ( apparently ) overestimated gains while occurrence underestimated losses . The higher number of gain-to-loss ratios in abundance models is an expected outcome as we are accounting for evolutionary processes such as gene duplications , gene rearrangements , and HGT that are known to increase the representation of genes in genomes . Ancient genes have more time to multiply and increase their genomic abundance compared to newly emergent genes . In contrast , occurrence merely describes the presence or absence of genes and provides a simplified view of the overall landscape of change . Another explanation is the possible existence of methodological artifacts when dealing with genomic occurrence in parsimony analysis that excludes most of the ancient FFs as non-informative characters , when these are present in all proteomes . Moreover , occurrence fails to take into account the weighted contribution of ancient genes to the phylogeny and treats all characters equally . Thus trees built from abundance counts are better resolved at their base while trees built from occurrence behave poorly in this regard [27] . We emphasize that the focus of this study is to highlight the relative contribution of domain gains and losses in the evolution of superkingdoms and not to evaluate which methodology is preferable . The finding that domain gains are overwhelming and increase approximately linearly with evolutionary time in both models is remarkable and suggests that the appearance of novel domains is a continuous process ( Figures 4 and 5 ) . In our phylogenomic model , we rooted ToLs by character absence ( i . e . 0 ) using the Lundberg method . We assumed that proteomes became progressively richer during the course of evolution . However , this implicit assumption did not lead to an increased number of domain gains as character state changes in both forward ( e . g . 9 to 22 ) and reverse ( 12 to 5 ) directions were allowed and carried equal weights . Moreover , we evaluated the effects of ToL rooting on the calculations of domain gain and loss statistics by considering outgroup taxa instead of the Lundberg method . Superkingdom trees rooted with outgroup taxa led to similar tree topologies and supported the conclusion of overwhelming gains that we here report ( Figure S1 ) . However , we decided to exclude outgroup analysis from this study for two reasons . First , outgroups add an external hypothesis into the model and bias gains and losses by including artificial character changes in the most basal branches leading to outgroup taxa . Second , the selection of the most appropriate outgroups for each superkingdom is a complicated problem and is virtually impossible for the reconstruction of ToLs . However , it would be interesting to study the gain and loss dynamics at different levels of the SCOP hierarchy such as the FSF and F levels of structural abstraction . We expect that patterns reported in this study will remain robust regardless of the SCOP conservation level and will extend the analysis to FSF in a separate publication . We used maximum parsimony to search for the best possible tree and described the evolution of 420 free-living proteomes using the entire repertoire of 2 , 397 FFs as phylogenetic characters . We note that parsimony is most appropriate ( and gives superior performance ) for this kind of analysis as it performs better when the characters are evolving under different evolutionary rates [100] . Moreover , rescaling of raw abundance values into 24 possible character states considerably reduces the likelihood of convergent evolution . Reconstructing evolutionary history of species and studying domain emergence and loss patterns is a difficult problem complicated by a number of considerations ( e . g . taxa and character sampling , biases introduced by organism lifestyles , ecological niches of organisms , and non-vertical evolutionary processes ) . We attempted to eliminate these problems by reconstructing whole-genome phylogenies , sampling conserved FF domains as characters , excluding parasitic and facultative parasitic organisms from study , and by using multistate phylogenetic characters . However , we realize that no method is free from technical and logical artifacts . Our analysis largely depends upon the accuracy of phylogenetic reconstruction methods , current SCOP domain definitions , reliability of function annotation schemes , and literature for organism lifestyle . However , we expect that recovered results will remain robust both with data growth and improvement in available methods and that drastic revisions to existing databases would be unlikely . For that reason we caution the reader to focus on the general trends and main conclusions of the paper ( i . e . overwhelming gains and its consequences ) rather than the actual numbers and discrepancies between the phylogenomic methods . Quantifying gain and loss events on a global scale is a difficult problem and our work lays foundations for more and improved studies in the future . We propose that grouping of protein domains into FFs provides a reliable character for a global evolutionary analysis that involves large number of proteomes . FF domains are both sufficiently conserved and informative to explore the many branches on the ToLs . The age and distribution of FFs in organismal groups is biased and carries the power to unfold superkingdom history and explain important structural and functional differences among superkingdoms . Based on our data , we propose the primacy of domain gains over losses over the entire evolutionary period , ongoing evolutionary adaptations in akaryotic microbes , evolution of emerging eukaryotic species by domain loss , an early origin for Archaea , and endosymbiosis leading to mitochondria as a crucial event in eukaryote diversification . Each of these conclusions is important for reconstructing the evolutionary past and predicting evolutionary events in the future .
|
Proteins are made up of well-packed structural units referred to as domains . Domain structure in proteins is responsible for protein function and is evolutionarily conserved . Here we report global patterns of protein domain gain and loss in the three superkingdoms of life . We reconstructed phylogenetic trees using domain fold families as phylogenetic characters and retraced the history of character changes along the many branches of the tree of life . Results revealed that both domain gains and losses were frequent events in the evolution of cells . However , domain gains generally overshadowed the number of losses . This trend was consistent in the three superkingdoms . However , the rate of domain discovery was highest in akaryotic microbes . Domain gains occurred throughout the evolutionary timeline albeit at a non-uniform rate . Our study sheds light into the evolutionary history of living organisms and highlights important ongoing mechanisms that are responsible for secondary evolutionary adaptations in the three superkingdoms of life .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"genomics",
"genome",
"evolution",
"biology",
"computational",
"biology"
] |
2014
|
Global Patterns of Protein Domain Gain and Loss in Superkingdoms
|
Since the beginnings of domestication , the craniofacial architecture of the domestic dog has morphed and radiated to human whims . By beginning to define the genetic underpinnings of breed skull shapes , we can elucidate mechanisms of morphological diversification while presenting a framework for understanding human cephalic disorders . Using intrabreed association mapping with museum specimen measurements , we show that skull shape is regulated by at least five quantitative trait loci ( QTLs ) . Our detailed analysis using whole-genome sequencing uncovers a missense mutation in BMP3 . Validation studies in zebrafish show that Bmp3 function in cranial development is ancient . Our study reveals the causal variant for a canine QTL contributing to a major morphologic trait .
Canine skull shape variation among dog breeds is in large part a human-created phenomenon , occurring through artificial selection and consolidation of desired traits . Morphological distinction between wolves and dogs dates as far back as 31 , 000 years ago [1] , [2] . Changes in skull shape are a key feature of dog domestication , foreshadowing the wide variety of shapes displayed by modern dog breeds . Skull shapes differ tremendously from one another , so much so that such differences are breed-defining . Two such skull shapes are brachycephaly ( “shortened head” , e . g . Bulldog , Pug , Boxer ) and dolichocephaly ( “elongated head” , e . g . Greyhound , Saluki , Collie ) , which are named after their resemblance to human cephalic disorders . Although canine cranial shape is subject to multigenic control [3]–[5] , the molecular underpinnings of this variation remain poorly defined . Candidate gene studies failed to uncover compelling causal variants of canine brachycephaly [6]–[8] . Airorhynchy ( dorsal bending of the snout; a feature common to brachycephalic breeds ) and midface length was previously correlated with polyglutamine and polyalanine repeat length of the transcription factor RUNX2 [4] . More recently , genome wide association scans ( GWAS ) and homozygosity mapping have converged on chromosome 1 ( CFA1 ) as a locus that is highly associated with brachycephaly , implicating a 296 kb haplotype that spans THSB2 and intergenic sequence proximal to SMOC2 [3] , [9] , [10] . Here we present data indicating that at least five genetic loci are responsible for the cranioskeletal differences that differentiate dolichocephalic and brachycephalic dog breeds . Our conclusions are based on a GWAS that coupled craniometric breed-sex averages collected from 533 modern specimens from museum and private collections with the genetic profiles of 576 purebred dogs ( 62 breeds ) assayed via single nucleotide polymorphism ( SNP ) chips . To identify candidates of phenotype causality , we filtered genetic variants derived from whole genome sequencing of eleven different breeds . This led to discovery of a compelling candidate for causality at the CFA32 QTL: a derived missense mutation in BMP3 that is nearly fixed among small , brachycephalic dog breeds . To evaluate the functional potential of this variant in vivo , we turned to zebrafish . We show that Bmp3 is indispensable for normal craniofacial development in zebrafish , and comparison of missexpression assays using BMP3 and its canine variant suggests enhanced activity in the latter . Together , our data reveal for the first time the molecular underpinnings of a quantitative trait , selected by dog fanciers to modulate a prominent morphological trait in domestic dogs .
To capture the three-dimensional morphological complexities present among modern dogs , we digitized 51 stereotyped landmarks from 533 skulls representing 120 breeds and four gray wolf subspecies ( Figure 1A–1D , Figure S1A–S1E , Tables S1 and S2 ) . As most skulls used in our study originated from museums , we selected only those specimens with unambiguous breed designations , sex status , and recent time of death ( within the past 40 years ) for use in this study . Using MorphoJ software [11] , we identified four principal components that accounted for nearly 75 . 5% of shape variance , with the majority of variance explained by the first component ( PC1 = 59 . 4% , PC2 = 8 . 2% , PC3 = 4 . 2% , PC4 = 3 . 8% ) . PC1 describes profound changes in rostrum length and angle , palate and zygomatic arch width , and depth of the neurocranium: essentially the continuum of cranioskeletal features that extend between dolichocephalic and brachycephalic breeds ( Figure 1E , 1F ) . Since purebred dogs must conform to specific morphological standards [12] , morphological traits like skull shape became highly uniform by breed , permitting association studies using one set of samples for genotyping and others for phenotyping . This strategy of using breed stereotypes has proven successful in mapping a number of canine morphologic traits by independent groups [3] , [13] , [14] . Using breed allele frequencies collected by the CanMap project [3] , we conducted genome-wide scans of QTLs associated with breed-sex averages for PC1 ( 1–10 specimen ( s ) /breed/sex , mean n = 3 , Table S3 ) . Initially , we scanned for PC1 associations using an additive linear regression model ( Figure S2A , Table S4 ) [15] . Size correction in the regression suggested potential confounders ( compare Figure S2A and S2B ) on CFA10 and 15 , which were previously associated with body size [3] , [13] , [14] , [16] . As expected , addition of log ( neurocranium centroid ) breed-sex values as a covariate removed those associations ( Figure S2B , see Materials and Methods for more details ) . False associations derived from breed relatedness were excluded using GEMMA [17] . Discounting associations on CFA10 and 15 , we identified six PC1-associated regions of interest indicated by SNPs at CFA1 . 59832965 , CFA5 . 32359028 , CFA24 . 26359293 , CFA30 . 35656568 , CFA32 . 8384767 , and CFAX . 44401786 ( −log10 ( P ) = 6 . 13–17 . 9 , Figure 2A , Table S4 ) . Of note , a suggestive association on CFAX was also observed , marked by SNP CFAX . 104724717 . Including a neurocranium centroid size covariate in the mixed-model removed associations at CFA10 and 15 , as well as those on CFA30 , 32 , and X . 44401786 and enhanced the association on CFAX . 104724717 to significance ( Figure 2B , Table S4 ) . Since nearly all extreme brachycephalic breeds used in our study are also small breeds , and therefore substantially related to small , non-brachycephalic breeds , we reasoned that use of a size covariate in the mixed-model was overcorrecting associations that could be driven by diminutive breeds [3] , [18] , [19] . To reduce the contrast in relatedness among our study population , we reran the mixed-model using only breeds with a log ( neurocranium centroid ) below the 50th percentile . This resulted in recovery of the CFA32 QTL , as well as new associations marked by SNPs at CFA9 . 50988217 and CFA13 . 26492600 . Although the association on CFA30 remained below threshold for statistical significance , its association markedly improved ( Figure 2C ) . When brachycephalic breeds were removed from the mixed-model , all aforementioned markers dropped below significance except for CFA5 . 36476657 ( Figure 2D ) . Summarizing these findings , QTLs on CFA1 , 5 , 24 , 32 , and X ( X:104724717 ) account for skull shape changes that occur along the continuum of canine brachycephaly-dolichocephaly . Additional associations reside on CFA9 , CFA13 , CFA30 , and CFAX ( X: 44401786 ) , though their instability across mixed-model scans suggest they are either allometric in nature , driven by variation that is marginally represented by the breed composition present in our GWAS , or possibly false positives ( Figure 2A–2D , Table S4 ) . Because shape variation is the result of artificial selection , we expected critical loci to be marked by reduction of observed heterozygosity ( Ho ) and elevated genetic differentiation ( FST ) , hallmarks of selective sweeps [20] , [21] . Among autosomes , QTLs on CFA1 , 5 , 30 , and 32 displayed particularly strong reductions in HO among brachycephalic breeds , relative to dolichocephalic breeds ( HR , see Materials and Methods ) . Sliding HR windows corresponding to these QTLs placed with the smallest <0 . 2% of the distribution . Among sliding window FST averages , windows corresponding to CFA1 , 5 , 24 , 30 , and 32 placed within the top 99 . 6% of the distribution ( Figure 2E–2F , Figure S3A–S3I , and Table S5 ) . We focused on the CFA32 QTL because it was the second most highly associated , non-allometric locus in our initial analysis ( Figure S2A ) , it showed compelling evidence of selection , and unlike the CFA1 QTL , it was previously unexplored [9] . Haplotype sharing at this locus among six of the seven extreme brachycephalic breeds , including the French Bulldog , Bulldog , Boston Terrier , Pekingese , Pug , and Brussels Griffon defined a critical interval spanning 190 kb ( 8 , 152 , 258–8 , 342 , 370 , Table S6 ) . Although this region included just two genes , both were excellent candidates: cGMP-dependent protein kinase 2 ( PRKG2 ) and bone morphogenetic protein 3 ( BMP3 ) [22]–[27] . To identify variants within the critical interval , we used whole-genome sequence analysis from eleven dog breeds of widely varying skull shapes ( unpublished data ) . Notably , brachycephalic breeds including a Pekingese and a Bulldog were among the eleven , enabling the evaluation of phenotype association with genotype at nearly every position . Initial examination of variant calls in the 190 kb critical interval revealed over 2 , 000 polymorphisms ( Table S7 ) . Of particular interest , allelic differences between the Bulldog and Pekingese extended downstream of 8 , 237 , 936 , suggesting a recombination breakpoint in the Pekingese . Confirmation of this breakpoint among 25 additional Pekingese reduced the critical interval to 85 kb ( 8 , 152 , 258–8 , 237 , 937 kb ) ( Figure 3A , Table S8 ) . Within the 85 kb interval , 48 variants that met one or more standard criteria were retained for further evaluation ( Figure 3B ) . Only one variant remained a compelling candidate for causality: a SNP at 8 , 196 , 098 that encodes a missense mutation in BMP3 , changing a phenylalanine to a leucine ( BMP3F452L or F452L ) . The Protein Specific Scoring Matrix ( PSSM ) for TGF-β superfamily members indicates that position 452 is nearly invariably occupied by an aromatic amino acid such as tyrosine or phenylalanine ( PSSM raw frequency = 0 . 84 ) and PolyPhen-2 substitution modeling predicted that the F452L substitution is likely damaging ( HumDiv = 1 . 0 , HumVar = 0 . 97 ) [28] . Moreover , F452 flanks highly conserved residues predicted to reside at the receptor-ligand interface [29] . Finally , expanded genotyping among 842 dogs from 113 breeds revealed that the BMP3F452L mutation is nearly fixed among extreme brachycephalic breeds . Furthermore , the PC1 scores of most carrier breeds fall between wolves ( ancestral ) and extreme brachycephalic breeds ( Table S9 ) . BMP3's role in cranioskeletal development is enigmatic in terms of molecular interactions and function . BMP3 antagonizes other BMPs and Activins through binding the ActRIIb receptor , and in vivo , BMP3 appears to restrict bone growth [23] , [30] , [31] . However , the absence of a knockout mouse craniofacial phenotype suggested that BMP3 function might be subtle , dispensable , or divergent to other mammals . We therefore assayed BMP3 function using the zebrafish model . Based on peptide similarity and synteny to CFA32 ( 96 . 4% identical within mature protein , 60 . 5% overall ) , the BMP3 ortholog was identified on zebrafish chromosome 5 . Endogenous expression of zebrafish bmp3 is highly dynamic , first appearing during mid-somitogenesis as ubiquitous expression throughout the head , brain ventricles , and as was shown previously , the posterior somites ( data not shown ) [32] . After 48 hours post fertilization ( hpf ) , bmp3 expression emerges in pectoral fins , the pharyngeal arch region , heart , and jaw structures ( Figure 4A–4D , data not shown ) . Prechondrogenic expression of bmp3 among cranial structures suggests a role for Bmp3 in cranioskeletal development . To formally test this hypothesis , we knocked down endogenous Bmp3 activity via injection of translation-blocking antisense morpholino oligonucleotides ( MO ) . Strikingly , MO-injected embryos demonstrated severe deficiencies in jaw development ( Figure 4E , 4H , 4K ) . Alcian blue staining revealed loss or hypoplasia of multiple cartilage elements that form the viscerocranium and neurocranium ( Figure 4F , 4G , 4I , 4J , 4L , 4M ) . Cartilage defects are specific to loss of Bmp3 activity since injection of two non-overlapping MOs produced identical craniofacial phenotypes , as did co-injection of both MOs at concentrations insufficient to cause phenotypes when injected alone ( data not shown ) . These results indicate that Bmp3 is required for zebrafish craniofacial development , and indicate that Bmp3's role in craniofacial development is ancient . Furthermore , overexpression assays using BMP3 , as well as other TGFβs , indicate that variation at the F452L residue has context-dependent effects on these molecules' activities ( Figures S4 , S5 ) .
Distortion of the skull , as observed among brachycephalic and dolichocephalic dog breeds , affects bones presumably derived from endochondral and intramembraneous ossification . We show that the genetic basis of this distortion is complex , relying on the contributions of at least five QTLs . We propose that the BMP3F452L variant was selected by dog fanciers for its influence on skull shape , but the specific aspects of cranioskeletal development that the F452L variant affects within the brachycephalic skull remain unclear . Previous studies , as well as ours , indicate that the CFA1 QTL is highly associated with canine brachycephaly and is robust to size-stratified GWAS ( Figure 2A and 2C , data not shown ) , suggesting that the underlying causal variant at this locus is shared by both large and small brachycephalic breeds [3] , [9] . Homozygosity mapping also implicated selective sweeps on CFA1 , as well as CFA26 , among Boxers , French Bulldogs , and Bulldogs [10] . Despite different morphometric approaches , skulls specimens , and utilization of CanMap genotypes profiles , our QTLs overlap with those reported by Boyko et al . for snout length ( CFA1 , 5 , 32 , X ) , cranial vault depth ( CFAX ) , palate width ( CFA30 ) , and zygomatic arch width ( CFA24 ) [3] . The associations that we report on CFA9 and 13 were revealed following size-stratified scans , raising caution regarding the implementation of mixed-model scans among domesticated populations whose traits and relatedness are difficult to disentangle . Notably , a snout ratio QTL on CFA9 was previously reported by Jones et al . in a study that also used breed stereotypes as phenotypes; our data independently replicates their finding [14] . We chose the zebrafish model to validate our GWAS results based on its rapid development , gene conservation , and flexibility for rapidly knocking down and overexpressing gene products of interest . Though loss-of-function using zebrafish indicates an ancient role for Bmp3 during craniofacial development , ontogenetic differences between teleost and amniote cranial development limit the extent to which specific phenotypic features can be recapitulated in both zebrafish and dogs . Bmp3−/− mice described by Daluiski et al . have excessive trabeculation of the long bones , but defects in the cranial bones were not reported [23] . Interestingly , when authors of this study moved the Bmp3 null allele to an inbred background , Bmp3−/− mutants died perinatally due to lung defects . A preliminary craniofacial analysis of E18 . 5 embryos suggests that a number of morphogenesis defects occur in the mutants ( unpublished data , JJS and KLM; personal communication with KLM ) . In dogs , the BMP3 mutation is but one of at least five QTLs that modulate canine skull shape variation . Thus , it is possible that genetic interactions with other QTLs enhance or act permissively to BMP3F452L's effects on cranioskeletal development . Microdeletions that include or flank BMP3 are described in humans [33] . Although craniofacial abnormalities associated with these microdeletions were attributed to loss of PRKG2 , our results suggest that haploinsufficiency for BMP3 might also contribute to the clinical features of 4q21 syndrome . Furthermore , isolated BMP3 dysfunction could be the basis of human cephalic conditions whose genetic etiologies remain unknown . The development of modern dog breeds is one of the most extensive genetic experiments ever conducted . Their existence allows us to exploit breed-average phenotypes for genetic analysis . In the past , the extensive linkage disequilibrium inherent to artificial selection often hindered the process of fine mapping causal variants in the dog [16] . We overcame this limitation using whole-genome sequencing to comprehensively evaluate candidate variants . Combining the resulting insights with the functional utility of zebrafish , we identified a causal mutation underlying a quantitative trait in the dog . Together these approaches have allowed us to extend the paradigm of leveraging breed-average phenotypes to include the identification of causal mutations . We can now work towards assembling the full inventory of genes associated with vertebrate cranioskeletal shape , in turn illuminating evolutionarily conserved mechanisms of cranioskeletal development in our own species .
Fifty-one measurements were captured using an Immersion MicroScribe Digitizer G2X running Microscribe Utility Software and Diagnostics ( v5 . 0 . 0 . 2 ) . In total , 533 canid skulls representative of 120 breeds and 4 gray wolf subspecies located in museums and private collections were documented . Dorsal and ventral landmark datasets were captured separately and merged based on landmarks in common between datasets ( landmarks 1 , 2 , 28 , and 29 ) using File Converter software ( Klingenberg lab ) . Procrustes fit , PCA , and residuals were generated using MorphoJ [11] . Residuals of nonallometric shape were calculated as implemented in MorphoJ ( v1 . 03a ) using linear regression ( pooled by sex and breed ) , with symmetric component and log ( neurocranium centroid ) corresponding to dependent and independent variables , respectively . Ten thousand permutations were performed . Refer to Figure S1 to see landmarks used by MorphoJ to calculate neurocranium centroid . A covariance matrix based on residuals was analyzed by PCA . GWAS was performed using a subset of the CanMap dataset of genotypes [3] . In total , 72 breed-sex averages of PC1 were assigned to CanMap breeds . In 30 instances , only one skull per breed-sex was measured . In such cases , the actual PC1 score was used for CanMap phenotype assignments . Log ( neurocranium centroid ) values were similarly assigned and used in subsequent analyses as a size covariate for PLINK and GEMMA association analyses ( see next section ) . Skull surface scans ( 1 Pug , 1 gray wolf ) were done by Konica Minolta ( 3D Sensing Labs , Ramsey , NJ ) . Decimated scans were loaded into Landmark Editor ( v3 . 6 ) [34] . Skull morphing was done using PC1 landmark coordinates exported from MorphoJ . Coordinate files used for morphing were generated from representatives of dolichocephalic and brachycephalic breeds ( a Collie and Pug ) . Base pair positions stated throughout refer to CanFam2 ( Broad/May 2005 ) coordinates . Single marker and haplotype association analyses were done using PLINK ( v1 . 07 ) [15] or mixed model GEMMA ( v0 . 91 ) [17] where specified . CanMap markers used in the analysis included SNPs with missingness <0 . 10 and minor allele frequency >0 . 01 . In the full dataset ( all breeds with breed-sex PC1 averages ) , 61 , 270 SNPs were analyzed by PLINK from 576 dogs representing 62 American Kennel Club-recognized breeds . In the mixed-model , ∼36 , 685 SNPs were analyzed . Breeds used in size-stratified analyses are listed in Table S2 . Significantly associated SNPs surpassed Bonferroni correction at the 0 . 05 level ( −log10 ( P ) > = 5 . 86 ) . HO was calculated by treating CanMap breeds at the polar extremes of PC1 as two comparisons populations ( Pug , Pekingese , Boston Terrier , Shih Tzu , Brussels Griffon , French Bulldog , Bulldog , Boxer , Cavalier King Charles Spaniel , Chihuahua versus Collie , Borzoi , Saluki , Scottish Deerhound , Bloodhound , Greyhound , Scottish Terrier , Doberman Pinscher , and Irish Wolfhound ) . FST was calculated treating brachycephalic breeds ( listed above ) as a single subpopulation . HO , HR ( the ratio of dolicho- and brachycephalic HO ) , and FST values were calculated using custom R scripts . fastPHASE was used to generate haplotype frequencies by breed , using CanMap genotypes using the clustering parameter k = 15 [35] . “Extreme brachycephalic breeds” were designated as such if both PC1 breed-sex averages exceeded 0 . 15 . This cutoff was chosen based on the obvious jump in magnitude of PC1 values ( see Figure 1E , Figure S6 ) . Breeds that meet this classification include the Pug , Pekingese , Boston Terrier , French Bulldog , Bulldog , Brussels Griffon , and Shih Tzu . DNA used in our study was extracted from blood samples as previously described [16] . In addition to whole-genome sequencing ( see below ) , BMP3 and PRKG2 were Sanger sequenced using six brachycephalic and six dolichocephalic breeds ( data not shown ) . The BMP3 8 , 196 , 098 C/A transversion was sequenced in an expanded panel composed of 847 dogs from 113 breeds . Primers were designed with a melting temperature ( Tm ) ranging between 68–72°C , GC content ranging between 20–80% , length ranging between 18–32 nucleotides , and included 5′ M13 tags ( Table S10 ) . PCR products for sequencing were generated with a 2-step thermocycler program: PCR products were sequenced using a standard protocol [16] . During the course of SNP discovery , we discovered errors in the reference genome sequence for canine BMP3 , producing two early stop codons in the first exon . Sequencing of 13 dogs , including the individual from which the reference genome sequence was derived , indicates these stop codons are the results of errors in the reference sequence . Paired-end libraries were prepared from DNA from eleven dogs of breeds with widely varying skull shapes . Sequencing was conducted on an Illumina HiSeq 2000 sequencer to a depth of 5 . 6–8 . 5× per dog using manufacturer protocols . The resulting 101-base paired-end sequences were mapped to the genome ( CanFam2 release May 2005 ) with bwa version 0 . 5 . 9-r16 with read trimming set to 15 . SNPs were called with samtools mpileup version 0 . 1 . 18 and custom R scripts [36]–[39] . Thirteen SNPs in the PC1-associated region overlap with the CanineHD Genotyping BeadChip ( Illumina cat . no . WG-440 ) . DNA from four dogs was assayed with the chip; all resulting genotypes were identical in the deep sequencing and chip results . Four hundred and fifty-two SNPs were identified in the critical interval ( 85 . 7 kb between 8 , 152 , 258 and 8 , 237 , 937 ) , and subjected to further filters . Genotypes with a genotype quality score below 8 were reset to “unknown . ” We performed association analysis using PLINK with options specifying an additive model omitting the Scottish Terrier , a dolichocephalic breed that appears to be an outlier [15] . After correcting for multiple testing , no SNPs were significantly associated due to limited statistical strength of the test . SNPs in the 5th percentile for association scores were retained . Cross-species conservation was assessed by the UCSC phastCons4way calculations [40] downloaded November 30 , 2011 , which is generated by using the phastCons program to score the extent of conservation between dog , human , mouse and rat . SNPs with a phastCons4way score above 0 . 7 were retained . SNPs in an exon or within 20 bases of a splice junction were retained . Morpholino knockdown experiments of bmp3 used two translation blockers ( MO1: 5′-TGACAGCGATCCATGCTGGAGGTGC-3′ , MO2 5′-CGGGACTATGGAAGCTGATCTA-3′ ) , which overlapped by one nucleotide . Morpholino injections used 5 . 1 ng ( MO1 ) or 7 . 5 ng ( MO2 ) , as determined by titrations . Zebrafish bmp3 ( IMAGE Id 7052011 ) and human BMP3 ( Origene clone SC302990 ) cDNAs were sequenced and determined to be full length . Missense F→L mutations for mouse GDF1 [41] , human Bmp3 , and zebrafish bmp2b [42] were introduced using site-directed mutagenesis and confirmed by sequencing . Zebrafish bmp3 wt and F→L cDNAs were PCR-amplified using gene-specific primers with attB sites . PCR products were subcloned into entry and destination vectors ( pCSDest ) using Gateway recombination , as previously described [43] , [44] . To construct the human BMP3 expression vector , we PCR-amplified the TGF-β signaling domain using primers with XbaI and XhoI restriction sites . PCR products were ligated into an expression vector bearing the Xenopus BMP2 prodomain , as such heterologous fusion constructs were previously shown to enhance propeptide cleavage and biological activity [41] . mRNA was synthesized using Ambion's SP6 mMessage kit from plasmid that was linearized with Not I . Embryo analyses of RNA injections were done based on injections of the following amounts: 25–300 pg human BMP3 mRNA , 25–300 pg mouse Gdf1 mRNA , 1–100 pg zebrafish bmp2b . mRNA overexpression assays were repeated three or more times at each stated concentration , unless stated otherwise . In situ hybridization was completed as described in Thisse and Thisse 2008 [45] , except probes were hydrolyzed for 2 minutes at 65°C , the hybridization solution contained 5% dextran sulfate , and the anti-DIG-AP incubation and subsequent washes were performed in Malic Acid Buffer rather than PBST . Alcian blue stains were done as previously described by Schilling et al . [46] , except that staining solution was composed of 0 . 15% Alcian blue , 50% EtOH , and 0 . 1 M HCl ( pH = ∼1 ) . Embryos were imaged using Zeiss Axio Imager . M1 , Zeiss SteREO Lumar v12 , or Leica M216F compound microscope . Zeiss Axiovision v4 . 8 . 1 software was used for image capture . Nonspecific background and dissection debris were removed from images of Alcian blue cartilage dissections using Adobe Photoshop CS3 . All plots were generated using custom scripts , in conjunction with R Cran packages ggplot2 [38] , reshape2 [39] , and RColorbrewer [47] . Manhattan plot and Q-Q plot scripts were adapted from examples posted on the blog “Getting Genetics Done” [48] . Post-processing of plots was done using Adobe CS4 Creative Suite softwares Photoshop , InDesign , and Illustrator . Informed consent was obtained for all collected dog samples . All animal protocols ( dog and zebrafish ) were approved by the Animal Care and Use Committees of the Intramural Program of the National Human Genome Research Institute at the National Institutes of Health or by Animal Care Committee of the Hospital for Sick Children Research Institute . Wild canids samples were graciously provided by Dr . Robert Wayne , in accordance with UCLA Approved Animal Care and Use Committee Policies .
|
As a result of selective breeding practices , modern dogs display a multitude of head shapes . Breeds such as the Pug and Bulldog popularize one of these morphologies , termed “brachycephaly . ” A short , upward-pointing snout , a massive and rounded head , and an underbite typify brachycephalic breeds . Here , we have coupled the phenotypes collected from museum skulls with the genotypes collected from dogs and identified five regions of the dog genome that are associated with canine brachycephaly . Fine mapping at one of these regions revealed a causal mutation in the gene BMP3 . Bmp3's role in regulating cranial development is evolutionarily ancient , as zebrafish require its function to generate a normal craniofacial morphology . Our data begin to expose the genetic mechanisms unknowingly employed by breeders to create and diversify the cranial shape of dogs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genome-wide",
"association",
"studies",
"animal",
"genetics",
"genetic",
"mutation",
"gene",
"function",
"animal",
"models",
"genome",
"sequencing",
"developmental",
"biology",
"model",
"organisms",
"mutation",
"types",
"skeletal",
"development",
"morphogenesis",
"biology",
"zebrafish",
"genetics",
"genomics",
"genetics",
"and",
"genomics"
] |
2012
|
Variation of BMP3 Contributes to Dog Breed Skull Diversity
|
The prefrontal cortex is centrally involved in a wide range of cognitive functions and their impairment in psychiatric disorders . Yet , the computational principles that govern the dynamics of prefrontal neural networks , and link their physiological , biochemical and anatomical properties to cognitive functions , are not well understood . Computational models can help to bridge the gap between these different levels of description , provided they are sufficiently constrained by experimental data and capable of predicting key properties of the intact cortex . Here , we present a detailed network model of the prefrontal cortex , based on a simple computationally efficient single neuron model ( simpAdEx ) , with all parameters derived from in vitro electrophysiological and anatomical data . Without additional tuning , this model could be shown to quantitatively reproduce a wide range of measures from in vivo electrophysiological recordings , to a degree where simulated and experimentally observed activities were statistically indistinguishable . These measures include spike train statistics , membrane potential fluctuations , local field potentials , and the transmission of transient stimulus information across layers . We further demonstrate that model predictions are robust against moderate changes in key parameters , and that synaptic heterogeneity is a crucial ingredient to the quantitative reproduction of in vivo-like electrophysiological behavior . Thus , we have produced a physiologically highly valid , in a quantitative sense , yet computationally efficient PFC network model , which helped to identify key properties underlying spike time dynamics as observed in vivo , and can be harvested for in-depth investigation of the links between physiology and cognition .
The prefrontal cortex ( PFC ) is a key structure in higher-level cognitive functions , including working memory , rule and concept representation and behavioral flexibility [1–6] , and has been linked to impairments of these functions in psychiatric disorders like schizophrenia [7–10] or attention-deficit/hyperactivity disorder [11] . Our understanding of the computational and dynamic mechanisms underlying these cognitive functions , their neuromodulation , and their aberrations in psychiatric disorders , is still very limited , however . Computational network models are a highly valuable tool for driving forward such an understanding , as data from many different levels of experimental analysis can be integrated into a coherent picture . With respect to psychiatric conditions , it is of particular importance that models incorporate sufficient biological detail and exhibit physiological validity in order to serve as explanatory tools . Psychiatric conditions like schizophrenia are characterized by a multitude of abnormalities in diverse cellular and synaptic properties , transmitter systems , and neuromodulatory input [7–10] . Moreover , pharmacological treatment options target the neurochemical and physiological level , yet they are supposed to change functionality at the behavioral and cognitive level . It is thus crucial to gain insight into the explanatory links between behavioral functions and the underlying neurobiological “hardware” , a task that requires sufficient physiological detail in the model specification , in particular realistic assumptions about anatomical structure and cell type diversity . Ultimately , the physiological validity of a computational model ought to be reflected in the degree to which it can reproduce and predict detailed aspects of the neural activity observed in vivo . That is , from a statistical perspective , one may define a good , physiologically valid model as one that accurately ( i . e . , quantitatively ) captures distributions compiled from the electrophysiological activity ( spiking , field potentials , membrane voltages ) produced by networks in vivo , but not necessarily as one that captures every detail of membrane biophysics or receptor kinetics . In our perception , such requirements are currently not met even by sophisticated cortical network models which do include a lot of biophysical detail [12–14] , as these are often only loosely compared to in vivo data or test only specific aspects of those . In this work , we present a computational network model of the PFC which has high physiological validity and predictivity both at the single-neuron- ( in vitro ) and at the network- ( in vivo ) level , yet is still simple enough to be computationally tractable . Its anatomical structure , neural , and synaptic properties are completely derived from the experimental literature and our own experimental data . The activity of the network is compared with a range of statistics derived from in vivo data , including spike trains , local field potentials , and membrane potential fluctuations . The model turns out to reproduce these data quantitatively , and also exhibits robustness with respect to moderate changes in parameters .
The network model introduced in Materials and Methods aims to combine computational tractability with physiological validity . This balance is achieved by embedding a simple , reduced two-dimensional single neuron model into a realistic network architecture that is derived from the experimental literature . All model parameters were directly estimated from our own in vitro data and the experimental literature ( see Materials and Methods for details ) , and no specific parameter tuning was necessary to bring the network model closer to in vivo-like behavior . At the single-cell level , the network is based on an approximation ( simpAdEx [15] ) to the adaptive exponential integrate-and-fire model ( AdEx [16] ) which yields closed-form expressions for instantaneous and steady-state firing rates , thus allowing for fast and fully automatized fitting to f-I and V-I curves from physiologically recorded cells ( Fig 1A ) . We had shown previously that this cell model is able to accurately predict spike times of recorded neurons driven by in vivo-like fluctuating currents not used for model fitting [15] ( Fig 1B ) and , like the full AdEx [17] , can generate a wide range of spike patterns . In vitro recordings from ∼200 L2/3 and L5 pyramidal cells , fast-spiking and bitufted interneurons from the medial PFC of adult rodents were used to generate a distribution of model cells that reflects the diversity of neurons in the real PFC ( see Materials and Methods for details ) . The resulting model parameters ( Table 1 ) follow broad distributions ( Fig 1C ) , mostly of Gaussian shape , with the exception of ΔT and τw which are best described by a Gamma distribution , and b which follows an exponential distribution ( red curves in Fig 1C indicate distributions from which model parameters were drawn ) . Anatomically , the network is divided into two laminar components , representing the superficial layers L2/3 and deep layer L5 ( Fig 2A ) . Neurons are distributed over the five cell types in each layer based on estimates from the literature ( Table 2 ) . The neurons are randomly connected with different connection probabilities pcon for each pair of cell types according to the literature [18–27] , including local clusters of higher connectivity [28 , 29] . The neurons are assumed to be organized in a single column and horizontal spatial distance is not taken into account . However , all neurons receive a constant background current ( i . e . , without fluctuations ) that represents synaptic connections from outside the network , both within and outside the same column ( see section “Admissable and realistic range of input currents” below ) . Since these currents were constant , all irregularity was produced intrinsically within the simulated network . Neurons are connected by conductance-based synapses ( AMPA , GABAA and NMDA ) with kinetics estimated from electrophysiological data , short-term synaptic plasticity [30] that is matched to the types of the connected neurons [31 , 32] , synaptic delays and random failure of synaptic transmission [33–36] . Distributions of synaptic weights ( log-normal [37] ) and delays ( Gaussian ) were extracted from the literature ( Table 3 ) . The average connection strength ( connectivity pcon times synaptic peak conductance gmax ) between pyramidal cells and interneurons in the different columns and layers is indicated by the width of the arrows in Fig 2A . Wherever possible , we used data from the rodent prefrontal cortex , or at least agranular cortices such as the motor cortex , which in rodents shows a similar layered anatomy as the PFC . Apart from the missing granular layer 4 , specific features of the rodent PFC that are modeled here include an increased fraction of reciprocal compared to unidirectional connections [32] , longer NMDA time constants than in other areas [38 , 39] , and a uniquely prefrontal distribution of short-term synaptic plasticity properties for connections among pyramidal cells [32] . To assess whether the network model can reproduce the dynamics of real prefrontal neurons in vivo , we compared measures computed from the model with those from electrophysiological data , as well as with a number of findings from the literature . Unless otherwise stated , we simulate a single column with 1000 neurons and apply a constant DC current of 250 pA to all pyramidal cells and 200 pA to all interneurons . These currents are the only parameters that are not directly obtained from experimental data . As discussed below , appropriate values for these currents were derived by inferring from lumped-population input simulations the amount of current produced by a network of realistic size , set up with the very same structure as the explicitly modeled network . Spike-train statistics . All experimentally recorded spike trains ( kindly provided by Dr . Christopher Lapish , Indiana University Purdue University , Indianapolis , see [40] for details ) were first segregated into statistically stationary segments to yield estimates of spike train statistics that reflect in vivo baseline activity , free from task-related responses ( not modeled here ) or other potential confounds [41] . For consistency , the same procedure was applied to the simulated spike trains , although , strictly , these were stationary by simulation setup . From all jointly stationary segments , the mean 〈ISI〉 , coefficient of variation CV , and autocorrelation function of the inter-spike intervals ( ISIs ) were computed for each individual spike train , as well as the zero-lag cross-correlation CC ( 0 ) between pairs of neurons ( Fig 3 ) . The in vivo data show very low zero-lag cross-correlations between neuron pairs ( 2 . 4 ⋅ 10−4 ± 2 . 5 , mean ± SD ) and CV s near one ( 1 . 04 ± 0 . 33 ) , consistent with the proposal of an “asynchronous-irregular” ( AI ) state of cortical dynamics ( although the correlations theoretically proposed for the AI state are usually even at least one order of magnitude larger than obtained here [42] ) . The average single-cell ISIs follow a monotonically decreasing distribution with a mean comparable in size to the standard deviation ( 570 ± 610 ms ) , but with a heavy tail that is better described by a log-normal or beta-2 distribution [43] rather than an exponential distribution . The autocorrelation function shows a rapid decay with small negative flanks ( half-width at half maximum: 10 . 1 ± 1 . 1 ms , minimum: 64 . 6 ± 69 . 9 ms , mean ± SD ) . Without further tuning of network parameters beyond their derivation from slice-physiological and anatomical data , all these in vivo statistics are well reproduced by the model ( Fig 3 ) . Two-sample Kolmogorov-Smirnov tests did not find notable differences between experimental and simulated distributions in any of the statistics ( CV : p = 0 . 26 , KS ( 29 ) = 0 . 28; mean ISI: p = 0 . 4 , KS ( 29 ) = 0 . 23; CC: p = 0 . 4 , KS ( 29 ) = 0 . 24 ) , indicating that simulated distributions were not statistically distinguishable from the experimental ones . The asynchronous- irregular firing with low rates is also seen in the raster plot of spike times ( Fig 3B ) . Low fraction of spiking neurons and layer-dependent firing rates . Fig 3B reveals a relatively low fraction of spiking pyramidal cells in both layers—only 22% of the cells emitted more than 10 spikes during the 30s of simulated time , which will be used as the definition of “spiking neurons” throughout the paper , in line with [44–46] . Comparing the neural and synaptic parameters of those neurons which fire at a sufficiently high rate ( > 0 . 33 Hz ) and those which do not ( ≤ 0 . 33 Hz ) , we find that only the rheobase ( and the cell parameters that contribute to it ) differs between the two populations: Spiking neurons have rheobases at the lower end of the distribution ( 42 . 9 ± 2 . 1 pA ) , compared to 69 . 0 ± 1 . 6 pA for non-spiking neurons ( mean ± SEM; p = 3 . 5 ⋅ 10−20 , t ( 997 ) = 9 . 4 , two-sided t-test ) , some of them even firing spontaneously ( called “generator neurons” [47] ) . While neurons firing at very low rates may go undetected using extracellular single-unit recordings , recording techniques that are less biased toward spiking neurons , such as calcium imaging or in vivo patch-clamp , often reveal a large fraction of neurons that are mostly silent ( “dark matter theory” of neuroscience , [44–46] ) . Consistent with these results , the fraction of neurons with more than 10 spikes rarely exceeded 40% in simulations with in vivo-like firing patterns ( see section “Admissible and realistic range of input currents” below ) . This can be explained by the way the neurons are activated: While most neurons receive a background current above their rheobase , the high firing rates of the interneurons ( Fig 3B ) lead to an average membrane potential in the pyramidal cells below the firing threshold ( mean difference: -17 . 3mV , range: -37 . 2 to -2 . 2mV for the example shown in Fig 3 ) that is occasionally kicked above threshold by random fluctuations . This means that the firing rate is mostly determined by the amplitude of the fluctuations of the membrane potential ( see below for statistics ) . These results are qualitatively conserved across the range of input currents for which the overlap between experimental and simulated distributions is reasonably high . Membrane potential and local field potential statistics . In addition to the spike data , we also compared the membrane potential statistics and LFP signals between simulation and experiments . For the simulated network , we observed a broad range of membrane potential fluctuations ( after removing spike events; Fig 4A; 3 . 28 ± 0 . 72 mV , mean ± SD; range between 0 . 72 mV and 11 . 23 mV ) . We compared this distribution of standard deviations with those from in vivo patch-clamp recordings from 10 putative pyramidal cells during up-states in anesthetized adult rodent PFC ( kindly provided by Dr . Thomas Hahn , Central Institute of Mental Health and BCCN Heidelberg-Mannheim ) . The simulated distribution is less than one SEM away from the average of the experimental distribution ( pooled over all data sets ) for most bins , and a Kolmogorov-Smirnov test ( see Materials and Methods ) does not show a significant difference ( p = 0 . 45 , KS ( 29 ) = 0 . 23 ) . The range of membrane potential fluctuations in the model and in the recordings used here is also consistent with values found in the literature [48 , 49] . The local field potential ( LFP ) in the model was estimated as the sum of all synaptic currents ( allowing excitatory and inhibitory currents to partially cancel ) . This is a reasonable approximation to the standard model of the LFP [50] under the assumption that all neurons are confined in a small volume of cortical space . We computed the power spectral density of this model-derived signal and of the LFP signals obtained from the in vivo recordings ( Fig 4B ) . Up to a constant offset ( that has been removed in the figure ) , the spectrum of the simulated LFP is less than one SEM away from the average estimated from the experimental recordings ( from awake , behaving animals , also provided by Dr . Christopher Lapish [40] ) at most of the frequencies . Both spectra follow a 1/f power law for frequencies below 60 Hz and change their scaling behavior for higher frequencies , consistent with LFP spectra described in the literature [51–53] ( the fluctuations in the simulated curve are stochastic in nature , i . e . there is no systematic deviation from the 1/f behavior across different simulations ) . For frequencies beyond 60 Hz , the experimental spectrum is well described by a 1/f2 power law , while the simulated one rather follows a 1/f3 relation . Both scaling behaviors have been reported in the literature ( 1/f2 [52 , 54] , 1/f3 [51] ) , and the difference may result from the simplifications made in the computation of the simulated LFP , e . g . neglecting the spatial integration of currents in extracellular space or the contribution of active currents [14] . Transient information transfer and the role of neuronal heterogeneity . We next examined how neurons in L2/3 and L5 would respond to a simple stimulus simulated by a brief series of spikes at high rate ( 250 spikes within 5 ms ) from a virtual ( not explicitly simulated ) “input population” connected to 10% of the pyramidal cells in L2/3 ( cf . Table 3 ) . The stimulus induces a number of spikes in L2/3 , and with a short delay also in L5 ( Fig 5A ) . The delays ( L2/3: 8 . 9 ± 1 . 1 ms; L5: 17 . 7 ± 1 . 2 ms , mean ± SD ) are similar to values that have been reported in the literature ( e . g . 3 . 4 ± 0 . 5 ms in L2/3 and 16 . 6 ± 1 . 2 ms in L5 [55] ) . Note that these delays are significantly longer than the fixed synaptic delays ( below 2 ms , see Materials and Methods ) and arise from the dynamics of the neurons and the kinetics of the synapses ( c . f . [56] ) . For a sufficiently strong stimulus ( e . g . 500 spikes within 5 ms ) , the neurons in L2/3 show a brief period ( 100–150 ms ) of persistent activity ( Fig 5B ) . The transmission of transient stimuli between layers crucially depends on the heterogeneity of the neuronal parameters . With a 80% reduction in the variance of all parameter distributions ( but no change in the means ) , the stimulus only elicits a response in L2/3 , but is not transmitted to the output layer L5 anymore ( Fig 5C ) . Indeed , L2/3 activity is almost independent of neuronal variability , whereas the number of spikes in L5 systematically decreases as the standard deviation of neuronal parameters is reduced ( Fig 5D ) . To further examine the transmission dynamics , we reproduced an in vitro experiment with suppressed inhibition [57] which showed that input in L2/3 resulted in an epileptiform spread of activation across the whole network under this condition , whereas the same input in L5 did not . We mimicked this setup by reducing the inhibitory synaptic weights in the network to 30% of their original values and inducing a strong stimulus ( see above ) in each of the two layers , while varying the peak conductance gmax of the synaptic connection between the mimicked Poisson input population and the network . For moderate connection strengths ( gmax = 2 ) , only a fraction of the network responds , and the number of spikes elicited by the network is much larger if the stimulus is injected in L2/3 ( 404 ± 116 , mean ± SD ) compared to a stimulus in L5 ( 118 ± 33 ) . Higher connection strengths ( gmax = 20 ) reliably drive the network into an “epileptic state” ( transient high-rate response from all neurons in the network ) for a stimulus in L2/3 . In contrast , this state was never reached for an input in L5 , consistent with the experimental results in [57] . In the previous section we showed that the model can reproduce a wide range of characteristics of neural activity in vivo . Here , we assess how the reproduction quality of in vivo-like behavior depends on those parameters of the model which were only loosely constrained by experimental data . We restrict this analysis to the spike series statistics 〈ISI〉 , CV and CC ( 0 ) . Admissible and realistic range of input currents . The background currents I = [ I ex L23 , I ex L5 , I inh L23 , I inh L5 ] have so far been treated as free parameters , as such estimates are difficult to obtain or at least have not been reported experimentally . We address this in two ways: First , we systematically vary these four currents and assess the similarity between experimental and simulated spike time distributions using Kolmogorov-Smirnov statistics as before . Second , we estimate the required background currents from the simulation itself , using the assumption that the simulated network is embedded in a larger , but structurally identical network from which these currents originate . Fig 6A shows the Kolmorgorv-Smirnov test statistic DKS as a function of Iex and Iinh , where I ex L23 = I ex L5 = I ex and I inh L23 = I inh L5 = I inh ( see below for a discussion of laminar differences in the input currents ) . The figure reveals that the overlap between experimental and simulated distributions is acceptable ( p > 0 . 05 for the two-sample Kolmogorov-Smirnov test , i . e . failure to reject the null hypothesis H0 of equal distributions for the two samples , see Materials and Methods ) for a wide region of Iex and Iinh values ( delimited by the black isocline in Fig 6A , associated with DKS values below 0 . 4 ) . More specifically , simulated CV and mean ISI distributions become indistinguishable from their experimental counterparts as Iex increases , while the overlap with the ISI distribution decreases again for very high Iex values ( Fig 6A , left inset ) . Both CV and mean ISI deviate from the experimental distributions as Iinh increases . CC , on the other hand , matches well with the experiments for high Iinh values ( Fig 6A , lower inset ) . As mentioned above , the fraction of firing neurons is quite low in most networks showing in vivo-like firing patterns , typically between 20 and 30% , as shown in Fig 6B ( blackly delimited region gives the empirically acceptable parameter regime copied from Fig 6A ) . The ratio of inputs into the two layers , I ex L23 / I ex L5 and I inh L23 / I inh L5 , does not have a strong influence on these results within the tested range ( mean DKS ± SEM: 0 . 31 ± 0 . 05 , 0 . 32 ± 0 . 05 and 0 . 35 ± 0 . 06 for ratios of 1 , 2 and 4 , respectively ) , but does of course affect the relative firing rate between the two layers . In vivo experiments found that firing rates are considerably higher in L5 compared to L2/3 pyramidal cells ( 3–20 times [27] ) . This condition is fulfilled in our model as long as L2/3 receives less or the same input as L5 ( Fig 6C ) . To estimate which range of I values could be realistically assumed , we tested whether a substantially larger network than the 1000-neuron-network simulated here would produce mean synaptic currents that are large enough to self-sustain in vivo-like activity ( i . e . within the blackly circumscribed regions in Fig 6A ) . In this case , the activity in the large network and the small network ( the latter driven by the currents from the larger one ) would be indistinguishable , and the in vivo-like activity would be supported by the larger network . We increased the size of the network either by changing the density of neurons or by adding input from nearby columns ( see “Estimation of background currents” in Materials and Methods ) . Fig 7 shows the mean synaptic current into pyramidal cells and interneurons in L2/3 and L5 that would result from the reduced equivalent-population input models described in Materials and Methods if the network size was varied through the number of columns ( Fig 7A ) or the density of neurons within columns ( Fig 7B , both figures showing currents averaged over the values of the other independent variable , i . e . neural density or number of columns , respectively ) . The shaded areas show the ranges for Iex ( blue ) and Iinh ( red ) within which these currents would produce in vivo-like activity ( DKS < 0 . 4 ) . Note that it is sufficient that one of the two layers receives a current above the lower bound , as it will push the other layer into the right regime by cross-layer synaptic connections . The upper bound , on the other hand , may not be exceeded by either of the two layers , as this would push the other layer beyond its upper bound as well . It is apparent that these conditions are fulfilled already for ( spatially ) relatively small networks ( ∼ 5 columns ) , and currents saturate as network size grows further ( Fig 7A ) . By increasing the neuron density , on the other hand , the input currents increase monotonically over a wide range ( Fig 7B , averaged over all column numbers ≥ 5 ) . Mean synaptic currents sufficient to drive the network into the experimentally observed regime arise for densities between 19 , 000 and 44 , 000 neurons per mm3 . This range overlaps with densities found in anatomical studies ( 30 , 000 to 90 , 000 neurons per mm3 [58–60]; horizontal dotted lines ) . Variation of synaptic parameters . We attempted to estimate all synaptic parameters from data reported in the literature . Given that these come with some uncertainty and variation , however , we explored how sensitive the network behavior is with respect to changes in mean synaptic peak conductances and their distribution , synaptic time constants , and the GABAA reversal potential . All these parameter variations were performed for a range of different background currents and averaged results are reported . The GABAA reversal potential E rev GABA was initially set to -70mV , which is well within the range of the values reported in the literature [19 , 24 , 55 , 61] . Within the physiologically reasonable range from -90 to -60mV [62] , the divergence between simulated and experimental distributions ( as assessed by the KS test statistic ) increases with E rev GABA ( Fig 8A ) . At the same time , the standard deviation of the membrane potential decreases . The time constants of the synaptic kinetics also turned out to be important for the agreement with in vivo data: While small changes are acceptable , both very fast and very slow GABAA kinetics strongly diminish the agreement with the experimental data ( DKS = 0 . 99 for τon = 0 . 6 ms and τoff = 8 ms and DKS = 0 . 85 for τon = 12 ms and τoff = 160 ms ) . The NMDA time constants have less effect , unless they are very strongly increased ( DKS = 1 . 0 for τon = 17 . 2 ms and τoff = 300 ms , compared to values of τon ≤7 ms and τoff ≤ 100 ms reported in the literature [38 , 39] ) . The effects of the mean synaptic peak conductances are shown in Fig 8B . While small to moderate changes ( ± 50% ) have no significant effect , a strong decrease in the inhibitory synaptic efficiencies leads to a significant mismatch with the in vivo statistics . Apart from the mean , we also analyzed how the distribution of the synaptic peak conductances affected in vivo-like behavior by either reducing the variability or drawing them from a normal rather than a log-normal distribution ( conserving mean and standard deviation ) . Reducing the variability of the synaptic weights increased the mismatch between empirical and simulated distributions ( Fig 8C , gray line ) . Surprisingly , just changing the form of the underlying distribution from log-normal to normal , without changing its mean or standard deviation , had a similarly strong effect as a pronounced reduction in standard deviation ( Fig 8C , black line ) , so both the variability as well as the functional form of the synaptic conductance distribution are crucial for reproducing spiking dynamics as observed in vivo . Without the long tail of the log-normal distribution , the network activity becomes much more synchronized ( CC ( 0 ) : 0 . 017 ± 0 . 006 , mean ± SD ) and exhibits strong bursts ( CV : 1 . 81 ± 0 . 09 , mean ± SD ) , while mean firing rates are not much affected ( 〈ISI〉: 420 ± 558 ms , mean ± SD ) .
The current model has a strong focus on its tight connection to data . Many existing network models of the neocortex are based on neurobiological findings as well [13 , 14 , 63–65] , but the present model differs from them in two respects: The strict way in which the in vitro data is used to fix or systematically infer every detail of the model , and , more importantly , the quantitative test of the model’s validity on a wide range of in vivo findings . Recently , a few studies have also moved in this direction . Fisher et al . [66] proposed a model for the short-term memory circuit in the oculo-motor system of the adult goldfish . They fitted the model simultaneously to a range of anatomical , physiological and behavioral data . This approach gives a coherent picture of this particularly well-defined non-cortical system . Furthermore , Potjans and Diesmann [27] proposed a model of a sensory cortex network where the connection probabilities are thoroughly derived from in vitro studies . While the neuron parameters are generic and homogeneous , their focus is on the precise laminar and horizontal organization of the synaptic connections . They compare the results of their simulations with the baseline firing rates and flow of transient information through the different layers in vivo . These comparisons to experimental data are qualitative in character , as it is the case for most existing large-scale models of cortical networks [12–14 , 67] . However , a few recent studies also made statistical comparisons on partial aspects of physiological data [68 , 69] . It would be interesting to assess these models on a wider range of in vivo data as we proposed here , to see which degree of biological detail is sufficient to predict their key properties . An important simplification made in the present model is the reduction to two laminar components , leaving out layer 4 and 6 as well as the long-range fiber bundles and interneurons in layer 1 . While layer 4 is missing in rodent PFC , layer 6 is only weakly connected to the other layers in our reference connectivity maps , which are based on the motor cortex [57 , 70 , 71] . Thus , its inclusion in the network should not have a major impact on the results shown here . This is probably different in sensory networks , where layer 6 strongly interacts with both pyramidal cells and interneurons in layer 4 [72] . The model exhibits a low fraction of spiking neurons , consistent with results from recording methods such as calcium imaging , which are not biased towards high firing rates ( “dark matter theory” of neuroscience [44–46] ) . As described above , this may partly result from the variance-driven firing of the neurons: The membrane potential is on average well below the spiking threshold , but large fluctuations still lead to occasional spiking . The size of the fluctuations and the low-rate , Poisson-like firing ( CV ≈ 1 ) of the neurons is consistent with the high-conductance state theory [48] and balanced-state theory [42 , 73] . We note that the irregular and highly asynchronous firing of the neurons [74] observed here is a generic property of the network that simply emerged from its parametrization through in vitro and anatomical findings . There are two main determinants of the high-fluctuation regime of the model: First , variability in the membrane potential requires variability in the synaptic parameters and in particular , the fat tail of the log-normal distribution of the synaptic weights . Second , the range between the firing threshold Vup and the GABAA reversal potential E rev GABA must be sufficiently large , because below E rev GABA , all synaptic currents depolarize the cell , so the dynamical range for a balanced , variance-driven state is constrained between these two values . Using the multivariate distributions of neuron parameters obtained from our in vitro recordings , we also observed that decreased cellular heterogeneity has a profound effect on the processing of transient stimuli . It prevents the transmission of stimulus-induced activity from L2/3 to L5 . This phenomenon can be understood if one considers the rheobase distribution: Reduced heterogeneity removes those neurons that originally had a very low or even negative rheobase . These are the ones which are highly susceptible to even small inputs and form a small but significant fraction of L5 neurons that were activated by the transient synaptic input from the L2/3 cells . Given that L5 provides the majority of output to other brain areas , impaired transfer of stimuli to this layer may lead to major impairments in information processing . Thus , apparently quite subtle changes in the distributional properties of synaptic and cellular parameters ( not affecting their means ) may lead to major changes in network dynamics and functional connectivity among columns or areas , effects that have been proposed to underlie major psychiatric conditions like schizophrenia [8 , 9] . By varying the total input from a virtual population designed according to the same principles as the actually simulated network , we provided evidence that a larger network than the one actually simulated with anatomically realistic neuron densities should be capable of self-sustaining in vivo-like spiking modes . Although we did not demonstrate self-consistency in a strict sense , we have shown that the background currents into the smaller , simulated network needed to yield in vivo-like behavior are consistent with the range produced by a much larger network of anatomically reasonable size . For currents within the blackly delimited region of Fig 6A , the spike train distributions are statistically indistinguishable from the in vivo statistics , and the background currents that would result from scaling up the simulated network to anatomically realistic size lie exactly within this regime . This analysis implies that in vivo-like activity can be self-sustained in a larger network with the same anatomical layout as explicitly simulated here , as it has been observed for instance in deafferented cortical slabs [75] , while e . g . the thalamus or other sub-cortical structures may provide transient , stimulus-related input or modulate the overall activity of the network [76] . Interestingly , the currents produced by this procedure are much higher in L5 compared to L2/3 ( Fig 7 ) , as required for the much higher firing rates observed in vivo [27] . At first glance , this seems counterintuitive , as L2/3 neurons receive input from neighboring columns , while L5 neurons do not ( see “Estimation of background currents” in Materials and Methods ) . However , L5 also receives strong inputs from L2/3 , while the inverse projections are much weaker ( Fig 2A ) . Thus , once L2/3 neurons receive enough input from other columns to spike , they drive L5 much stronger than themselves . In terms of space , input from just a few columns is sufficient to drive the network , as connectivity rapidly decays over the cortical extent . Nevertheless , a single column is not sufficient for driving the network because of the higher fraction of excitatory synapses in long-range connections and the more local connectivity of interneurons . This is consistent with recent experimental studies [60 , 77] and earlier results from deafferented cortical slabs [75] ( but see [78] ) . In this study , we have focused on the resting state of the network . However , it may also be used as a foundation for more functional investigations of cognition . For instance , the clusters of increased synaptic connectivity may serve as building blocks for cell assemblies [29] which can be used to represent behavioral rules [3 , 40] or transient stimuli that need to be kept in working memory [64] . Moreover , the fast and fully automatized framework for fitting the neuron model to in vitro data [15] opens a convenient way to test the network effects of genetic or pharmacological manipulations: Recordings from neurons that underwent such a manipulation can be used by the very same fitting procedure as employed for wildtype or control cells , resulting in different parameter sets that can be plugged into the network to assess their implications for network behavior . Likewise , this could be done for the synaptic parameters using paired recordings and recent methods to fit the parameters of short-term synaptic plasticity models to these data [79] . In summary , we have provided a prefrontal cortex network model here with single cells and synapses strictly parametrized through in vitro electrophysiological findings ( no specific tuning or adjustment of synaptic currents to compensate for simulated network size ) , with realistic cellular and synaptic heterogeneity , and with a structural layout derived from anatomical data . We have then systematically compared the full network activity to a number of spiking and correlation statistics from in vivo multiple single cell recordings in awake rodents , as well as LFP data from these animals , and estimates of membrane potential fluctuations from in vivo patch-clamping . Our model is therefore highly validated at the in vivo physiological level , yet it is computationally efficient by virtue of its computationally comparatively simple single unit design . We therefore hope that this network model can serve as a valuable tool in the further study of how physiological and anatomical properties relate to cortical network dynamics , and ultimately cognition , and how alterations of these properties may give rise to symptoms observed in various psychiatric conditions .
Neuron model . Single neurons were modeled by the simplified adaptive exponential integrate-and-fire neuron ( simpAdEx ) introduced in [15]: C · d V d t = - g L · ( V - E L ) + g L · Δ T · e ( V - V T Δ T ) + I - w = : w V - w ( 1 ) d w d t = 0 for w - w V > τ m τ w w V Θ ( V T - V ) · 1 - τ m τ w d w V d V d V d t otherwise ( 2 ) if V > V up then V → V r and w → w r = w + b if w = 1 + τ m τ w w V then w → 1 - τ m τ w w V , where C is the membrane capacitance , gL a leak conductance ( with reversal potential EL ) , τm and τw are the membrane and adaptation time constants , respectively , Θ denotes the heavy-side function , and wV is the V-nullcline of the system as defined in Eq 1 . Like the full AdEx [80] , this model consists of one differential equation for the membrane potential V ( including an exponential term with slope parameter ΔT , which causes a strong upswing of the membrane potential once it exceeds VT ) , and one for an adaptation variable w , and can reproduce a whole variety of different spiking patterns [15] . A spike is recorded whenever V crosses Vup , at which point the voltage is reset to Vr and spike-triggered adaptation is simulated by increasing w by a fixed amount b . The simpAdEx was derived from the full AdEx based on phase-plane considerations , effectively dissecting the dynamics into three different regimes ( defined through their distance from the V-nullcline , see Eq 2 ) , each of them approximated in a way that allows for closed-form expressions for the instantaneous and steady-state firing rates . This enables fast and efficient fitting of the model to f-I and I-V curves as commonly used to characterize the electrophysiological behavior of cells in vitro [15] ( Fig 1A ) . We had shown previously that this model , although estimated from f-I and I-V curves only , can predict spike times under in vivo-like conditions with high accuracy from physiological recordings not used for model fitting [15] . Different from [15] , the upper voltage limit Vup was initially estimated from the inflection point of the voltage traces . This makes Vup an absolute firing threshold ( as in the leaky integrate-and-fire neuron ) and leaves Vth as a free parameter for the subthreshold dynamics , resulting in a shallower exponential rise to the spike , more akin to what would be expected from the action of persistent sodium channels [81] or L-type calcium channels [82] . We estimated neuron models for a large number of in vitro recordings from different cell types from the prefrontal cortex of rats and mice , namely layer 3 ( n = 34 ) and layer 5 pyramidal cells ( n = 108 ) , fast-spiking ( n = 32 ) , and bitufted ( n = 22 ) interneurons . Additionally , we extracted statistics ( means and variances ) about f-I curves and subthreshold dynamics of Martinotti cells from the literature [26 , 83–88] , and used these to construct 100 sets of f-I and I-V curves drawn from Gaussian distributions instantiated by the empirically estimated parameters . For each data set drawn from these distributions , Martinotti cell models were estimated . The pool of estimated models for each cell type defines a multivariate parameter distribution for each type of neuron , from which the final parameter sets for the 1000 neurons used in the network simulations were drawn . This joint parameter distribution for each cell type was initially modeled as a multivariate Gaussian , where marginal distributions not of Gaussian shape ( as estimated from the empirical data ) were first Box-Cox-transformed to adhere with the Gaussian assumptions . In a second step , the Box-Cox transform was inverted to regain the non-Gaussian shape of the marginal distributions ( red curves in Fig 1C ) . The mean values and standard deviations of all model parameters for the different cell types are given in Table 1 . Network anatomy and connectivity . The network is divided into two laminar components , representing the superficial layers L2/3 and the deep layer L5 ( Fig 2A ) . The network also includes a horizontal organization into distinct columns which are typically about 300μm wide [89] , so the model is in principle suited to study information transfer between columns . For most part , however , the present analyses focuses on a single column which was found to be sufficient to reproduce in vivo-like resting-state activity , provided a source of constant external input ( see below ) . The relative numbers of pyramidal cells and interneurons in each layer were taken from [58] who studied the rat motor cortex , as such data are not available for the PFC . Following [58] , 47% of all cells were modeled as L2/3 pyramidal cells ( L2/3-E ) , 10 . 4% as L2/3 interneurons ( L2/3-I ) , 38% as L5 pyramidal cells , and 4 . 6% as L5 interneurons . With regards to the specific types of interneurons and their distribution across layers , we followed [90] and [91] and defined local interneurons ( IN-L ) with projections within the same layer and column as fast-spiking cells , cross-layer interneurons ( IN-CL ) as bitufted cells , and far-reaching interneurons ( IN-F ) with projections both outside of their column and layer of origin as Martinotti cells [90] ( Table 1 ) . The cross-column cells ( IN-CC ) have been classified as large basket cells [90 , 92 , 93] , with electrophysiological properties resembling those of pyramidal cells [94] . Therefore , we used the same parameter distributions as for the pyramidal cells in the respective layer for this cell class . Markram et al . [90] also estimated the relative numbers of different types of interneurons within each cortical layer . Together with the classification above , these data result in the full distribution of cell types summarized in Table 2 . Neurons were randomly connected with distinct connection probabilities pcon for each pair of cell types as derived from a survey of about 40 studies , e . g . [95–99] , most of which are reviewed in [22] and [27] , except [18–21 , 23–25] and [26] . Most of them performed whole cell or dual sharp electrode recordings in vitro in various neocortical regions of rats and mice . We also included a few studies using monkeys , ferrets and cats , as there are more studies from PFC in these species and some parameters were not available in rodents . Connection probabilities were further adjusted jointly with the connection weights to match data from photostimulation experiments as explained in detail below [57 , 70 , 71 , 100 , 101] . Pyramidal cells within the same layer form clusters of increased connection probability as defined by the “common neighbor rule” [28 , 29] which states that the connection probability of two neurons increases linearly with the number of neurons they are both connected to . Furthermore , a fraction of 47% of the connections was specified as reciprocal [32] , since the proportion of reciprocal connections was experimentally observed to be significantly higher than chance [32 , 37] . For cross-column projections , connection probabilities exponentially decay with the distance from the column of origin . Data from rodent studies [29 , 60 , 77 , 102 , 103] suggest a wide range of spatial decay constants . We use the median from these studies , which is 114μm for pyramidal cells and 95μm for interneurons . Apart from the recurrent synaptic connections within the network , we also introduced constant background currents that are fed into all neurons and which differ in strength for pyramidal cells and interneurons in layer 2/3 and 5 . Appropriate values for these four streams of background inputs were determined using reduced equivalent-population input models ( see below ) . It is emphasized that there was no source of external noise fed into the network , i . e . the external inputs consisted of just constant ( DC ) currents . Thus , all variability observed in the network arises from its internal dynamics . Synaptic properties . Neurons were connected through conductance-based AMPA- , GABAA- , and NMDA-type synapses , with kinetics modeled by double exponential functions [104] IX=gXmax s ( V ) ∑tspa ( tsp ) ( e− ( t−tsp−τD ) /τoffX−e− ( t−tsp−τD ) /τonX ) ( V−ErevX ) with s ( V ) ={ 1 . 08 ( 1+0 . 19·exp ( −0 . 064V ) ) −1for X=NMDA1otherwise ( 3 ) where X ∈ {AMPA , GABAA , NMDA} . The reversal potential Erev is set to zero for AMPA and NMDA , and to −70 mV for GABAA [19 , 24 , 55 , 61] . The onset and offset time constants τon and τoff are set to 1 . 4 ms and 10 ms , respectively , for AMPA [39 , 94] , 3 ms and 40 ms for GABAA and 4 . 3 ms [105] and 75 ms [38 , 39] for NMDA . NMDA conductances exhibit a nonlinear voltage-dependency s ( V ) due to their magnesium block at lower voltages [106] . Synaptic transmission delays τD were drawn from Gaussian distributions with means and standard deviations depending on the pair of connected cell types , with parameters derived from the same electrophysiological literature as the connection probabilities ( see below; Table 3 ) . Synaptic delays were chosen to increase linearly with distance from the target column [107] , τD ( d ) = τD ( 1 + d ) , where d is the number of columns separating the connected neurons . Synapses were also equipped with short-term plasticity dynamics implemented by the corrected version [108] of the Tsodyks and Markram model [30] a k = u k · R k ( 4 ) u k = U + u k - 1 ( 1 - U ) exp ( - Δ k - 1 / τ fac ) ( 5 ) R k = 1 + ( R k - 1 - u k - 1 R k - 1 - 1 ) exp ( - Δ k - 1 / τ rec ) . ( 6 ) These recursive equations describe the dynamics of the relative efficiency a ( tspk ) across series of spikes , with initial conditions u1 = U and R1 = 1 , where tspk is the interval between the ( k − 1 ) th and the kth spike . Model parameters U , τrec and τfac were specified according to [31] and [32] who differentiated between facilitating ( E1/I1 ) , depressing ( E2/I2 ) or combined ( E3/I3 ) short-term dynamics , for both excitatory ( E ) and inhibitory ( I ) connections ( Fig 2B , right panel; Table 4 ) . The cell types of the pre- and postsynaptic neurons determine which of these classes is used for each individual combination ( Fig 2B , left panel ) . Synaptic inputs were further subject to release failures with a probability of 30% [33–36] . Distributions of peak conductances ( “synaptic weights” ) gmax for each cell population were derived in two steps . As the first step , initial estimates were obtained from the anatomical and electrophysiological literature ( see above ) . Generally , peak conductances were adjusted such that they reproduced log-normal distributions of postsynaptic potential ( PSP ) amplitudes as reported in [37] ( means and standard deviations given in Table 3 ) . For excitatory synapses , only the AMPA conductances are specified this way , while NMDA conductances are given by 1 . 09 times the respective AMPA peak conductance [38 , 39] , with both AMPA and NMDA synapses activating after the same delay . For synaptic connections where peak conductances were not directly available from the surveyed literature , estimates were obtained in one of the following ways: 1 ) Missing estimates for specific interneuron types were replaced by estimates from other interneuron types where possible . 2 ) Missing estimates for inhibitory connections within one layer were replaced by those from another layer , rescaled such that they followed the same between-layer ratio as the excitatory inputs . 3 ) If only means but no standard deviations for the distribution of synaptic parameters were available , we used standard deviations from another layer scaled according to the ratio of the means between layers ( missing values of connection probabilities or synaptic delays were estimated in the same way ) . Finally , for cross-column projections , synaptic weights were assumed to decay with the same exponential course ( space constant of 114μm for pyramidal cells and 95μm for interneurons ) as taken for the connection probabilities themselves ( see above ) . In a second step , since by far most of the studies cited above have been performed in sensory areas , data from laser scanning photostimulation ( LSPS ) [57 , 70 , 100 , 101] and genetically targeted photostimulation [71] studies from motor cortex were used to obtain values closer to PFC . Specifically , all connection probabilities and synaptic weights were scaled such that the total input to each cell type would match the one observed experimentally in these studies . To compute this scaling factor s i , j I , the product pcon ⋅ gmax of connection probability and synaptic weight was assumed to be proportional to the quantity ILSPS obtained in the experimental studies [70] , yielding s i , j I = I LSPS ( i , j ) / I LSPS p i , j · g i , j / p con · g max , ( 7 ) where |⋅| denotes the sum over all matrix elements . pcon and gmax are then multiplied element-wise by s I , such that pcon ⋅ gmax agrees with ILSPS . The scaled values of all parameters are given in Table 3 . The average connection strength ( as defined by pcon ⋅ gmax ) between pyramidal cells and interneurons in the different stripes and layers is coded by the arrow width in Fig 2A . Simulation details . All simulations were performed in customized C code written by the authors . Differential equations were numerically integrated using a 2nd-order Runge-Kutta method with a maximum time step of 0 . 05 ms , and all spikes , synaptic , and external events were exactly timed by adjusting the time steps accordingly . More specifically , whenever an incoming spike or a change in external currents occurs within the default time step , the time step is reduced accordingly and all equations are updated at the precise time of that event . Neurons were initialized with V i ( 0 ) = E L i and wi ( 0 ) = 0 for all i . MATLAB-based routines were used for parameter estimation and network analysis . All software is publicly available at https://www . zi-mannheim . de/index . php ? id=626 and on the freely available repository ModelDB ( http://senselab . med . yale . edu/ModelDB/ ) . The constant background currents I used in the simulations are assumed to replace missing synaptic input from the surrounding network not explicitly simulated . A network of sufficient size should be able to produce these amounts of current inherently ( with physiologically realistic synaptic efficacies as used here ) and thus self-sustain its in vivo-like activity . To test this idea , we computed the amount of current that is produced by a larger network N that is set up and connected exactly the same way as the actually simulated network n and compared it to the range of background currents that are required for in vivo-like activity . The currents IN were modeled as the time-averaged synaptic currents that are elicited in a single neuron for each cell type ( with the averaged cellular and synaptic parameters of all neurons of that type ) in response to a bombardment of spikes drawn from a Poisson distribution that mimics a large number of input neurons , reflecting its input connectivity . For a Poisson input spike train with the same firing rate as in the original network , this yields the same synaptic current In as in the full simulation . Larger networks N are simulated by using a higher number of inputs , resulting in higher overall input spike rates . Because connection probabilities decay with distance between neurons [89] , we independently tested two ways to increase the number of input neurons in the network: By increasing its spatial size L ( measured in number of columns NC , each LC = 300μm in diameter ) or its within-column neuron density D . For cross-column input , we assumed a radial distribution of inputs [60 , 89] and an exponential decay of connection probabilities with distance ( lE = 114μm , lI = 95μm , see above ) . Only pyramidal cells in L2/3 as well as cross-column ( IN-CC ) and far-reaching ( IN-F ) interneurons project across columns . Specifically , the number of input neurons N syn i projecting onto a neuron of cell type i is given by N syn i ( D , L ) = f C i · N D i ( 0 , L C ) + f N i · N D i ( L C , L ) with N D i ( L 1 , L 2 ) = 2 π D · p con i ∫ L 1 L 2 x · exp ( x / l i ) d x . ( 8 ) p con i is the fraction of neurons that connects to cell type i . N D i ( L 1 , L 2 ) denotes the number of neurons within a hollow cylinder defined by the inner radius L1 and the outer radius L2 and the height of 1 mm [60] that are connected to a neuron of cell type i at the center of this cylinder , given a neuron density D . Thus , the two terms represent input from within the same column ( up to LC ) and from outside that column ( up to the full radius L ) . The ratio of pyramidal cells and interneurons that project beyond a single column ( Table 2 ) is reflected in different scaling factors for excitatory and inhibitory connectivity for input from within ( fC ) and from outside the same column ( fN ) . The resulting background currents IN are then computed independently for the four main cell types—pyramidal cells and interneurons in layer 2/3 and 5 . Two in vivo data sets were used for comparison with simulation results ( kindly provided by Dr . Christopher Lapish , Indiana University Purdue University , Indianapolis and Dr . Thomas Hahn , Central Institute of Mental Health and BCCN Heidelberg-Mannheim ) . For spike trains and local field potentials , extracellular multiple single-unit recordings were obtained from the rat’s anterior cingulate cortex ( ACC ) while they were performing an eight-arm radial maze task [40] . Stationary periods ( largely free from motor or sensory responses ) were obtained from 381 units using a previously described stationarity-segmentation method [41] . For the voltage traces , we used patch-clamp recordings from anaesthetized rodents ( see [109 , 110] for details ) . Spike trains , voltage traces and local field potentials from the network simulation and the in vivo data were analyzed the same way . For each model cell or recorded unit , mean and coefficient of variation ( CV ) of the interspike interval ( ISI ) distributions were computed . Autocorrelations of ISI series and the zero-lag cross-correlation CC ( 0 ) between ISIs from pairs of spike trains were computed as well according to the procedures described in [41] to correct for non-stationarities . All analyses were restricted to spike trains of at least 10 spikes to yield sensible estimates of single-cell statistics without cutting off too much of the low-rate tail from the distributions ( see Results section for a more empirical justification ) . Similarity among simulated and experimentally obtained distributions was assessed by two-sample Kolmogorov-Smirnov ( KS ) tests , where test statistic DKS is bounded between zero ( complete overlap ) and one ( maximally dissimilar distributions ) . Underlying distributions were inferred through kernel-density estimation [111] ( implemented by the function “ksdensity” in MATLAB’s statistics toolbox ) . As KS test statistics may depend on sample size but different simulations vary in number of spikes , we limited the number of data points to a sufficiently low , common value ( 30 ) , repeated KS tests with 100 random drawings , and report averages across obtained p values and KS statistics . The overall similarity of a simulation data set with the experimental spike data is quantified by conducting the test for the mean ISI , the CV and the CC ( 0 ) distributions , and reporting the minimal p or the maximal DKS value of those three ( i . e . , the value associated with the largest difference between the compared distributions ) . Finally , we visualize the statistical overlap of distributions by plotting shaded areas representing the SEM around the mean at each value of the experimental distributions , which are computed from the 100 bootstrap samples as indicated above .
|
Computational network models are an important tool for linking physiological and neuro-dynamical processes to cognition . However , harvesting network models for this purpose may less depend on how much biophysical detail is included , but more on how well the model can capture the functional network physiology . Here , we present the first network model of the prefrontal cortex which has not only its single neuron properties and anatomical layout tightly constrained by experimental data , but is also able to quantitatively reproduce a large range of spiking , field potential , and membrane voltage statistics obtained from in vivo data , without need of specific parameter tuning . It thus represents a novel computational tool for addressing questions about the neuro-dynamics of cognition in health and disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"action",
"potentials",
"medicine",
"and",
"health",
"sciences",
"neural",
"networks",
"prefrontal",
"cortex",
"membrane",
"potential",
"brain",
"electrophysiology",
"neuroscience",
"ganglion",
"cells",
"network",
"analysis",
"interneurons",
"computer",
"and",
"information",
"sciences",
"animal",
"cells",
"cellular",
"neuroscience",
"cell",
"biology",
"pyramidal",
"cells",
"anatomy",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"neurophysiology"
] |
2016
|
A Detailed Data-Driven Network Model of Prefrontal Cortex Reproduces Key Features of In Vivo Activity
|
Prostate cancer patients often have increased levels of psychological stress or anxiety , but the molecular mechanisms underlying the interaction between psychological stress and prostate cancer as well as therapy resistance have been rarely studied and remain poorly understood . Recent reports show that stress inhibits apoptosis in prostate cancer cells via epinephrine/beta2 adrenergic receptor/PKA/BAD pathway . In this study , we used experimental data on the signaling pathways that control BAD phosphorylation to build a dynamic network model of apoptosis regulation in prostate cancer cells . We then compared the predictive power of two different models with or without the role of Mcl-1 , which justified the role of Mcl-1 stabilization in anti-apoptotic effects of emotional stress . Based on the selected model , we examined and quantitatively evaluated the induction of apoptosis by drug combination therapies . We predicted that the combination of PI3K inhibitor LY294002 and inhibition of BAD phosphorylation at S112 would produce the best synergistic effect among 8 interventions examined . Experimental validation confirmed the effectiveness of our predictive model . Moreover , we found that epinephrine signaling changes the synergism pattern and decreases efficacy of combination therapy . The molecular mechanisms responsible for therapeutic resistance and the switch in synergism were explored by analyzing a network model of signaling pathways affected by psychological stress . These results provide insights into the mechanisms of psychological stress signaling in therapy-resistant cancer , and indicate the potential benefit of reducing psychological stress in designing more effective therapies for prostate cancer patients .
Psychological stress has been implicated in cancer for almost 2 millennia . It has been observed that psychological stress may contribute to cancer initiation and progression [1] , [2] . However , the causal relationship between stress and cancer remains poorly understood [3] , largely because of limited information about how stress could influence tumor development and drug resistance [4] , [5] . Our recent experiments in an animal model [5] demonstrated that injections of epinephrine or immobilization stress counteracted the anti-tumor effects of PI3K inhibitors on prostate cancer xenografts in mice . Based on these observations , we hypothesized that psychological stress activates anti-apoptotic signaling in prostate cancer cells and , as a result , contributes to the progression of prostate cancer and chemotherapeutic resistance in advanced prostate cancer . Our experiments [5]– have demonstrated that tumor-promoting effects of stress depend on phosphorylation of BAD , a member of the BH-3 only subfamily of Bcl2 proteins . BAD is phosphorylated at Ser112 through the epinephrine-beta2 adrenergic receptor ( β2AR ) -PKA-BAD anti-apoptotic signaling pathway [5]–[7] . BAD can also be phosphorylated by other signaling pathways . For example , epidermal growth factor ( EGF ) triggers phosphorylation of BAD at Ser112 through the EGFR-Raf-MEK/ERK-KinaseX pathway and at Ser136 through the Rac-PAK pathway [8] , whereas activated PI3K transmits signals to Ser136 through AKT activation , and also regulates Ser112 via an unidentified mechanism partially dependent on Akt [8] . To extend analysis of interactions between stress and apoptosis beyond single linear pathway , we used a systems biology approach to study interactions between stress-activated signaling and a regulatory network that controls apoptosis in prostate cancer cells . Several mathematical models of apoptosis regulation have been developed . A Boolean model of apoptosis [9] was proposed to qualitatively analyze the central intrinsic and extrinsic apoptosis pathways and connected pathways . Continuous modeling based on kinetic laws , such as the law of mass action and Michaelis-Menten kinetics , is an alternative approach . Constituted by differential equations , a model of the signaling pathways governing apoptosis [10] demonstrated that inhibition of caspase 3 and caspase 9 resulted in an implicit positive feedback and in bistability . Recently , a mathematic model of Src control on the mitochondrial pathway of apoptosis [11] was designed and fitted to experimental data , used for theoretical design of optimal therapeutic strategies . However , no models have examined interactions between signaling activated by psychological stress , apoptosis , and drug resistance , particularly , resistance to drug combination therapy [12]–[14] . We developed a systems biology model to examine the role of psychological stress in apoptosis regulation and therapeutic sensitivity , and to further analyze the associated signaling pathways activated by stress hormones . By comparing predictive power of two different models with or without the role of Mcl-1 , we predicted that in addition to BAD phosphorylation Mcl-1 expression could be upregulated by stress/epinephrine signaling to inhibit apoptosis . Overall our modeling showed that stress/epinephrine signaling interfered with apoptosis induced in prostate cancer cells by combinations of signal transduction inhibitors .
BAD is a convergence point for several anti-apoptotic signaling pathways in prostate cancer cells . Phosphorylated BAD is critical for the anti-apoptotic effects of such signaling pathways , while dephosphorylated BAD has pro-apoptotic effects . Stress , EGF and PI3K can activate independent signaling pathways that phosphorylate BAD ( Figure 1 ) . These signaling pathways form a convergent network that control apoptosis via BAD phosphorylation . We modeled these signal transduction networks using a system of ordinary differential equations ( ODEs ) to describe the dynamic phosphorylation and dephosphorylation of each protein in the pathways . The model was built according to Michaelis-Menten kinetics [15] using Hill functions [16] , [17] . Our experimental data ( Figure 2 ) demonstrated that the phosphorylation of ERK1/2 peaks under the stimulation of EGF , and then decreases within 1 hour due to the short term signaling of the epidermal growth factor receptor ( EGFR ) [18] . Therefore , we described the de-phosphorylation rates of each protein in the EGFR-Ras-ERK1/2-KinaseX pathway to be dependent on both its phosphorylation and dephosphorylation level and time course as in Equations ( 1–4 ) below , ( 1 ) ( 2 ) ( 3 ) ( 4 ) Where and are maximal activation velocities and Michaelis activation coefficient of each protein by its upstream regulator , respectively . By multiplying the constant dephosphorylation coefficient with time t , Equations ( 1–4 ) can reproduce the signaling curves with peaks followed by later declines . The other signaling regulations regarding phosphorylation or activation of Rac , PAK , PI3K , AKT , PKA , cAMP , PKA , CREB , S112BAD and S136BAD were also modeled by ODEs using Hill functions as described below in Equations ( 5–13 ) , where the dephosphorylation rates were modeled as constants calculated by ensuring the existence of the steady states of these proteins ( see Materials and Methods ) . ( 5 ) ( 6 ) ( 7 ) ( 8 ) ( 9 ) ( 10 ) ( 11 ) ( 12 ) ( 13 ) We then fitted unknown parameters in the model to the experimental data ( see Materials and Methods ) . The estimated parameter values involved in the modeled signaling pathways are listed in Table S1 . Figure 3 shows that the simulations are consistent with the experimental data ( mean squared error between the simulated and experimental data = 0 . 1211 ) . Next we linked the BAD phosphorylation signaling pathways established above to apoptosis percentage . Recently , the preliminary experimental study in our lab indicated that , besides BAD , Mcl-1 may be also involved in stress-mediated apoptosis regulation [19] , [20] . Thus , we considered one model based on BAD phosphorylation only ( see Equation 14 . 1 below ) and one based on both BAD phosphorylation and stabilization of Mcl-1 ( see Equation 14 . 2 which accounts for the potential role of stress-induced activation of CREB , leading to increased transcription of Mcl-1 independent of BAD phosphorylation ) . ( 14 . 1 ) ( 14 . 2 ) ( 15 ) where is cell survival percentage and is apoptosis percentage . is the apoptosis rate in prostate cancer cells . represents total BAD . The additive incorporation of and in Equation ( 14 ) implies that phosphorylation at either S112 or S136 is sufficient to inhibit pro-apoptotic function of BAD , as previously observed [6] , [21] . The potential role of Mcl-1 will be verified by examining its predictive power . The unknown parameters in Hill functions including , , , , , , , and apoptosis rate , , were fitted to our experimental data ( Figure S1 ) by a procedure similar to that above ( see Materials and Methods ) ; estimated values are listed in Table S2 . We did not explicitly model the regulation of apoptosis by some proteins or transcription factors ( e . g . BclXL , BAX and BAK [22] ) involved downstream of our considered pathways . In an implicit approach , indicated by fitting to the evolution of experimental apoptosis percentage ( Figure S1 ) , we modeled the time-dependent nonlinear regulation of apoptosis by multiplying to the right hand of the equation , which resulted in a better data fit . Figure 4 shows prediction of apoptosis percentage in the model with Mcl-1 compared to the experimental data ( mean squared error = 0 . 0221 ) . Here , we theoretically analyzed the stability of the developed system . Let denote the vector of functions in the right hand of the Equations ( 1–15 ) with the vector of proteins phosphorylation considered . Then is Lipschitz continuous with respect to uniformly in the range for any finite T , therefore the developed system continuously depends on the initial values and parameters [23] . We then performed a sensitivity analysis for the estimated parameters ( see Materials and Methods ) . Each parameter was increased by 1% from its estimated value , and then we obtained the time-averaged percentage change of each variable value . All sensitivity values were not more than 1 . 4327% ( Figure 5 ) . The sensitivity analysis result confirmed that the developed system is conserved through the modest parameter changes and our model is rather robust . Currently several inhibitors that target the BAD upstream signaling network are in clinical trials , including PI3K inhibitors ( e . g . LY294002 ( LY ) , CAL-101 , BKM120 , and GDC-0941 ) , EGFR inhibitors ( e . g . gefitinib , erlotinib HCl ) and MEK inhibitors ( e . g . AZD6244 , GSK1120212 ) . As shown in Figure 1 we considered 8 pharmacological and dominant negative inhibitors of signaling downstream of the EGF , PI3K and psychological stress pathways: the PI3K inhibitor LY294002 , the EGFR tyrosine kinase inhibitor AG1478 , the Rac inhibitor N17Rac , the PAK inhibitor DN-PAK1 , the RAF inhibitor C4BRaf , the MEK/ERK1/2 inhibitor DN-MEK1 , the PKA inhibitor PKI-GFP , and BADS112A as a functional equivalent of an inhibitor of BAD phosphorylation at S112 [5] , [6] . The inhibition effect of LY294002 ( LY ) was modeled in Equation ( 7 ) using an inhibition Hill function . Inhibition effects of the other inhibitors were also modeled by multiplying an inhibition Hill function to the maximal reaction velocity ( see Equations 1–3 , 5 , 6 , and 10 , respectively ) . We integrated these inhibition effects by redefining each as: ( 16 ) where is the Michaelis-Menten constant indicating the concentration of drug that decreases the maximal reaction velocity to half the original value without drug treatment . In this work we normalized the concentration of drug to . As a result , the non-dimensional value of the drug concentration became . Thus , we did not introduce any additional parameters into the model . Mutant BADS112A inhibits the anti-apoptotic role of phosphorylated pS112BAD by decreasing the relative ratio of phosphorylated pS112BAD to dephosphorylated S112BAD that binds BclXL and promotes apoptosis . Therefore , we assumed that BADS112A decreases the relative level of steady phosphorylation of S112BAD , which was modeled by integrating drug effects into the dephosphorylation rate of S112BAD as follows , ( 17 ) To investigate the potential role of Mcl-1 transcription in anti-apoptosis , we compared two different models of anti-apoptosis regulation: one based on BAD phosphorylation only , and one based on both BAD phosphorylation and stabilization of Mcl-1 , as modeled in Equations 14 . 1 and 14 . 2 , respectively . The predictions of apoptosis percentage under different treatments or conditions were compared to the experimental data [6] , [8] ( Figure S1 ) . Experimental data were normalized to the same experimental environment . The prediction of apoptosis percentage for EGF&LY&C4BRaf&DNPAK1 in the first model ( Figure 6A ) was not consistent with the experimental data . The second model ( Figure 6B ) improved validation and presented better predictive power , and emphasized the potential role of Mcl-1 in anti-apoptotic effects of emotional stress/epinephrine . Our selected model , the second model with Mcl-1 , predicted that LY294002 , LY294002 & C4BRaf , LY294002 & DNPAK1 , LY294002 & DNPAK1 & C4BRaf should have similar effects on the percent apoptosis of cancer cells , which was consistent with the experimental data ( Figure 6B ) . The prediction that LY294002 plus BADS112A would produce the best pro-apoptotic effect was experimentally validated . Moreover , addition of EGF or activation of PKA signaling by epinephrine inhibited apoptosis induced by a single inhibitor or a combination , shown both in the model and experimentally . The agreement between the predicted and the experimental results confirmed that our model can quantitatively predict apoptosis percentage of prostate cancer cells under various treatments and different conditions . Then we investigated the effects of combined signaling inhibitors on apoptosis percentage with or without EGF and/or epinephrine . Since the combination of more than 3 drugs is less realistic for clinical purposes and may lead to unknown side effects , we limited our considerations to a combination of two inhibitors . The dose of each inhibitor in the pairs was set as 1 , so the total dose of each combination was 2 , which was the same for one single inhibitor “combined” with this inhibitor itself . Figure 7A shows the apoptosis percentages induced by inhibitor combinations under conditions without EGF and epinephrine . The signaling pathways stimulated by EGF and psychological stress were inactivated and the apoptosis percentage was effectively promoted by all inhibitors . LY294002 showed a strong pro-apoptotic effect as a single treatment or combined with other inhibitors , and BADS112A had less effect . Figure 7B shows the combinatorial effects of inhibitors with EGF but no psychological stress . The apoptosis percentages were decreased compared to Figure 6A . However , LY294002 combined with BADS112A demonstrated a much stronger pro-apoptotic effect compared to other combinations . Figure 7C shows the effects of inhibitor combinations plus epinephrine . Pro-apoptotic effects of all combinations of inhibitors , except for BADS112A with LY294002 , were inhibited by stress-activated signaling . Finally , when both EGF and epinephrine were present , pro-apoptotic effects of all inhibitor combinations were substantially decreased ( Figure 7D ) . These results demonstrate variability of apoptosis induction by different combinations of inhibitors , in the presence of agents that activate anti-apoptotic pathways . Based on our modeling , the combination of BADS112A and LY294002 produces the greatest effect on promoting apoptosis in prostate cancer cells . Therefore , we tested whether this combination of BADS112A and LY294002 is synergistic [24] , [25] . We first adopted the Loewe additivity [26]–[28] to quantitatively evaluate and examine the synergism of LY294002 plus BADS112A . The Loewe combination index is defined as a ratio of total effective drug dose ( combination versus single drug ) required to achieve a given effect as follows: ( 18 ) where d1 ( BADS112A ) and d2 ( LY ) are the doses in the combination isobologram with respect to the percentage of apoptotic cells . and represent the concentration of BADS112A and LY294002 with respect to promoting apoptotic cells by percentage , respectively . CILoewe<1 , CILoewe>1 and CILoewe = 1 indicate Loewe synergy , antagonism , and additivity , respectively . Figure 8 shows that 25% isobologram of BADS112A and LY294002 ( blue curve ) bows inward , indicating CILoewe<1 . Therefore the combination of BADS112A and LY294002 is synergistic regarding the 25% apoptosis isobologram . To calculate the Loewe index requires solving a reverse problem based on an isobologram . Thus , this approach requires a high computing cost and consideration for specific isobolograms . Another quantification method for combination therapies is Bliss independence [27] , [29] . But the calculation of this qualification index resulted in negative expected apoptosis percentage values of combined inhibitors , which is not realistic . Thus , indicated by ( but different from ) the Bliss index , we defined a new combination index as follows: ( 19 ) where is apoptosis percentage induced by doses of inhibitor 1 , and is apoptosis percentage induced by doses of inhibitor 2 . is the apoptosis percentage promoted by combined inhibitor 1 and inhibitor 2 with dose and dose , respectively . With the same total doses , if the combined inhibitors produce a greater effect than both single inhibitor 1 and inhibitor 2 , the index considers that these two inhibitors work synergistically . Therefore , the index considers the combination as a synergism effect if CI <1 , as antagonism if CI>1 , and otherwise additivity . We evaluated dose-dependent synergism of combined BADS112A and LY294002 as defined in Equation ( 19 ) with or without psychological stress . In the simulation , the dose of each inhibitor ranged from 0 . 01 to 100 . In the no or low psychological stress environment , BADS112A plus LY294002 has a synergistic effect , but in the high psychological stress environment , the synergism pattern switched ( Figure 9 ) . The synergism pattern was divided into two regions: one with CI<1 indicating synergism and another with CI≥1 corresponding to antagonism or additivity . Therefore , psychological stress triggered the synergism pattern switch to a dose-dependent combination synergism . Under the high psychological stress condition , only if the doses of BADS112A and LY294002 were high enough , did their combination produce synergism with respect to promoting apoptosis of cancer cells . Stress could decrease the efficiency of anti-cancer therapy ( Figure 1 ) . A dose-dependent response of BADS112A and LY294002 combination therapy in Figure S2 further demonstrates drug resistance induced by psychological stress . When the stress ( or epinephrine ) was absent , the apoptosis percentage was slightly affected by the doses of LY294002 and BADS112A and remained at a high level . While when the psychological stress emerged , high doses and low doses of LY294002 resulted in different levels of apoptosis percentage , even when combined with the high doses of BADS112A . The drug resistance induced by stress was consistent with the switch of synergism pattern as demonstrated above . We then examined the differences in signaling pathways with or without psychological stress with combination therapy . When there was no psychological stress , the epinephrine-β2AR-cAMP-PKA-CREB signaling pathway was not activated . PI3K-AKT pathway was inhibited by LY294002 , and the relative phosphorylation of BAD at S112 and S136 was repressed to a low level around 0 . 1 ( Figure 10 ) . When psychological stress was introduced , the epinephrine-β2AR-cAMP-PKA signaling pathway was activated leading to phosphorylation of BAD at S112 , which counteracted the repression of BAD phosphorylation at S112 induced by LY294002 and BADS112A . As a result , the relative phosphorylation of BAD at S112 returned to a higher level . In addition to phosphorylation of BAD at S112 and s136 , Mcl1 ( or CREB ) activated by stress signaling could also inhibit apoptotic , so percent apoptosis in cancer cells was decreased compared to the no-stress condition ( Figure 10 ) . Therefore , the differentially activated signaling pathways stimulated by psychological stress , leading to both BAD phosphorylation and Mcl-1 activation , were responsible for the drug resistance and synergism pattern switch in combination therapy .
Our modeling strategy successfully captured key kinetic features of the underlying signaling pathways discussed above . We did not describe the kinetics in the pathway by linear equation based on mass action law , since the detailed reaction was unclear and ignorable . Instead , we incorporated by Michaelis-Menten kinetics using the Hill function [16] , [17] to integrate less critical reaction details . Based on experimental data , we phenomenologically modeled the rate of change for dephosphorylation of proteins in EGFR-ERK1/2 pathway and apoptosis regulation to be time dependent . The simulation results ( Figure 3 ) were consistent with experimental data ( Figure 2 ) , which suggested the fundamental signaling networks used in this work were reliable . In future work , we will integrate elements downstream of BAD , such as BclXL , BAX and BAK [22] , to investigate a more detailed mechanism related to stress interactions in prostate cancer . The anti-apoptotic role of BAD phosphorylation mediated by emotional stress has been well studied . Recently , our lab found that , besides BAD , Mcl-1 may be also involved in stress-mediated apoptosis regulation ( Hassan et al unpublished data ) . Here , we applied a systems biology approach to investigate the potential role of Mcl-1 stabilization in anti-apoptotic effects of emotional stress/epinephrine , which was verified by comparing the predictive power of two different models with or without the role of Mcl-1 . The selected model with better predictive power will be used to explore effects of stress on Mcl-1 in our ongoing experiments . Effects of drugs on apoptosis varied depending on which components in the signaling network were targeted . This was due to kinetic asymmetry of different signaling pathways . As shown in Figure 1 , PI3K/AKT pathway can phosphorylate BAD at both S112 and S136 [8] , so the inhibition of this pathway by PI3K inhibitor LY294002 could induce more cell death compared to other inhibitors , such as N17Rac or DN-PAK1 , that target pathways that phosphorylate only one site of BAD . Expression of phosphorylation-deficient mutant BADS112A can also effectively promote apoptosis [5] , [6] . Finally , drug-induced signaling network remodeling is an important and interesting question for future work . Psychological stress and anxiety are often experienced by prostate cancer patients . The increased psychological stress that can result from cancer progression and diagnosis strengthens the activation of anti-apoptotic signaling pathways [19] , as demonstrated in our simulation , which could decrease therapy efficiency and shift drug combinations from synergy to antagonism . These results also suggest the need for deeper analysis of the role of stress-related signaling in other therapy-resistant cancers . In summary , we developed a dynamic network model of signaling pathways that control apoptosis in prostate cancer cells to study the role of psychological stress on prostate cancer therapy , and justified the role of Mcl-1 stabilization in anti-apoptotic effects of emotional stress . A drug resistance and synergism switch was revealed in our model , and the associated signaling mechanisms were explored .
We collected data at both the molecular and cellular levels . The molecular data regarding protein phosphorylation included two sets of Western blotting images [6] , [7] , both done in LNCaP cells . In the first experiment ( Figure 2A ) , cells were treated with 50 µm LY294002 for 2 hours followed with increasing concentrations of epinephrine ( 0 . 01–1000 nm ) for 1 h . BAD phosphorylation at Ser112 and CREB phosphorylated at Ser133 were measured . The second set of data ( Figure 2B ) contains the time course of protein phosphorylation in cells treated with LY294002 followed with EGF 2 hours later . Phospho-Ser473 Akt , phospho-Thr308 Akt , total Akt , phospho-ERK1/2 , total ERK1/2 , phospho-Ser112 BAD , and total BAD were measured for the indicated times . LY294002 , inducing dephosphorylation of HA-BAD at Ser112 and Ser136 , was followed by Western blot analysis ( Figure 2B ) . We quantified the Western blotting data using ImageJ software and the normalized values were listed in Table S3 . For the first experimental data set , there were 8 conditions with or without LY294002 treatment and with increasing concentrations of epinephrine . The concentration of phosphorylated BADs112 was normalized to the control condition without LY294002 treatment and epinephrine . Phosphorylated CREB was normalized to the maximal concentration , since the concentration in the control condition was minimal . For the second experimental data set , there were 10 treatment conditions with different time periods of LY294002 and EGF treatment . Concentrations of phosphorylated S473Akt , S112BAD , and S136 BAD were normalized to the control condition ( neither LY294002 nor EGF treatment ) . Phosphorylated ERK1/2 concentration was normalized to total ERK1/2 concentration , since the concentration of phosphorylated ERK1/2 under the control condition was almost zero . These data were used to estimate the parameters in Equations ( 1–13 ) . The cellular level data from [6] , [8] were apoptosis percentages determined by counting at least 350 cells in several randomly chosen fields for every treatment . Considering that the experimental data were conducted in different experimental environments , we scaled the data in Figure S1B , C to the data in Figure S1A to ensure the apoptosis percentages under the treatments of LY294002 and LY&EGF were at the same levels . Treatments of LY , LY&EGF , and LY&VIP were used to estimate the parameters in Equation ( 14 ) ; the remaining data were used to validate model predictions . We measured apoptosis by several independent methods: 1 ) caspase assay – a quantitative assay that measures activity of effector caspase 3 against fluorogenic substrate DEVD-amc [6]; 2 ) time-lapse video microscopy- a quantitative assay that follows morphological changes of individual cells over 24 hours [6]; 3 ) western blotting for apoptosis markers– cleaved caspase 3 , caspase 7 and cleaved PARP , this is qualitative assays that confirms activation of caspases and cleavage of physiological substrate in dying cells [30]; 4 ) immunofluorescent staining for active caspase 3 and release of cytochrome c from mitochondria– a specific hallmark of apoptosis [30] , [31]; 5 ) TUNEL assay , this assay detects cleaved DNA – specific hallmark of apoptotic cell death [20] . Of these methods caspase assay and time lapse video microscopy are considered most appropriate to quantitatively measure apoptosis . Other methods confirm that cell death is indeed by apoptosis mechanism . These methodologies are consistent with published “Guidelines for the use and interpretation of assays for monitoring cell death in higher eukaryotes” [32] . We estimated the unknown parameters in the model by fitting the simulation results to the experimental data described above . Equation ( 20 ) was employed for parameter estimation by minimizing the fitness error between the experimental and simulated data , ( 20 ) where and represent the simulated and experimental data with parameters under condition , respectively . stands for the parameter space , in which the search space for each parameter was preset in a range according to the experimental observations and Michaelis-Menten kinetics . According to the experimental data ( Figure 2 ) , we set the initial values of Equations ( 1–13 ) as the vector ( 0 , 0 , 0 , 0 , 0 , 0 , 1 , 1 , 0 , 0 , 0 , 1 , 1 ) in the simulation . To further reduce the numbers of unknown parameters , the parameters , , were calculated by ensuring the existence of the steady states of the system , for example , the dephosphorylation rate of PI3K , d8 , was set as ( 21 ) where and are steady states of PI3K and AKT which are assumed as 1 equal to their initial concentrations , respectively . The remaining parameters , including , ( = 1 , 2 , … , 13a , 13b ) and d1 , d2 , … , d4 , were estimated using the above optimization procedure , for a total of 37 parameters in Equations ( 1–13 ) that were estimated by fitting to 56 experimental data points under different conditions . Similarly , 7 parameters in Equation ( 14 ) were estimated by fitting to 27 experimental data points . A genetic algorithm [33] was adopted to minimize the cost function in Equation ( 20 ) . The system of nonlinear ODEs was numerically solved using the 4th Runge-Kutta method . The model simulation and result analysis were performed in MATLAB R2007b ( MathWorks , USA ) . Parameter sensitivity analysis examines whether a system is preserved to the modest parameter changes and quantitatively explores the sensitive parameters . We used parameter sensitivity analysis to study the relationship between the proteins , apoptosis percentage and the variations for each parameter value . The relative sensitivity coefficient [34] of a variable at time t with respect to a parameter was computed by: ( 22 ) Time-averaged sensitivities were calculated according to ( 23 ) where is an equal partition of . In the simulation , was set as 100 and as 10 . Each parameter was increased by a small perturbation , for instance 1% , from its estimated value , and then we obtained the time-averaged percentage change of each variable value .
|
Psychological stress and anxiety are often experienced by prostate cancer patients , but the underlying mechanisms of interactions between psychological stress and cancer development , as well as drug resistance , are unclear . Here , we employed a systems biology approach to study interactions between stress-activated epinephrine/beta2 adrenergic receptor/protein kinase A signaling and a regulatory network that controls apoptosis in prostate cancer cells . We developed a dynamic network model of signaling pathways that control apoptosis in prostate cancer cells and quantitatively evaluated the effects of stress-activated signaling on apoptosis induced by drug combinations . Experimental data were used to guide modeling , to fit the unknown parameters and validate the model . Based on our model we found that epinephrine/beta2 adrenergic receptor/protein kinase A signaling can decrease drug efficiency , and can shift the effect of drug combination from synergy to antagonism . We also predicted that in addition to BAD phosphorylation Mcl-1 expression could be upregulated by stress/epinephrine signaling to inhibit apoptosis . This study provides insights into the mechanisms of psychological stress signaling in therapy-resistant cancer , and suggests that reducing psychological stress could help to make prostate cancer treatment more effective .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
Systems Modeling of Anti-apoptotic Pathways in Prostate Cancer: Psychological Stress Triggers a Synergism Pattern Switch in Drug Combination Therapy
|
Recently available cancer sequencing data have revealed a complex view of the cancer genome containing a multitude of mutations , including drivers responsible for cancer progression and neutral passengers . Measuring selection in cancer and distinguishing drivers from passengers have important implications for development of novel treatment strategies . It has recently been argued that a third of cancers are evolving neutrally , as their mutational frequency spectrum follows a 1/f power law expected from neutral evolution in a particular intermediate frequency range . We study a stochastic model of cancer evolution and derive a formula for the probability distribution of the cancer cell frequency of a subclonal driver , demonstrating that driver frequency is biased towards 0 and 1 . We show that it is difficult to capture a driver mutation at an intermediate frequency , and thus the calling of neutrality due to a lack of such driver will significantly overestimate the number of neutrally evolving tumors . Our approach provides precise quantification of the validity of the 1/f statistic across the entire range of all relevant parameter values . We also show that our conclusions remain valid for non-exponential models: spatial 3d model and sigmoidal growth , relevant for early- and late stages of cancer growth .
Darwinian evolution in cancer has been the subject of intense research in the past decade . In particular , the problem of distinguishing driver mutations that carry a selective advantage from passenger mutations , and their role in shaping intra-tumor genetic heterogeneity has come to the fore [1–5] . Determining which mutations in cancer are drivers and which are passengers is one of the most pressing questions in cancer genomics , as identification of new driver mutations can contribute to development of new targeted therapeutics [6 , 7] and personalized medicine [8] . Numerous methods for classifying driver and passenger mutations and measuring selection in cancer have been developed , including those that identify driver genes based on how frequently they are mutated [2] , specific mutation patterns [9 , 10] , and dN/dS ratios [1 , 11] . These methods can reliably identify driver genes mutated in a high proportion of tumors of a given type ( >20% ) ; using such methods to find less common drivers would require a large number of cancer samples [12] , and drivers unique to a single or a small number of patients could still be missed . Several recent papers attempt to measure the magnitude of selection operating during cancer evolution using the frequency distribution of subclonal mutations in an individual patient’s cancer . In a seminal paper , Williams et al . used mutant allele frequencies to conclude that a significant fraction ( ~1/3 ) of cancers evolve neutrally [13] . Subsequent studies focused on quantifying the strength of selection and distinguishing it from “effectively neutral” cancer evolution [14 , 15] . These works [13 , 15] are based upon the assumption that drivers that arose after cancer initiation will be present at a macroscopic but clearly subclonal frequency ( i . e . “detectable” ) , which will make the cumulative mutant allele frequency look different to the 1/f power law expected from neutral evolution . Here we use a branching process model of cancer evolution to derive a formula for the probability of detection of a subclonal driver , and test the validity of the proposed 1/f statistic across all relevant parameter combinations .
We consider a two-type stochastic model of cancer evolution ( Fig 1a ) . In the model , cancer is initiated by a single transformed cell . Progeny of this cell follow a branching process with birth rate b and death rate d . We set r = b − d > 0 , so the population grows if it survives initial stochastic fluctuations . In addition , cancer cells can obtain a new driver mutation with rate u ( see Materials and methods ) . Cells with the driver mutation replicate with rate b1 and die with rate d1 smaller than b1 ( Fig 1a and 1b ) . The subpopulation of driver-carrying cells has therefore a net growth rate r1 = b1 − d1 , and we assume that r1 > r so that the additional driver increases the net growth rate by the factor c = r1/r > 1 . We define g = c − 1 as the relative increase in the growth rate due to the driver . We are interested in the frequency of cancer cells that carry the driver mutation . In a neutral process ( g = 0 ) , mutation frequency stabilizes and remains approximately constant at large times [16] . For g > 0 ( driver with a selective advantage ) , the frequency of cells with the driver increases from ≈0 to ≈100% during tumor expansion . We denote by F[α] the probability that subclonal driver frequency f is smaller than α . We show ( Materials and methods ) that this cumulative frequency distribution reads F[α]≈∫0Mcub1exp ( -cub1X ) [1-exp ( -crb1α ( 1-α ) c ) XcM1-c]dX ( 1 ) where denotes M is number of cells in the tumor . Formula ( 1 ) is in excellent agreement with exact computer simulations of the branching process ( Materials and methods ) . Recently , similar two-type processes were studied by Kessler and Levine [17] and Cheek and Antal [18] , who derived generating functions [18] and probability distributions for the size of the mutant population in the case of small mutant frequencies [17] in the general asymmetric ( different fitness of wildtype and mutant cells ) Luria-Delbruck model with death . In contrast , our formula ( 1 ) describes the probability distribution for a single driver subclone that has reached non-negligible frequencies . We note that Eq ( 1 ) concerns the cancer cell frequency of a subclonal driver , and not the variant allele frequency obtained from genomic analysis . As noted recently [19] , variant allele frequency does not automatically indicate a certain cancer cell frequency ( due to contamination with normal cells and variable ploidy ) , and using allele frequencies instead of cancer cell frequencies to detect selection can be an additional source of bias . We assume that a subclonal driver mutation can be detected , and able to skew the 1/f power law expected from neutral evolution , when its cancer cell frequency is between 20% and 80% . This range is much wider than the range 24% to 48% used in Williams et al . [13] ( mutant allele frequency range 12% to 24% ) . Thus , the probability that a driver can be detected is given by Pdet=F[0 . 8]-F[0 . 2] ( 2 ) For moderate levels of selection , e . g . , when the additional driver mutation increases the growth rate by g = 30% , the probability that the driver mutation is in the detectable range ( [0 . 2 , 0 . 8] ) is <15% for population sizes up to M = 109 cells , and remains below one third for M ≤ 1011 cells ( Fig 2a ) . For other cases considered here ( 70% and 100% increase in the net growth rate ) , the chance of detecting the subclonal driver is always <60% and—for a broad range of tumor sizes—less than 30% . The parameters used in Fig 2a are from Bozic et al . [20] , and are typical for a moderately aggressive cancer ( net growth rate 0 . 01/day ) . We show that the situation is qualitatively similar for faster and slower growing cancers in Fig 2b and 2c . In summary , for moderate levels of selection ( g = 30% ) , the chance of detecting a subclonal driver is small for almost any tumor size , and very strong selection ( g = 100% ) will be detectable only in small cancers . Strong selection ( g = 70% ) will be detectable at intermediate-size , moderately growing tumors; large , fast-growing tumors; or small , slow-growing tumors . Most notably , for all parameter values , even in the parameter regimes where the probability of detecting the subclonal driver is the highest , it is still below 60% ( Fig 2a–2c ) . Our model suggests that detecting deviation from neutral evolution is challenging , as there is a significant chance that a subclonal driver will not be in the detectable range . The reason is that the frequency of cells with the new driver is biased toward 0 and 1 . When the tumor is small , the fraction of driver-carrying cells is very close to zero , as there has not yet been enough time for the fitter subpopulation to expand ( Fig 2d ) . In contrast , for large tumors , driver-carrying subpopulation has already expanded and completely dominates the population , so its frequency is close to 100% ( Fig 2e ) . Interestingly , for sizes at which the chance of detecting the subclonal driver is highest ( close to 60% ) , the frequency distribution is almost flat ( Fig 2f ) . In Fig 2 we used a previously estimated driver mutation rate u = 10−5 per day [21] . To explore the effect of a higher or lower driver mutation rate on our conclusions , we first used recently published genomic data to determine an upper bound ( 10−3 ) and a lower bound ( 10−7 per day ) on the driver mutation rate ( Materials and methods ) . We next performed a numerical grid search on the space of all parameters ( driver mutation rate u , relative growth rate advantage of a driver g , net growth rate of tumor cells r , division rate of cells with the driver b1 , and final number of tumor cells M ) . A wide range of values is taken for each parameter , including driver mutation rate u between 10−7 and 10−3 per day , and growth advantage of a subclonal driver g between 1% and 200% ( see Materials and methods for more details ) . The grid search demonstrated that the probability of detection of a subclonal driver is always below 60% , and that subclonal driver frequency is biased towards 0 and 1 across the entire range of reasonable parameter values of the carcinogenic process ( Materials and methods , Fig 3 ) . The intuitive reason behind this result is that the probability density function for subclonal driver frequency is convex across this entire parameter range ( examples in Fig 2d , 2e and 2f ) . The well-mixed model discussed so far does not include spatial constrains experienced by solid tumors . We thus extended our analysis to a similar two-type process in three-dimensional space , using a lattice-based computer model [5] . We considered a version of the model in which cells that occupy points of a 3d lattice replicate to empty neighboring sites , die , and mutate , but do not migrate [5] . Normal cells replicate and die with rates b , d , whereas mutant cells replicate with rate b but die with rate d1 . Similarly as before , the ratio of net growth rates of cells with and without the driver is given by c = 1 + g = ( b − d1 ) / ( b − d ) >1 . To simulate a large number of tumors for realistic sizes , we scaled the parameters of the model so that 1 site corresponded to 100 cells , and mutation probability was 100x larger . S1 Fig shows the probability of detection of a subclonal driver as a function of tumor size for medium , fast and slow tumor growth , and for three levels of driver potency ( g = 30% , 70% and 100% increase in net growth rate b-d ) . These results demonstrate that our main conclusion that many drivers elude detection remains true in the 3d model . In particular , out of 36 parameter combinations evaluated , the probability of detection of a subclonal driver was below 60% in 33 out of 36 cases , and in the remaining 3 cases the probability was 63% , 67% and 68% . The probability of detection was 40% or below for one half of the parameter sets . Of note , we observe that in the 3d model there is a higher chance of detection of moderate drivers ( g = 30% ) compared to the well-mixed model , and the chance of detection peaks at lower tumor sizes compared to the well-mixed case . The growth of tumor populations is complex and may change throughout the carcinogenic process . For example , initial growth may be slower than exponential due to tissue constraints and nutrient availability , exponential-like after angiogenesis , and may slow down again when the tumor is very large . To model this more complex scenario , we use the 3d model described above to model the first , avascular stage , before tumor reaches 106 cells [22 , 23] . As 106 cells is typically at the lower end of tumor detectability , it is unlikely that drivers will be detected at this stage . To model the later stages of tumor growth ( exponential and slow-down ) , we employ the following system of differential equations: dxdt=rx ( 1-x+yK ) dydt=cry ( 1-x+yK ) Here x is the size of the type-0 and y is the size of the type-1 ( driver ) population , c = 1+g >1 is the ratio of their initial growth rates and K is the carrying capacity of the tumor . We use a deterministic model because both wild-type and driver populations are likely to be large at the end of the spatial phase . To combine the 3d and the above model , we record the sizes of the type-0 and type-1 populations obtained from the 3d simulation when total population size reaches 106 cells , and use them as initial conditions for the system described above , which we solve numerically . We show results for probability of driver detection in this sigmoidal model in S2 Fig . In this sigmoidal growth model , in contrast to the well-mixed and 3d models , driver fraction does not approach 1 for large tumors . Instead , it will reach a stable frequency that depends on the initial driver fraction and model parameters . For all parameter combinations we evaluated , the chance that this final driver frequency is between 0 . 2 and 0 . 8 is below 63% . We note that the sigmoidal model increases the chance of detection of moderate drivers ( g = 30% ) compared to the well-mixed model , but decreases the chance of detection of strong drivers ( g = 70% and g = 100% ) , as they will expand to the carrying capacity quickly and not leave much room for the type-0 population . In this model , selection will in general be more detectable for slower growing compared to faster-growing tumors . In sum , our results imply that in the sigmoidal model , low to moderate levels of selection will be most detectable at the final ( close to carrying capacity ) stage , whereas strong drivers will be most detectable at the beginning of the exponential-like phase . We note , however , that sigmoidal growth models may produce different behaviors depending on the specifics of the competition between the two populations—for example not only initial growth rates but also the levels of growth inhibition may differ between populations , and cell turnover while the population is at carrying capacity may lead to competitive exclusion of the less fit population . Finally , we also considered how the appearance of a second driver within the first driver population would influence probability of detection of either driver for realistic parameter values in the well-mixed model . To that end , we derived the expression for the probability that the second driver will be in the detectable range , assuming the first driver is not detectable , P2≈∫0 . 81f1 ( α1 ) ( F2 ( 0 . 8/α1 ) -F2 ( 0 . 2/α1 ) ) dα1 Here f1 is the probability density of the first driver frequency ( Eq ( 7 ) in Materials and methods ) and F2 is the cumulative probability for the frequency of the second driver in the first driver population ( Eq ( 1 ) with appropriate parameter values ) . For the same parameters as in Fig 2 , and assuming that second driver increases the net growth rate by the same absolute amount as the first , we show that the probability that driver frequency is in the range [0 . 2 , 0 . 8] is always below 69% ( S3 Fig ) . In addition to this upper bound , we also calculate the average probability of driver detection across all parameter values evaluated in S3 Fig . The average detection probability is 43% when one includes driver mutations in cancer cell frequency ( CCF ) range [0 . 2 , 0 . 8] . The average driver detection probability for the CCF range [0 . 24 , 0 . 48] from the original Williams et al . study [13] is 18% .
In sum , the fact that no subclonal driver is present at intermediate frequencies cannot be taken as proof of neutral evolution . It can simply be a consequence of population dynamics which creates only a short window during which the driver mutation can be detected but has not yet dominated the population . Tarabichi et al . [24] and McDonald and colleagues [25] simulated tumor evolution in which they explicitly include selection , and showed that , even in models with selection , mutant allele frequency can exhibit the 1/f power law behavior , resulting in incorrect calling of neutrality . In response , Williams and colleagues [26 , 27] argue that the example simulations from Tarabichi et al . [24] and McDonald et al . [25] that were incorrectly classified as neutral use extreme parameter values or correspond to either strong and early selection ( a driver mutant quickly sweeps to fixation ) , or weak and late selection ( driver mutants unable to reach detectable frequencies ) . In contrast , we show here that , for almost any driver mutation rate and selection strength , whenever we look at the mutant frequency spectrum of a tumor , it is likely either too early and the driver is present at a very low frequency , or it is already too late , and the driver is present in almost all cells of the tumor . Importantly , even if we manage to obtain the mutant frequency spectrum during the optimal window for detection , there is still significant chance ( close to half ) that the subclonal driver will not be in the detectable range . Thus , even though multiple studies [13 , 28] ( including ours [16] ) have confirmed that the experimental allele frequency spectrum of many cancers agrees with the spectrum of a neutral model in a certain frequency range , we argue that this agreement should not be taken as evidence of neutral evolution . Simulations of branching processes of cancer evolution for realistic tumor sizes and parameter values are computationally expensive . To circumvent that , studies often use small population sizes , death rate of cancer cells much smaller than the birth rate , and only examine a small set of different parameter values . In contrast , our mathematical results ( formula 2 ) can be quickly evaluated for realistic parameter values , including all biologically plausible values of selection , mutation , birth and death rates , and population size . Furthermore , our results explain why the deviation of the mutant allele frequency from the 1/f power law in an intermediate frequency range is not a sensitive statistic for detecting subclonal selection in models of exponentially growing cancer populations: mutational frequency distribution of a subclonal driver is convex and thus always biased toward 0% or 100% frequency . In this paper , we study a process in which we explicitly include selection , and show that a subclonal driver , though present , may often fail to change the VAF distribution expected from neutral evolution . We note that , if there is no subclonal driver present , i . e . when driver mutation frequency is precisely f = 0 or f = 1 , the neutral test developed by Williams et al . [13] will be correct . However , due to the finite resolution with which we can distinguish different mutation frequencies using current sequencing techniques , it will be difficult if not impossible to determine that a mutation is present at precisely 0 or 1 frequency ( e . g . experimental f = 0 . 9 may correspond to f = 1 due to sequencing errors and vice versa; low frequency mutations may have experimental f = 0 due to insufficient sequencing depth ) . Our conclusions do not contradict the results of many recent genomic studies that find large subclones in majority of sequenced cancers [29–31]; many of these subclones may in fact be lacking functional driver mutations and/or can be a consequence of genetic drift . For example , a recent genomic study of chronic lymphocytic leukemia [30] ( CLL ) reports that the majority of macroscopic ( >10% cancer cell frequency ) CLL subclones seem to be passenger subclones that lack selective advantage over their parent subclones . Our results are in agreement with the recent Williams et al . study [15] , which found that 21% of colon cancers , 29% of gastric cancers and 53% of metastases they examined had evidence of differentially selected ( driver ) subclones . These estimates are in line with our predictions of how likely it would be to detect selection even if 100% of tumors were non-neutral . On the other hand , Nik-Zanai et . al . [32] find dominant subclones ( >50% CCF ) in all 21 breast tumors they studied , and argue that these dominant subclones are likely to have been selected ( i . e . that they are driver subclones ) . We argue that detecting subclonal drivers is in general challenging; if all dominant subclones from Nik-Zanai et al . [32] did in fact contain drivers , it may mean that these tumors have evolved differently from the model and parameters we have assumed in this paper . Yates et al . [31] find subclonal driver mutations in 15/50 ( 30% ) of breast cancers they studied; Turajlic et al . [33] found subclonal driver mutations in 120/216 ( 56% ) of primary and metastatic renal cancers they sequenced . The fractions of samples with detectable subclonal driver mutations in these two studies are in line with the average driver detection probability in our model with two drivers , which is 43% . The two-type model we have studied here is a simplification of the process of driver accumulation in tumors . Our model is deliberately oversimplified to allow for analytic treatment and to develop an intuition why subclonal drivers may elude detection . In reality , detecting subclonal drivers will be even more difficult due to experimental uncertainties , confounding passenger mutations , and a contribution from contaminating non-cancer cells . For example , we and others [13 , 15] implicitly assume that genomic analysis of biopsy samples identifies true cancer cell frequencies of mutations in the entire tumor population . However , significant spatial heterogeneity may exist in solid tumors , and a minor subclone may be dominant in certain spatial regions of the tumor and overrepresented in a tumor biopsy , resulting in possible misleading conclusions derived from the mutational frequency spectrum of an unrepresentative sample [34] . There is a debate about mini-drivers in cancer [35] , referring to mutations that are in between strong and/or very frequently mutated drivers on the one hand , and neutral passenger mutations on the other . Our study demonstrates that strong drivers ( growth rate increase g = 70% ) are most likely to be detected from macroscopic subclonal frequencies for moderately-growing tumors and a typical driver mutation rate ( u~10−5 ) . For higher driver mutation rates and/or slower-growing tumors , growth advantage of most likely detectable drivers decreases . In particular , our results demonstrate that mini-drivers will most likely be detectable in very slow-growing tumors , that are already close to or at carrying capacity . Finally , our conclusions are relevant not only to cancer but more generally to the problem of measuring selection when an expanding subpopulation of fitter cells coexist with “wild-type” cells , such as growing bacterial populations acquiring de novo resistance to antibiotics or adapting to a new environment .
We study a two-type continuous time branching process that starts with a single type-0 cell . With rate b , type-0 cells divide into two identical daughter cells . Death rate of type-0 cells is d , with b > d . In addition , type-0 cells can receive an additional driver mutation with rate u . We will assume that the driver mutation rate is very small , on the order of u ~ 10−5 ( per day ) . Cells with the additional driver divide with rate b1 and die with rate d1 , again with b1 > d1 . The net growth rate of cells with the additional driver , r1 = b1 − d1 , is greater than the net growth rate of type-0 cells , r = b − d . We will denote the ratio of the two growth rates r1 and r by c = r1/r > 1 . Let X0 be the number of type-0 cells at the appearance of the first successful cell with a driver ( whose progeny survives stochastic fluctuations ) . The progeny of this cell forms the type-1 population . Total population size is the sum of type-0 and type-1 cells . We are interested in the probability distribution of the fraction fsub of type-1 cells in the population when total population size is M . Typical size of M that we will consider is 108 − 109 cells . If we let X be the size of the type-0 population and Y the size of the type-1 population when total population size is M , then fsub = Y/M and M = X+Y . Survival probability of a cell with the additional driver mutation is r1/b1 = cr/b1 . Thus the "suc- cessful" driver mutation rate ( the rate at which driver cells with surviving progeny are produced ) is us = ( cr/b1 ) u . On the other hand , we have shown before that the arrival of mutations which appear with rate us in type-0 cells , can be viewed as a Poisson process with rate us/r on the size of the type-0 population [37] . More precisely , we use the fact that the number of mutations produced by the type-0 population by the time it reaches population size M is distributed as Poisson ( M us /r ) . This result was derived using heuristic arguments by Iwasa et al . [38] and Bozic and Nowak [37] , and proved recently by Cheek and Antal [18] . The size of the type-0 population when the first type-1 cell appears , X0 , is therefore exponentially distributed with rate ( c/b1 ) u . Thus X0 will be of the order of b1/ ( cu ) , which is typically much larger than 1 , but much smaller than M . We note that , in line with multiple studies of similar branching processes [39 , 37] , we use a continuous approximation to the size of the type-0 population at the time of mutant appearance . As typical driver mutation rate is u~10−5 , driver mutation is expected to appear when type-0 population contains ~105 cells , so this approximation is justified . We will measure time from the appearance of the first type-1 cell . Let t be the time when total population size is M . Since X0 is typically very large , the population of type-0 cells at time t can be well approximated by X ≈ ert X0 . On the other hand , since type-1 cells started a surviving population at time 0 with a single cell , for the population of type-1 cells we have [39] Y → V1ecrt for large t , where V1 is an exponentially distributed random variable with rate cr/b1 . In other words , Y ≈ V1 ( X / X0 ) c . It follows that M=X+Y≈X+V1 ( X/X0 ) c M1-c≈ ( X/M ) M1-c+V1 ( X/M ) cX0-c 1≈ ( 1-fsub ) +V1 ( 1-fsub ) cX0-cMc-1 fsub ( 1-fsub ) c≈V1X0-cMc-1 ( 3 ) Here we used the fact that X / M = 1 − fsub . On the other hand , P[fsub≤α]=P[fsub ( 1-fsub ) c≤α ( 1-α ) c] , ( 4 ) since x/ ( 1-x ) c is a function that increases as x increases from 0 to 1 . Thus from ( 3 ) and ( 4 ) we have P[ fsub≤α|X0 ]≈P[ V1X0−cMc−1≤α ( 1−α ) c ]=P[ V1≤α ( 1−α ) cX0cM1−c ]=1−exp ( − ( cr/b1 ) α ( 1−α ) cX0cM1−c ) ( 5 ) Finally we have P[fsub≤α]≈∫0Mcub1exp ( -cub1X0 ) [1-exp ( -crb1α ( 1-α ) cX0cM1-c ) ]dX0 ( 6 ) We show the excellent agreement of formula ( 6 ) and exact computer simulations of the process in Fig 4 . Let H ( α ) =cub1exp ( -cub1X0 ) [1-exp ( -crb1α ( 1-α ) cX0cM1-c ) ] To calculate the probability density function , f , for the frequency of subclonal driver , we note that f ( α ) =ddαP[fsub≤α]=ddα∫0MH ( α ) dX0 Using Leibniz’s rule we obtain f ( α ) =∫0MddαH ( α ) dX0 Finally , probability density function for the frequency of a subclonal driver is given by f ( α ) = ( cb1 ) 2ruM1-c ( 1-α ) -1-c ( 1+ ( -1+c ) α ) ∫0MX0cexp ( -cb1 ( uX0+rM1-cX0c ( 1-α ) -cα ) ) dX0 ( 7 ) The estimated driver mutation rate u ∼ 10−5 used in Fig 2 comes from Bozic et al . [21] . In that paper , it was estimated that there are 377 driver genes in the human genome , and an average of 90 positions per driver gene that , if mutated , will result in a functional driver mutation . In addition , it was assumed that the mutation rate per base pair per cell division was 5 · 10−10 , leading to a driver mutation rate on the order of 10−5 per cell division . Since then , new estimates have become available for both the number of driver genes and the point mutation rate in tumors . For example , Vogelstein et al . [9] used mutation patterns to estimate that there are 138 driver genes discovered so far . Similarly , Davoli et al . [10] analyzed patterns of mutational signatures in tumors and estimated 570 driver genes . Lawrence et al . [12] used mutation frequencies and estimated 219 driver genes . They also performed a saturation analysis and showed that many new candidate cancer genes remain to be discovered beyond those they report . Recently , Martincorena et al . [11] used dN/dS ratio to determine genes under positive selection in cancer and estimated 203 driver genes . Based on the sum of these data , we set the upper bound on the number of driver genes to be 600 . On the other hand , if we only focus on strong drivers in a single cancer type , such as colorectal , the number of genes is significantly smaller . For example , Martincorena et al . [11] report 28 genes under significant positive selection in colorectal cancer . Thus we will set the lower bound on the number of significant driver genes of a single cancer type to 20 . Blokzijl et al . [40] estimate that ∼ 40 mutations accumulate per year in the genome of multiple human tissues , including the small intestine , colon and liver , leading to a mutation rate of 0 . 1/day per genome or ∼ 4 · 10−11 per base pair per day . This will be our lower bound for the point mutation rate . Recently , Werner and Sottoriva [41] used the change in the mean and variance of the mutational burden with age in healthy human tissues to estimate the mutation rate in the colon and small intestine; they obtained ∼ 4 * 10−9 per base pair per cell division . Assuming the value they used for time between stem cell divisions is one week , this leads to a mutation rate of ∼ 6 * 10−10 per base pair per day . Mutation rate in cancer can be increased 10–100 fold compared to normal tissues [42] , so we set the upper bound for point mutation rate to ∼ 100*6·10−10 = 6·10−8 per base pair per day . We obtain an upper bound for the driver mutation rate by multiplying the upper bounds for the number of driver genes and point mutation rate with the average number of driver positions , leading to uU = 600*6·10−8 *90∼10−3 per day . Multiplying our lower bounds for the number of driver genes and point mutation rate with the average number of driver positions leads to the lower bound for the driver mutation rate uL = 20*4·10−11 *90∼10−7 . Using formula ( 6 ) , we numerically evaluate P [0 . 2 < fsub ≤ 0 . 8] = F ( 0 . 8 ) − F ( 0 . 2 ) for the following ranges of parameters: ratio of net growth rates of cells with and without the driver , c , between 1 . 01 and 3 ( i . e . relative growth rate advantage of a driver , g , between 1% and 200% ) ; driver mutation rate , u , between 10−7 and 10−3 per day; final tumor size , M , between 107 and 1011 cells; division rate of cells with the driver , b1 , between 0 . 1 and 1 per day; and net growth rate of tumor cells , r , between 0 . 001 and 0 . 1 per day . These ranges are wide and include all meaningful parameter values . We create a grid by taking 100 equally-spaced values for each parameter within its range defined above ( parameter values for driver mutation rate u , tumor size M and net growth rate of tumor cells r are equally spaced in log space ) . We exhaustively evaluate all points on this 5-dimensional grid ( 1005 = 1010 parameter value combinations ) . We find that P [0 . 2 < fsub ≤ 0 . 8] < 0 . 6 holds everywhere , and that the frequency of a subclonal driver is always biased toward 0 and 1 .
|
Darwinian evolution in cancer is responsible for the emergence of malignant traits in initially benign tumors . As tumor cells divide , they accumulate new mutations and while most of them are “passengers” which do not confer any selective growth advantage , “driver” mutations endow cells with traits that contribute to cancer spread . Identifying driver mutations that are under selection in cancer can point to new targets for cancer therapeutics and open new avenues for personalized cancer treatment . It has recently been argued that the presence or absence of selection in cancer can be deduced from deviation of mutant allele frequencies from 1/f power law in an intermediate frequency range . Using a stochastic mathematical model of cancer evolution we derive a formula for the frequency of a subclonal driver and show that frequencies of cancer drivers are biased towards 0 and 1; thus most mutations will inevitably appear to be either neutral ( frequency ≈ 0 ) or clonal ( frequency ≈ 1 ) despite very different levels of selection . Consequently , the proposed 1/f statistic will significantly overestimate the number of cancers deemed to be evolving neutrally . Our work quantifies the validity of the proposed neutral evolution statistic across the entire range of all relevant parameter values .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"cancer",
"genomics",
"medicine",
"and",
"health",
"sciences",
"cancer",
"genetics",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"basic",
"cancer",
"research",
"oncology",
"mutation",
"probability",
"distribution",
"mathematics",
"computer",
"and",
"information",
"sciences",
"probability",
"theory",
"carcinogenesis",
"cancer",
"evolution",
"point",
"mutation",
"computer",
"modeling",
"cell",
"biology",
"gene",
"identification",
"and",
"analysis",
"genetics",
"mutation",
"detection",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"genomics",
"genomic",
"medicine"
] |
2019
|
On measuring selection in cancer from subclonal mutation frequencies
|
Originally a binary classifier , Lot Quality Assurance Sampling ( LQAS ) has proven to be a useful tool for classification of the prevalence of Schistosoma mansoni into multiple categories ( ≤10% , >10 and <50% , ≥50% ) , and semi-curtailed sampling has been shown to effectively reduce the number of observations needed to reach a decision . To date the statistical underpinnings for Multiple Category-LQAS ( MC-LQAS ) have not received full treatment . We explore the analytical properties of MC-LQAS , and validate its use for the classification of S . mansoni prevalence in multiple settings in East Africa . We outline MC-LQAS design principles and formulae for operating characteristic curves . In addition , we derive the average sample number for MC-LQAS when utilizing semi-curtailed sampling and introduce curtailed sampling in this setting . We also assess the performance of MC-LQAS designs with maximum sample sizes of n = 15 and n = 25 via a weighted kappa-statistic using S . mansoni data collected in 388 schools from four studies in East Africa . Overall performance of MC-LQAS classification was high ( kappa-statistic of 0 . 87 ) . In three of the studies , the kappa-statistic for a design with n = 15 was greater than 0 . 75 . In the fourth study , where these designs performed poorly ( kappa-statistic less than 0 . 50 ) , the majority of observations fell in regions where potential error is known to be high . Employment of semi-curtailed and curtailed sampling further reduced the sample size by as many as 0 . 5 and 3 . 5 observations per school , respectively , without increasing classification error . This work provides the needed analytics to understand the properties of MC-LQAS for assessing the prevalance of S . mansoni and shows that in most settings a sample size of 15 children provides a reliable classification of schools .
Schistosomiasis is a tropical disease caused by infection with Schistosoma parasitic worms . The disease burden of schistosomiasis is greatest in sub-Saharan Africa ( SSA ) which shoulders 85% of the global burden [1] , [2] , with school-age children as well as adolescent girls and women of childbearing age suffering the greatest consequences of infection [3] , [4] . The two main species responsible for schistosomiasis in SSA are Schistosoma haematobium , which causes urinary schistosomiasis , and S . mansoni , responsible for intestinal schistosomiasis . The World Health Organization ( WHO ) recommends a three-way classification ( ≤10% , >10 and <50% , ≥50% ) of the prevalence of schistosome infection to determine appropriate interventions for school-age children [4] , [5] . These classifications are generally made using classical statistical approaches with data collected in parasitological surveys of between 250 and 500 children in five to ten schools per ecological zone ( about 50 children per school ) [6] , [7] . However , this recommendation is based on logistical concerns more so than statistical ones . Sampling 50 children in multiple schools may be financially prohibitive and there is a need for rapid assesment methods for defining the distribution of infection in order to target control [8] . For identifying communities/schools with high prevalences of S . haematobium the WHO recommends the use of questionnaires of self-reported blood in urine or parasitological tests [9] . Concerns about the lack of a reliable questionnaire approach for S . mansoni has prompted researchers to explore alternative ways , including the use of lot quality assurance sampling ( LQAS ) , to reduce the sampling effort required to assess the prevalence and distribution of S . mansoni based on parasitological surveys [10] , [11] . A classification tool , LQAS has been used in a variety of settings to identify program areas as either “acceptable” or “unacceptable” with respect to a preestablished target [12] , [13] . As early as 2001 , Rabarijaona et al utilized LQAS to classify the prevalence of S . mansoni [14] , [15] . More recently , a number of studies have been published which utilize LQAS to provide a three-way classification of disease . In 2003 , Myatt et al used LQAS to provide a ternary classification of the prevalence of active trachoma in Malawi [16] . In order to provide a finer classification , Myatt specified two classic LQAS sampling plans with the goal of classifying areas as low/not low and high/not high . Areas classified as both “not low” and “not high” were classified as moderate . In addition , Myatt et al allowed for early stopping in the sampling procedure contigent upon reaching a maximum allowable number of failures in the sample . Brooker et al went on to apply this three-way LQAS to identify schools with a high prevalence of S . mansoni in Uganda , finding that an LQA sample size fo 15 ( n = 15 ) provided reliable classification of infection prevalence [8] . This study showed that the use of an LQAS-based classification system in high transmission settings could drastically reduce the cost of treatment when compared both to the conventional survey method and to blanket treatment without prior screening . These studies demonstrate that LQAS can be used to provide a more precise classification than the traditional method . The Brooker et al multiple classification scheme was evaluated by simulation from a large database of 202 Ugandan schools . While this represents an important step toward validating 3-way LQAS , the results are entirely contingent upon the data in the database and therefore provide little insight into how the same method will perform when applied to different regions with even minor deviations in the underlying distribution of disease . Moreover , the simulation approach to validation gives little guidance for designing a survey where prior information is either sparse or unavailable . If a three-way classification LQAS system is to be used in other settings , it is important to understand the statistical underpinnings of the methodology and provide guidelines for designing such surveys in other settings or for other diseases . The primary aim of the current work is the development of a unified Multiple Category-LQAS tool ( MC-LQAS ) and the validatation of its use in multiple setting in East Africa . Specifically , we outline the theoretical underpinnings of MC-LQAS system , focusing on classification into one of three categories , and provide guidelines for choosing design parameters . Next , we present the theoretical aspects of sequential sampling as employed by researchers in the field [8] , [16] , [17] , [18] . Sequential LQAS , known in the statistical literature as semi-curtailed sampling , allows for potential reduction in the sample size required to make a decision without impacting classification error . A worked example of the tool to assess the prevalence of S . mansoni in 388 schools in Kenya , Uganda , and Tanzania is given [19] , [20] , [21] . The primary design examined is that used by Brooker et al [8] and we discuss the analytical properties of this approach . Finally , we validate the multiple classification system against the standard approach to classification and show that the use of an LQAS based system can substantially reduce the necessary sample size , while providing valid information for selecting the appropriate intervention strategy .
Traditional LQAS calls for a random sample of n binary observations from a “lot” . If the number of successes in the sample , X , is less than or equal to a predefined decision rule , d , the locale is classified as unacceptable . Otherwise , the locale is classified as acceptable . The word “success” is a statistical convention but typically denotes a failure to meet an established criterion or receive an intervention . In the case of sampling for S . mansoni , the number of successes are cases of S . mansoni infection . If the number of infected cases exceeded a pre-determined level , then the lot is rejected and the school/community is identified as in need of mass treatment . A succinct summary of any LQAS design is the Operating Characteristic ( OC ) curve [22] . The OC curve depicts the probability of an acceptable classification against the true underlying prevalence , p . We assume that p represents the proportion in a given population with infection , such as S . mansoni infection . An example of an OC curve is given in Figure 1A with n = 15 and d = 7 . The choice of the sample size and decision rule are critical , as they determine the expected classification error in the procedure . Generally , n and d are chosen so that the probability of incorrectly classifying a locale as having low prevalence is less than or equal to α and the probability of incorrectly classifying a locale as having high prevalence is less than or equal to β . In many cases , practitioners associate the labels “low” and “high” with values of the prevalence below and above some value , p* , respectively . In practice , classification probabilitites are evaluated at upper and lower thresholds pU and pL , and the value of p* serves little purpose aside from informing the choice of these two parameters . For an appropiately chosen design , the values of the OC curve at p = pU and p = pL will be approximately equal to some predefined values 1−α and β , respectively . For example , the design depicted in Figure 1A is chosen so that at pL = 0 . 40 and pU = 0 . 60 , the probability of an acceptable classification is less than or equal to β = 0 . 20 to the left of pL and greater than or equal to 1−α = 0 . 80 , to the right of pU . Due to the monotonicity of the OC curve , it follows that for any p beyond the upper or lower thresholds , the probability of committing an error is no greater than α or β . An additional property of the OC curve is that it makes explicit the values of p for which LQAS runs a risk that is higher than the maximum of α and β; the value of the OC curve increases from β to 1−α as p increases from pL to pU . The area between pL and pU is commonly referred to as the “grey region” . Thus a locale which truly has prevalence p such that pL<p<pU will be classified as acceptable with probability somewhere in the range ( β , 1−α ) ( assuming β<1−α ) . The MC-LQAS procedure extends basic LQAS by classifying a sample against multiple decision rules . In the following we develop MC-LQAS for three-way classification although the method is generalizable to more than three categories . For three-way classification , we must choose a total of two decision rules , d1 and d2 . If the number of successes , X , out of a total of n observations is less than or equal to d1 , classify the prevalence as low . If X is greater than d2 , classify the prevalence as high . Otherwise , classify the prevalence as medium , the middle category . Analogous to the OC curve , for a specific design we can plot the probability of classification into each of the three categories against p to succinctly summarize the MC-LQAS design . Figure 1B shows the OC curve for a three-way classification procedure with n = 15 , d1 = 1 , d2 = 7 . Note that this is a simple extension of the two-way design discussed previously where we now allow for the “unacceptable” category to be parsed into “moderate” and “low” , thus making explicit the connection between this development and that of Myatt et al and Brooker et al [8] , [16] . Of note is the bell-shape of the curve for classification into the moderate category . The lack of monotonicity for this curve is one characteristic of MC-LQAS which sets it apart from LQAS and plays an important role with respect to choosing a design . As with LQAS , in practice we choose to control for potential misclassification at predetermined thresholds , which we call pL1 , pU1 , pL2 , and pU2 . These should be chosen so that pL1<p1*<pU1 and pL2<p2*<pU2 , and in practice it oftentimes makes sense to set p1* = pL1 and p2* = pU2 . To control for the amount of misclassification , we choose d1 and d2 so that the probability of correct classification remains high at these thresholds . That is , choose the decision rules so thatwhere δ1 , δ2 , δ3 and δ4 reflect the acceptable levels of potential error determined by the investigator . This is directly analogous to choosing upper and lower thresholds , pL and pU , in classical two-way LQAS with the notable exception of the moderate category , where we see that it is important to control for the possible error at two locations . This has to do with the aforementioned bell shape of the moderate OC curve . The lack of monotonicity makes it so that one must control for error at both pU1 and pL2 . We note that in the above formulation , we have ignored possible misclassification into the extreme categories . Depending on the distance between thresholds , misclassification into a non-contiguous category can be minimal for even small samples . Hence , for moderate sample sizes , we only worry about four possible misclassifications , which are those misclassifications into contiguous classes . In certain situations , it is possible to reduce the sample size needed to reach a decision by “sampling to the decision rule” . For example , suppose we define a traditional LQAS plan with a sample size n = 15 and d = 7 . Suppose further that during data collection we find that the first eight observations are successes . At this point , we need not sample further to know the resulting classification will be acceptable . The analytical properties of this type of sampling are neither well-documented nor well-understood in the public health literature . However , this process is referred to as semi-curtailed sampling in the statistics literature where it has been in use for the past fifty years [23] , [24] , [25] . The main benefit of this type of sampling is the potential to reduce the overall number of observations , or the Average Sample Number ( ASN ) , required to reach a decision . The semi-curtailed ASN is plotted as a function of the prevalence with n = 15 and d = 7 in Figure 1C and its derivation provided in the Appendix S1 . A feature of semi-curtailed sampling is that it preserves the OC curve , which means that the expected error rates are not affected [25] . Thus , there seems to be little drawback to employing semi-curtailed sampling when feasible to reduce the sample size . Indeed , one can benefit even more by adopting a curtailed sampling plan [23] . That is , one can terminate sampling either if the number of successes is too great or too few at a given point . To continue with our example , suppose instead that the first eight observations are failures . In this case , it is not possible to observe more than seven successes in the remaining observations , and sampling can also cease . The curtailed ASN plotted as a function of the prevalence with n = 15 and d = 7 is plotted in Figure 1C and its derivation is included in the Appendix S1 . Once again , the employment of curtailed sampling does not affect the OC curve . The notion of curtailed sampling is easily extended to MC-LQAS . For example , MC-LQAS also allows for the potential of early stopping by sampling to the decision rule , or semi-curtailed sampling . For example , when utilizing an MC-LQAS design with n = 15 , d1 = 1 and d2 = 7 , sampling can terminate with a high classification as soon as the number of successes excedes seven . The ASN for MC-LQAS when employing semi-curtailed sampling is equal to the ASN in traditional LQAS . The curtailed version of MC-LQAS is slightly different than its traditional counterpart in that it allows for early stopping for low , moderate , or high classifications . Continuing with our example , if the first thirteen observations are failures , then it follows that the lot will be classified as low irrespective of the remaining observations . Likewise , if in the first twelve observations are four successes and eight failures , then sampling can stop with a moderate classification , as neither low nor high classifications are possible at this point . The semi-curtailed and curtailed ASNs for an MC-LQAS design with n = 15 , d1 = 1 , and d2 = 7 are plotted as a function of the prevalence in Figure 1D . We note that the functional form of the ASN under curtailed sampling will generally be bi-model , which reflects the two areas of uncertainy or grey regions . It follows for the same reasons as in the traditional LQAS setting that the OC curves for MC-LQAS are not affected by sequential sampling of this nature . Proofs of these results are given in the Appendix S1 . In the following , we consider S . mansoni data reported in four different studies; two in Kenya [19] , [21] , one in Uganda [26] , and one in Tanzania [20] . In each study , a sample of school children in multiple schools were randomly selected to provide stool samples which were examined microscopically for the ova of S . mansoni , hookworm , Ascaris lumbricoides , and Trichuris trichiura . The number of schools sampled ranges from 21 [19] study to 199 [26] with school sample sizes as low as 21 and as high as 202 . In Figure 2 , the estimated prevalence of S . mansoni in each of the 388 schools , along with 95% exact binomial confidence intervals , are plotted for each of the four studies . We use these data to assess the performance of the MC-LQAS design with n = 15 , d1 = 1 , and d2 = 7 and compare with expected performance . This design differs slightly from that which was utilized in the 2005 Brooker study , where d1 = 2 [8] . Our current choice reflects a 2006 change in WHO guidelines which shifted the lower programmatic threshold from 20% to 10% [5] . Note that decision rules d1 = 1 and d2 = 7 corresponds to prevalence decision thresholds of 6 . 7% and 46 . 7% , respectively . To choose upper and lower thresholds , we assume that the desired probability of correct classification should be greater than or equal to 0 . 80 uniformly ( i . e . δ1 = δ2 = δ3 = δ4 = 0 . 20 ) . Under this assumption , we can solve for the upper and lower thresholds , yielding pL1 = 0 . 055 , pU1 = 0 . 188 , pL2 = 0 . 392 , and pU2 = 0 . 606 . Additionally , to assess the impact of increasing the sample size on classification agreement , we consider an MC-LQAS design with n = 25 , d1 = 2 and d2 = 12 . Using the same approach , we identified upper and lower thresholds of pL1 = 0 . 062 , pU1 = 0 . 164 , pL2 = 0 . 417 , and pU2 = 0 . 583 for this design . We generate 1000 MC-LQAS classifications of each school in the sample by repeatedly “sampling down” the individual data to 15 or 25 students and classifying each school based on these observations . To compare the classifications resulting from MC-LQAS with those that result from binning the full sample prevalence , we calculate for each simulation the weighted kappa statistic , which measures agreement between classification methods across locations [27] . We report the mean kappa statistic and interquartile range ( IQR ) across the 1000 simulations . In addition , we calculate the ASN in each simulation when employing both semi-curtailed and curtailed sampling plans and report the mean ASN and IQR across the 1000 simulations . Lastly , we calculate the proportion correctly classified as a function of the full sample prevalence . All simulations were conducted using R statistical software , version 2 . 11 . 1 [28] .
Figure 3A displays the average proportion of schools correctly classified as a function of the full sample prevalence . For expository purposes , we overlay the OC curve for this design , noting that the simulation results and expected curves coincide . Likewise , we display the average ASN under semi-curtailed ( Figure 3B ) and curtailed ( Figure 3C ) sampling as a function of the full sample prevalence and overlay the expected ASN curves . Once again , these quantities coincide , as expected . The results of our simulation study are presented in Table 1 . The overall agreement between the MC-LQAS with a sample size of 15 and full sample classifications was 0 . 87 ( IQR: 0 . 86–0 . 89 ) . Although not everyone agrees on the interpretation of the kappa statistic , values greater than 0 . 60 are commonly interpretted as implying “substantial” agreement , whereas values greater than 0 . 80 are thought to imply “almost perfect” agreement . For three of the four studies , the agreement between the MC-LQAS and the full sample classifications was high ( κ>0 . 75 ) [27] . On average the use of the MC-LQAS procedure resulted in either substantial or almost perfect agreement with these data . The notable exception was the Clarke et al study from Kenya , where the kappa statistic was 0 . 46 ( IQR: 0 . 34–0 . 58 ) when n = 15 . Of the four studies , the Clarke et al study had the fewest observations and fewest schools . Furthermore , of the 21 schools sampled in this study , 13 schools had full sample prevalence lying within one of the two grey regions where potential error is known to be high ( in comparison with 13 of 25 , 20 of 143 , and 49 of 199 in the Brooker , Clements , and Kabatereine studies , respectively ) . Thus , this MC-LQAS design is expected to be sub-optimal for this type of underlying distribution of prevalences . One might improve performance by increasing the sample size . The kappa statistics for all studies slightly increased when using a sample of size 25 ( Table 1 ) , although agreement in the Clarke study remained low with a kappa statistic of 0 . 52 ( IQR: 0 . 39–0 . 62 ) . The ASN for a maximum sample size of n = 15 when utilizing curtailed and semi-curtailed sampling was 12 . 90 ( IQR: 12 . 87–12 . 94 ) and 14 . 50 ( IQR: 14 . 49–14 . 52 ) , respectively . When the maximum sample size was increased to n = 25 , the ASN for curtailed and semi-curtailed sampling increased to 21 . 54 ( IQR: 21 . 50–21 . 59 ) and 24 . 2 ( 24 . 14–24 . 18 ) , respectively .
This work outlines a unified and systematic approach to designing Multiple Category-LQAS classification systems with application to the prevalence of S . mansoni in schoolchildren . Through simulation and using real data , we show it performs as well as existing methods in practice for classification of the prevalence of infection at a fraction of the sampling effort . Furthermore , for the first time in the public health literature , we have elucidated the theoretical properties of “sampling to the decision rule” , or semi-curtailed sampling , in LQAS , and extended these notions to multiple classification . Our validation study shows that an MC-LQAS design with n = 15 , d1 = 1 , and d2 = 7 provides classifications in near perfect agreement with the standard “binning” approach , yet using less than half as many observations . As expected , agreement between MC-LQAS and full sample classifications tends to be the worst for prevalences lying within the grey region ( as found in the Clarke study ) , where the risks of classification error are high . Our findings resonate with empirical results pointing to the reliability and potential cost-reduction associated with using LQAS for rapid assessment of S . mansoni [8] . Recent research suggests that an LQAS-based approach may also perform better than sophisiticated geostatistical modeling strategies with respect to correct classification , although at a higher cost per high prevalence school correctly classified [11] . Thus , while we have shown that MC-LQAS is a reliable tool for classification , investigators should continue to take care to choose the evaluative approach which best suits a given situation . A limitation of this study is the lack of consideration for diagnostic sensitivity and specificity . The standard method for diagnosis of S . mansoni is the Kato-Katz method , which has been criticized for having low sensitivity that varies depending on the intensity of infection in an individual [29] , [30] . Some studies have found sensitivities as low as 0 . 60 , which is a serious violation of the perfect diagnostic test assumption . Methods for estimating the prevalence of S . mansoni in the presense of variable sensitivity and infection intensity is an area of ongoing research [31] . A shortcoming of our study is that we ignore the underlying distribution of prevalence . In the event that prior information on the level or distribution of p is available , Olives and Pagano provide Bayesian methods for choosing the sample size and decision rule for traditional LQAS [32] . Olives discusses the same approach in the context of multiple classification in [33] , providing the basis for incorporating complex disease dynamics into the model . Although ignoring prior information does not impact the viability of our results , it is expected that incorporating this extra information would improve expected performance . A strength of our study is the principled treatment of curtailed and semi-curtailed sampling in LQAS . The ASN is a largely ignored piece of information that program managers can utilize to inform their choice of LQAS design . Note that curtailed sampling plans allow for early stopping with a classification of moderate prevalence , in addition to low and high . This is in contrast to other sequential LQAS designs used for multiple category classification in the literature , such as those used to classify transmitted HIV drug resistance [34] . In the context of the classification of S . mansoni prevalence , the use of curtailed designs will ultimately require fewer stool samples to be analyzed via microscopy . The reduction in sample size will be most pronounced in high prevalence schools , where as few as eight slides may need to be read before reaching a decision . Unfortunately in many cases , slides will be prepared for all participants and sent to the laboratory for microscopic inspection . Thus these savings are likely to be less pronounced in the field than in the laboratory . For other diseases , such as malaria and urinary schistosomiasis , where rapid diagnositc tests and dipsticks ( for haematuria ) are the modes of diagnosis , the use of curtailed sampling may be of more importance in the field . Further work is required to evaluate the use of MC-LQAS for sampling for several infections; for example the collection of stool samples to diagnose S . mansoni infection and urine samples to diagnose S . haematobium , using either dipsticks for the detection of haematruia or the urine filatration technique . How such an integrated approach compares to the use of questionnaire surveys for S . haematobium also needs to be investigated . LQAS as a tool has come to be associated with simplicity and versatility . MC-LQAS maintains these attributes so as to be useful to a wider audience of practitioners . Here we consider the case of S . mansoni , and show that as a tool for classification of the prevalence of infection , MC-LQAS is both reliable and adapatable . However , just as LQAS has had extensive use in multiple areas in health , we anticipate that this work will have implications reaching well beyond schistosomiasis for other infectious diseases , such as malaria . The design we describe allows for easy adaptation to other circumstances .
|
The control of schistosomiasis calls for rapid and reliable classification tools . This study evaluates the performance of one such tool , Lot Quality Assurance Sampling ( LQAS ) for assessing the prevalence of S . mansoni in African schoolchildren . We outline the design considerations and introduce novel sequential sampling plans for Multiple Category-LQAS . We use data from 388 schools in Uganda , Kenya , and Tanzania to assess the performance of LQAS as a tool for classification of S . mansoni infection into one of three classes: ≤10% >10 and <50% , ≥50% . Our findings suggest that an LQAS-based multiple classification system performs as well as the World Health Organization recommended methods at a fraction of the sampling effort . Our work validates LQAS as a rapid assessment tool and extends it to allow investigators to apply the method to control other infectious diseases .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"public",
"health",
"medicine",
"infectious",
"diseases",
"schistosomiasis",
"public",
"health",
"and",
"epidemiology",
"mathematics",
"statistics",
"biostatistics",
"child",
"health",
"parasitic",
"diseases"
] |
2012
|
Multiple Category-Lot Quality Assurance Sampling: A New Classification System with Application to Schistosomiasis Control
|
Cryptococcus neoformans ( Cn ) , the major causative agent of human fungal meningoencephalitis , replicates within phagolysosomes of infected host cells . Despite more than a half-century of investigation into host-Cn interactions , host factors that mediate infection by this fungal pathogen remain obscure . Here , we describe the development of a system that employs Drosophila S2 cells and RNA interference ( RNAi ) to define and characterize Cn host factors . The system recapitulated salient aspects of fungal interactions with mammalian cells , including phagocytosis , intracellular trafficking , replication , cell-to-cell spread and escape of the pathogen from host cells . Fifty-seven evolutionarily conserved host factors were identified using this system , including 29 factors that had not been previously implicated in mediating fungal pathogenesis . Subsequent analysis indicated that Cn exploits host actin cytoskeletal elements , cell surface signaling molecules , and vesicle-mediated transport proteins to establish a replicative niche . Several host molecules known to be associated with autophagy ( Atg ) , including Atg2 , Atg5 , Atg9 and Pi3K59F ( a class III PI3-kinase ) were also uncovered in our screen . Small interfering RNA ( siRNA ) mediated depletion of these autophagy proteins in murine RAW264 . 7 macrophages demonstrated their requirement during Cn infection , thereby validating findings obtained using the Drosophila S2 cell system . Immunofluorescence confocal microscopy analyses demonstrated that Atg5 , LC3 , Atg9a were recruited to the vicinity of Cn containing vacuoles ( CnCvs ) in the early stages of Cn infection . Pharmacological inhibition of autophagy and/or PI3-kinase activity further demonstrated a requirement for autophagy associated host proteins in supporting infection of mammalian cells by Cn . Finally , systematic trafficking studies indicated that CnCVs associated with Atg proteins , including Atg5 , Atg9a and LC3 , during trafficking to a terminal intracellular compartment that was decorated with the lysosomal markers LAMP-1 and cathepsin D . Our findings validate the utility of the Drosophila S2 cell system as a functional genomic platform for identifying and characterizing host factors that mediate fungal intracellular replication . Our results also support a model in which host Atg proteins mediate Cn intracellular trafficking and replication .
Over the past half-century , Cryptococcus neoformans ( Cn ) , an opportunistic encapsulated yeast , has emerged as the major causative agent of fungal meningoencephalitis in humans and animals worldwide [1] , [2] . The pathogen is found in a variety of environmental niches , including soil , vegetation , and avian excreta . Initial infection in the lungs results from the inhalation of fungal cells from the environment [1] , [2] . The most common clinical form of systemic Cryptococcus infection in humans is meningoencephalitis . An elucidation of the molecular mechanisms mediating the interaction of Cn with host cells will further the development of measures to control cryptococcosis [3] . Several Cn features , such as polysaccharide capsule synthesis [4] , [5] , melanin production [6] , [7] , and growth at host physiological temperatures [7] are associated with virulence . In addition , several signaling pathways , including the cAMP-PKA pathway , three MAP kinase pathways ( involving Cpk1 , Hog1 and Mpk1 ) , the Ras specific pathway , and the Ca2+-calcineurin pathway modulate Cn morphological differentiation , virulence , and stress responses [8] . Numerous Cn virulence factors have also been characterized [9] . Insights into host cell functions that mediate infection have also been garnered . For example , host toll-like receptors ( TLRs ) help recognize fungal pathogens such as Candida albicans , Aspergillus fumigatus and Cn [10] . Recognition of pathogen-associated molecular patterns ( PAMPs ) by TLRs , either alone or with other TLR or non-TLR receptors , induces signals responsible for activation of the host innate immune response [10] . Recent studies have demonstrated that MyD88 , but neither TLR2 nor TLR4 , plays a central role in host defense against Cn [11] , [12] . A variety of host cells , such as T helper type 1 ( Th1 ) [13] , B cells [14] , [15] , gamma-delta antigen receptor-bearing T ( γ δ T ) cells [15] , dendritic cells [16] , alveolar macrophages [17] and bronchial epithelial cells [18] are known to be involved in coordinating the host immune response to Cn infection . Cell-mediated host immunity to infection is characterized by immune cell activation and secretion of cytokines , including gamma interferon ( IFN-γ ) , IL-12 , IL-18 , IL-23 and tumor necrosis factor alpha ( TNF-α ) [1] . Despite these insights , the molecular mechanisms mediating host cell invasion , intracellular replication and escape of Cn from host cells remain obscure . The fate of internalized Cn cells is cell type-dependent . However , common patterns of intracellular trafficking are observed . After entry into host cells , Cn resides in early endosomal antigen 1 ( EEA1 ) -positive compartments [16] , and also interacts with membranes containing the endosome/lysosome markers CD63 [19] and lysosomal-associated membrane protein 1 ( LAMP-1 ) [16] , [20] , [21] . The pathogen is killed in the lysosomes of dendritic cells [16] . However , in macrophages , Cn cells can survive , replicate , and extrude from phagosome/phagolysosome compartments without inducing host cell death [21] , [22] . This extrusion of replicative Cn cells may constitute a mechanism for disseminating Cn among host cells or tissues while limiting detection by host surveillance systems [21] , [23] . Despite these observations , the subcellular pathways that Cn uses to enter , replicate and escape from host cells remain obscure . Recent studies have indicated that activation of the host cell autophagy pathway can control the replication of intracellular pathogens [24] , [25] , [26] . Autophagy is a highly conserved process in which cellular materials , including proteins and organelles , are engulfed into an autophagosome , a specialized double membrane bounded compartment that delivers material to lysosomes for degradation [27] . Nutrient starvation , endoplasmic reticulum stress , and pathogen attack and/or subversion can activate the autophagy machinery . Autophagosome biogenesis requires the activity of autophagy-specific proteins , including Atg5 , an ubiquitin-like protein . Atg5 regulates the conversion of microtubule-associated protein 1 light chain 3 ( LC3-I ) to a lipidated form LC3-II , which localizes to the autophagic membrane [28] , [29] . Membrane-bound LC3-II interacts with the adaptor molecules p62 and NBR1 , which capture cytoplasmic cargo earmarked for autophagic degradation [26] . Therefore , the amount of LC3-II within a cell reflects the amount of newly formed autophagic membrane and/or the number of autophagosomes . Previous studies have demonstrated that host autophagy may be involved in inhibiting the growth of intracellular pathogens , including intracellular bacteria [30] , [31] , [32] and parasites [32] , [33] . However , some microorganisms , including bacteria such as Porphyromonas gingivalis , Brucella abortus , and Coxiella burnetii [34] , and viruses [35] , [36] , [37] may subvert host autophagic processes to promote their intracellular survival and replication . We exploited Drosophila melanogaster S2 cells and RNAi technology to identify host factors that mediate the phagocytosis , intracellular replication and escape of Cn from host cells . Several features of the Drosophila cell system made it an attractive choice for these studies . First , the macrophage-like Drosophila S2 cell system has proven useful for elucidating evolutionarily conserved components of innate immunity [38] , [39] . Second , RNAi strategies can be used to efficiently deplete proteins in these cells . Third , S2 cells have been successfully employed to identify host factors for several microbial pathogens , including Listeria monocytogenes [38] , [40] , Brucella melitensis [41] and the fungal pathogen C . albicans [39] . Importantly , findings in insect cell models have been validated in their mammalian cell counterparts [41] , [42] , [43] , [44] . Thus , the insect cell system has proven useful for resolving physiologically relevant host factors in evolutionarily divergent organisms . Finally , some Cryptococcus species have been shown to reside in a variety of environmental niches , including in association with insects [45] , [46] . Moreover , both the caterpillar Galleria mellonella [47] and D . melanogaster [48] have been previously used as model hosts to study fungal virulence and host defense . Therefore , Drosophila cells constitute a host system from which biologically relevant information can be extracted . Here , we demonstrate that Drosophila S2 cells share striking similarities with their mammalian host cell counterparts in supporting the uptake , intracellular replication , cell-to-cell dissemination and escape of Cn cells . We also demonstrate that Cn interacts with host cell membranes containing the autophagosome marker LC3 , the endosome/lysosome marker LAMP-1 , and the lysosome marker cathepsin D in Drosophila S2 cells , murine J774 . A1 and RAW264 . 7 macrophages and mouse embryonic fibroblasts ( MEFs ) . Finally , we demonstrate that several autophagy-associated proteins mediate Cn infection in these systems , thereby implicating the autophagy pathway in supporting the intracellular lifestyle of the pathogen . These data provide the first functional genomic analysis of Cn host factors and provide new insights into mechanisms mediating this host-pathogen interaction .
Time-lapse microscopy of live cells provides a compelling approach for visualizing dynamic interactions between host and pathogen cells . Moreover , this approach has proven useful for demonstrating extrusion of Cn from host cells [21] , [22] . We employed time-lapse microscopy to determine whether Drosophila S2 cells can support infection by Cn . An analysis of time-lapse images revealed several important aspects of this interaction . First , S2 cells efficiently internalized Cn from the surrounding medium ( Figure 1A , yellow arrows; Supplemental Videos 1 and 2 at: http://www . youtube . com/user/deFigueiredoLab ) . Second , internalized Cn cells were observed to replicate within S2 cells ( Figure 1A , wide light-green arrows; Figure 1B , yellow arrows; Supplemental Video 1 ) with a doubling time of ∼2 to 3 hrs . Third , Cn cells were observed to extrude from host cells , either in groups or individually ( Figure 1A , red arrows and Supplemental Video 1 and 2 ) . Cn extrusion across the plasma membrane was rapid ( <2 min ) and was not accompanied by host cell lysis ( Figure 1A , Supplemental Videos 1 and 2 ) . Finally , Cn cell-to-cell dissemination by means of re-uptake of escaped Cn cells into new host cells ( Figure 1A , yellow arrows and Supplemental Videos 1 and 2 ) , direct cell-to-cell movement ( Figure 1B , red arrows ) and division of infected host cells ( Figure 1A , white arrows and Supplemental Video 2 in yellow frame ) was also observed . Although the viability of host cells was not compromised following extrusion , after long-term infection in cell culture ( >24 hrs ) , host cells lysed and intracellular Cn were released ( Figure 1A , t = 27∶36∶03 ) . Taken together , our time-lapse microscopy analyses indicated that key aspects of host-Cn cell interactions ( i . e . , pathogen uptake , replication , cell-to-cell dissemination , extrusion , and host cell lysis ) were shared by the Drosophila S2 host cell model and mammalian macrophages . To determine whether S2 cells constitute an evolutionarily conserved system in which to interrogate Cn interactions with host cells , we employed fluorescence microscopy to visualize eGFP or dsRed-expressing Cn cells ( AI132-GFP and AI100-dsRed , respectively ) during a time course of infection of S2 or J774 . A1 cells . First , we demonstrated that GFP or dsRed expression in Cn cells did not affect pathogen interactions with host cells ( Figure S1 A ) . Next , we employed fluorescence microscopy to analyze the replication of these Cn cells within host cells . For these experiments , we used Alexa 488-conjugated phalloidin or rhodamine-phalloidin to resolve the host cell actin cytoskeleton and thereby demarcate the region occupied by host cells . We then quantified the number of intracellular Cn cells ( Figure S1B , upper panel ) , and found that their number increased over a time course of infection in both S2 and J774 . A1 cells ( Figure 2 , A1 and B1 ) . Interestingly , we also observed Cn cells that appeared to be escaping from S2 cells ( Figure S1B , lower panel ) . These data supported results obtained from our live cell time-lapse microscopy studies ( Figure 1 and Supplemental Videos 1 and 2 ) and encouraged us to develop a system that can differentiate between Cn uptake , replication , and escape in the context of a functional genomic host factor screening experiment . Antibiotic protection assays have proven useful for identifying and characterizing interactions between intracellular bacterial pathogens and host cells . Despite the power of these assays , similar systems have not been developed for studying interactions between intracellular fungal pathogens and host cells . We therefore developed such an assay to monitor the intracellular replication and escape of Cn from host cells . Fluconazole is a widely used fungistatic agent that inhibits the growth and proliferation of Cn [49] . To test whether fluconazole differentially affects the intracellular or extracellular populations of Cn cells , we infected Drosophila S2 and assorted mammalian cell lines with Cn and then treated the infected cells with various concentrations of fluconazole . We then monitored the growth of the intracellular population by analyzing the number of colony forming units ( CFUs ) within host cells after lysis and plating onto solid medium . The extracellular populations were also measured by plating media collected from the culture . Finally , the viability of the host cells was assessed in trypan blue exclusion experiments [41] . Fluconazole ( 0 to 40 µg/ml ) had no effect on the proliferation/viability of host cells in culture as demonstrated previously [50] and shown in Figure S2A . However , the replication of Cn cells in fluconazole-treated ( 20 µg/ml ) fresh or conditioned medium ( CM ) was inhibited ( Figure S2B ) . After 15 hrs , the number of Cn cells decreased in fresh or conditioned mammalian tissue culture medium ( Dulbecco's Modified Eagle's Medium: DMEM ) at 37°C under 5% CO2 ( Figure S2B , red line ) , indicating that prolonged treatment of Cn with fluconazole in DMEM was toxic to Cn cells . However , the number of Cn cells in Drosophila fresh or conditioned medium remained at a similar level during 24 hrs of incubation at 28°C in ambient air ( Figure S2B , green line ) . When fluconazole was added to Cn infected S2 or J774 . A1 cells , levels of intracellular replication were similar during the first 15 hrs of infection ( Figure 2 , A2 and B2 , black lines ) . We were encouraged by these findings and exploited the fluconazole protection approach to evaluate the escape of Cn cells from infected host cells ( Figure S2C ) . When culture media was collected over a 15 hr time period from Cn-infected S2 or J774 . A1 cells that had been treated with fluconazole , the number of Cn cells in the media increased dramatically ( Figure 2 , A2 and B2 , green line ) . The fluconazole containing culture medium did not support the replication of Cn ( Figure S2B ) , which was consistent with the fungistatic nature of this compound . Moreover , fungal cells retained viability in this medium [49] . The dramatic increase of Cn cells in fluconazole-containing culture media could therefore be regarded as derived from a population of Cn cells that had replicated in and escaped from host cells . Therefore , our CFU analysis provided a means of quantifying the number of Cn cells that escaped from host cells over time . The number of Cn cells recovered from the extracellular media of infected J774 . A1 cell cultures decreased after 15 hrs ( Figure 2 , B2 , green line ) , consistent with our observation that prolonged incubation of Cn in fluconazole-containing DMEM impaired Cn viability ( Figure S2B , red line ) . Hence , experiments that exploited fluconazole to measure the amount Cn escape from mammalian cells were not conducted for more than 15 hrs . In contrast , a loss of Cn viability was not observed in the population of escaped cells that were recovered from the supernatants of S2 cell cultures ( Figure 2 , A2 , green line ) , again consistent with our observation that Cn viability remained at a similar level during 24 hrs of incubation in S2 cell culture supernatants that contained fluconazole ( Figure S2B , green line ) . Taken together , our data indicated that S2 and J774 . A1 cells displayed striking similarities in supporting infection by Cn ( Figure 1 and Figure 2 ) and that fluconazole protection assays could be used to assess the intracellular replication and escape of Cn from both host cell types . We hypothesized that the intracellular trafficking of Cn in S2 and J774 . A1 cells share similarities . To test this hypothesis , we used immunofluorescence microscopy assays to compare the intracellular trafficking patterns of AI100-dsRed in these cell types . Antibodies that recognize evolutionarily conserved Drosophila and murine proteins that reside in assorted subcellular compartments were used for these experiments ( Figure S3A , and [51] ) . These antibodies were shown to display minimal cross-reactivity with the pathogen in immunofluorescence microscopy experiments ( Figure S3B ) . The movement of Cn from the host cell plasma membrane to its replicative niche involves several sequential steps [16] , [19] , [20] , [21] . Upon entry , Cn cells were localized within a vacuole that contained the early endosome marker EEA1 ( Figure 3 , A3 and B3 ) . The maturing Cn-containing vacuole ( CnCV ) could also be sequentially stained with antibodies directed against the late endosome marker mannose-6-phosphate receptor ( M6PR ) , the late endosome/lysosome marker LAMP-1 , and the lysosome marker cathepsin D ( Figure 3 and data not shown ) . The pathogen was also observed to be tightly associated with membranes that contained the ER marker calreticulin ( Figure 3 , A12 and B12 ) . However , Cn interactions with the Golgi marker Grasp65 were not observed at the indicated time point ( Figure 3 , A15 and B15 ) . Similar findings were obtained in both S2 and mammalian cell lines ( Figure 3 ) . Quantitative analysis of these marker protein-positive CnCVs indicated that the host endocytic pathway was involved in the phagocytosis and intracellular trafficking of the pathogen ( Figure S4 ) . Therefore , these data indicated the route of intracellular trafficking of Cn in S2 cells , and also demonstrated that Cn intracellular trafficking and replication in S2 and mammalian cells share strikingly similarities . To further examine whether Drosophila S2 cells can serve as a model host cell system for studying host-Cn interactions , the infection phenotypes of several serotypes and mutant variants of Cn ( Table S1 ) were compared in J774 . A1 and S2 cells . These mutant strains were selected for study because they were derived from the wild-type ( WT ) H99 , displayed varying degrees of virulence in murine models of cryptococcosis , and possessed defects in distinct genetic pathways ( Table S1 ) . When the tested Cn strains were used to infect mammalian macrophages and S2 cells , similar patterns of pathogen uptake , replication and escape were observed ( Figure 4 and Figure S5 ) . For example , strain XL1601 displayed similar uptake , replication , and escape efficiencies in both cell systems ( Figure 4 ) . The number of XL1601 CFUs that were recovered from inside host cells was similar to that of the AI100-dsRed control ( Figure 4 , A3 and B3 ) . However , the size of the corresponding extracellular population was larger than the AI100-dsRed control population ( Figure 4 , A2 and B2 ) , consistent with the possibility that the enhanced capability of this mutant to escape from host cells contributes to its enhanced virulence in animal models of cryptococcosis [52] . The acapsular mutant cap59 displayed ∼6-fold higher rates of phagocytosis than capsular strains ( Figure 4 , A1 and B1 ) , consistent with previous observations [53] . These results indicated that WT and mutant strains of Cn displayed similar infection profiles in both mammalian and S2 cell systems , thereby supporting the conclusion that the S2 cell system provides a useful model host cell system . If Drosophila S2 cells are to provide a model system for elucidating evolutionarily conserved host molecules , then they must share Cn infection phenotypes with mammalian cells when host cell functions are disrupted . With this idea in mind , we examined whether Cn ( AI100-dsRed ) interacted similarly with S2 and J7774 . A1 cells that had been pre-treated ( 1 hr ) and then continuously incubated with a variety of compounds that disrupt host cell functions ( Table S2 ) . Fluconazole was added to the media immediately after infection so that the intracellular replication and escape of the pathogen from host cells could be independently monitored using fluconazole protection assays . We found that the tested compounds neither inhibited fungal growth in culture media ( Figure S6 , A and B ) nor compromised the viability/growth of host cells under the examined conditions ( data not shown and [41] ) . We also showed that pretreating ( 3 hrs ) Cn with these compounds at the indicated concentrations before host cell infection did not impair the entry , intracellular replication and escape of the pathogen from host cells ( Figure S6C ) . We examined whether compound-treated S2 and J774 . A1 cells supported similar patterns of pathogen uptake and replication . Treatment of either cell type with cytochalasin D ( CytD ) , which disrupts actin cytoskeleton polymerization , dramatically reduced the amount of Cn phagocytosis . However , this compound did not inhibit the escape of replicative Cn cells from host cells ( Figure 5 , A2 and B2 ) . When host cells were treated with LY294002 or 3-Methyladenine ( 3-MA ) , which disrupts class III PI3-kinase activity [54] , [55] , the entry , intracellular replication and escape of Cn from host cells were significantly reduced ( Figure 5 ) . Finally , bafilomycin A1 ( BAF ) , an autophagy inhibitor that disrupts vacuolar H+-ATPase activity and endolysosomal acidification [56] , significantly inhibited intracellular replication of internalized Cn cells . However , this compound had limited effect on the escape of Cn cells from mammalian macrophages ( Figure 5 ) . These data indicated that the infection profiles of Cn were similar in Drosophila S2 and J774 . A1 cells that had been treated with compounds that disrupt host cell functions . Encouraged by our findings that Cn infection of S2 cells recapitulates salient aspects of the corresponding infection of mammalian cells , we examined whether the S2 cell system and RNAi technology could be combined to identify host factors that mediate host-Cn interactions . We used a novel RNAi screening approach ( Figure S7 ) to perform a pilot screen for these host factors . Cn is known to replicate within a poorly characterized phagosome-derived compartment [21] , [22] , [57] , suggesting that host factors involved in membrane remodeling , trafficking , or phagosome biogenesis and maturation may be critical to the intracellular lifestyle of the pathogen . For our screen , we pre-selected 410 dsRNAs that targeted the knockdown of genes that ( 1 ) had been annotated to be associated with these events ( http://flybase . org/ ) and ( 2 ) were available in Release 1 . 0 of the Drosophila RNAi Library ( Open Biosystems , Huntsville , AL , USA ) . After two rounds of screening , sixty-two dsRNAs that significantly altered Cn infection of S2 cells were identified . Among the 62 candidates , 57 high priority hits ( Table S3 ) were identified after three rounds of screening . A hit was defined as a dsRNA treatment in which the relative infection index ( RIF , see the Materials and Methods ) differed by two fold of the standard deviation ( SD ) from the untreated control . We also screened 96 dsRNAs that were randomly picked from one of the 76 source plates in the Drosophila RNAi library . Because the manufacturer randomly arrayed dsRNAs into source plates , this strategy for selecting dsRNAs introduced no bias in gene function into this analysis . Six of the randomly selected dsRNAs impaired host cell growth . However , after two rounds of screening , dsRNA hits were not uncovered in this experiment . Therefore , we estimate the hit frequency in the whole Drosophila genome to be less than 1% , and note that our preselected set of dsRNAs was significantly enriched in dsRNAs that target Cn host cell functions . We analyzed our high priority hits in several ways . First , we classified hits into functional classes , including vesicle-mediated transport proteins , cytoskeletal proteins , and vacuolar proton-transport factors ( Figure 6A and Table S3 ) , according to the gene ontology system of biological and molecular function , cellular component , or protein domains as reported in FlyBase ( www . flybase . org ) . Interestingly , of the 57 high priority hits , twenty-eight hits ( 49 . 1% of the total hits recovered ) had been previously shown to mediate infection of S2 cells by fungal or intracellular bacterial pathogens , including Brucella melitensis [41] , Chlamydia caviae [43] , Chlamydia trachomatis [44] , Listeria monocytogenes [38] , [40] or Mycobacterium fortuitum [40] , [42] ( Figure 6B and Table S3 ) . Among the 28 hits , 11 genes ( 19 . 3% of the hits ) are known to be required for phagocytosis of C . albicans [39] , a fungal pathogen that causes oral and genital infections in humans and is also associated with morbidity and mortality in immunocompromised patients ( Figure 6B and Table S3 ) . Several host factors previously known to function in phagocytosis [39] , such as Rac1 , actins , and PI3-kinase ( class III ) and COPI coatomers , were identified as high priority hits ( Table S3 ) . Consistent with these findings , pharmacological disruption of the functions of these host factors with CytD or LY294002 reduced host cell phagocytosis of Cn cells ( Figure 5 , A1 and B1 ) . Finally , we identified 29 genes ( 50 . 9% of the high priority hits ) that had not been previously reported to be involved in mediating intracellular Cn infection ( Table S3 ) . Therefore , our screening strategy was sufficiently robust to uncover suspected host factors as well as factors with novel functions . Our screen uncovered several autophagy ( Atg ) genes , including Atg2 ( CG1241 ) , Atg5 ( CG1642 ) and Atg9 ( CG3615 ) , which play conserved roles in autophagosome biogenesis [58] . However , the engagement of host cell Atg proteins by Cn had not been previously reported . To determine whether elements of the host cell autophagy pathway regulate Cn interactions with macrophages , we exploited murine RAW264 . 7 macrophages as a host cell system . RAW264 . 7 cells provided an established model for interrogating Cn interactions with mammalian macrophages [9] . RNAi mediated gene knockdown can also be readily achieved in these cells , which was advantageous for our studies . We thus treated RAW264 . 7 cells with siRNAs to knock down the expression of the murine orthologs of several Drosophila Atg genes that were uncovered in our S2 cell screen . The targeted mammalian genes included Atg2a , Atg5 , Atg9a , Atg12 and LC3 . After validating gene knockdown by immunoblotting analysis ( Figure S8 ) , we infected the siRNA treated cells with Cn ( H99 ) , and then quantified the phagocytosis , replication , and escape of the pathogen using our fluconazole protection assay . When host cells that had been depleted of the indicated Atg proteins were infected with Cn , the uptake and/or replication of the pathogen were significantly decreased compared to controls that had been treated with scrambled siRNAs ( Figure 7 ) . These data supported findings obtained from our RNAi studies using the S2 cell system ( Table S3 ) that indicated the engagement of host cell autophagy proteins in supporting Cn infection . Our siRNA experiments in RAW264 . 7 cells suggested that Atg2 , Atg5 and Atg9 support Cn infection of both Drosophila S2 and mammalian cell systems . To further investigate the roles of Atg proteins in supporting pathogen infection , we employed confocal immunofluorescence microscopy to analyze the interaction of assorted Atg proteins with Cn cells during a time course of Cn infection of mammalian host cells . Upon internalization , Cn was observed in close apposition to membranes that were decorated with LC3 , Atg9a and Atg5 ( Figure 8 , A–D ) , suggesting that these autophagy elements are recruited to the vicinity of CnCVs . Interestingly , the intimate interactions between Cn and these Atg proteins during the early stages of Cn infection were observed in several mammalian host cell types , including J774 . A1 , MEF and RAW264 . 7 cells ( Figure 8 , A–D and data not shown ) . Quantitative analysis revealed that the recruitment of Atg proteins to the vicinity of CnCVs was reduced after 3 h . p . i . ( Figure 8 , A and B , white arrows and data not shown ) . These microscopy experiments therefore supported findings from our siRNA experiments . To further investigate the possibility that the activation of proteins in the autophagosome biogenesis pathway supports Cn infection , we examined the conversion of host LC3-I to LC3-II as an indicator of these events . As expected , overnight nutritional limitation of host cells induced the conversion of LC3-I to LC3-II , resulting in an increased ratio of LC3-II/LC3-I ( Figure 8E ) ; however , this conversion was not significantly increased over a 24 hr time course of incubation in the absence of Cn infection ( Figure 8E and data not shown ) . Immunoblotting results indicated that the level of LC3-II , and the ratio of LC3-II/LC3-I increased during Cn infection of WT J774 . A1 , RAW264 . 7 and Atg5+/+ MEF cells ( Figure 8E ) . However , the conversion of LC3-I to LC3-II was not detected in Atg5−/− MEFs at all of the tested time points ( Figure 8E3 ) . This latter finding was consistent with our observations that Cn cells did not reside in LC3-positive phagosomes/vacuoles in the Atg5−/− MEFs ( data not shown ) . Taken together , our data demonstrated that Cn infection of host cells involves interactions with LC3-positive phagosome membranes .
Yeast or basidiospore forms of Cn typically use the host respiratory system to gain entry into the host . Cn then rapidly disseminates to extrapulmonary tissues , including the brain , which accounts for the meningoencephalitis that is typically associated with cryptococcosis . Macrophages constitute important cellular targets of infection . Cn infects and replicates within alveolar macrophages [17] , [59] , and can also exploit macrophages to gain entry into the central nervous system via a Trojan horse mechanism in which these immune cells ferry ingested Cn cells across the blood brain barrier [60] , [61] . Understanding host mechanisms that support the survival and replication of the pathogen within macrophages is therefore important to understanding pathogen dissemination and disease progression . Non-vertebrate models have emerged as tractable experimental systems for illuminating host-pathogen interactions [62] , [63] . The current study indicates that the combination of macrophage-like Drosophila S2 cells and RNAi technology provide an appealing system for elucidating the roles of host factors in mediating the host-Cn interaction . Cn infection of S2 cells recapitulated salient aspects of mammalian macrophage cell infection . First , Cn infection of S2 and mammalian cells shared similar infection phenomena , including phagocytosis , intracellular replication , cell-to-cell dissemination , extrusion and host cell lysis profiles . Moreover , the intracellular trafficking and replication of Cn in murine macrophages and Drosophila S2 cells shared striking similarities , suggesting that the fungal pathogen subverts conserved host factors to secure an intracellular replicative niche and to disseminate . Second , Cn strains from different genetic backgrounds and mutants with altered virulence in mouse models of infection displayed similar patterns of uptake and replication in S2 and J774 . A1 cells . Third , similar patterns of Cn phagocytosis , intracellular replication , and escape were observed in host cells that had been treated with structurally diverse compounds that target specific host cell functions . These data suggested that both host cell systems share common molecular requirements to support these processes . Fourth , results from an RNAi screening experiment demonstrated that host factors previously known to be required for the phagocytosis of various bacterial and fungal pathogens [39] are also required for Cn infection of host cells ( Table S3 and Figure 6B ) . These data support the idea that evolutionarily conserved host functions support pathogen interactions with S2 and mammalian cells . Finally , Cn cells have been isolated from insects such as cockroaches and beetles [45] , [46] , which is consistent with the possibility that aspects of the observed interactions with S2 cells may recapitulate pathogen interactions with insects in the natural environment . It has been postulated that the pathogenic strategies of Cn in mammals evolved from interactions between Cn and its natural environmental predators [64] . Our data support the conclusion that S2 cells provide a useful model system for elucidating host-Cn interactions . Autophagy is a catabolic process in which cytosolic constituents are sequestered from the rest of the cytoplasm by the autophagosome [27] , which delivers the sequestered constituents to lysosomes for degradation . Autophagy is a tightly regulated process that is essential for survival , differentiation , development and homeostasis [25] , [26] . Several forms of autophagy have been identified that differ with respect to their physiological functions and the mode of cargo delivery to the lysosome [25] . Atg5 is essential for most forms of mammalian autophagy [65] , [66] , [67] . This protein mediates the conversion of LC3-I to LC3-II , which drives autophagosome biogenesis , and indicates activation of the autophagy pathway . Atg5 also directs the localization of LC3-II to autophagosomes [28] , [29] . Atg9a , the only transmembrane Atg protein in mammals , is proposed to mediate membrane transport to generate autophagosomes [68] . The Atg18-Atg2 complex is involved in the transport of Atg9 to autophagic membranes [69] , [70] . Finally , an alternative , Atg5/Atg7-independent macroautophagy pathway has been described in which conversion of LC3-I to LC3-II does not occur [71] , [72] . Therefore , autophagy can be engaged without recruiting the full complement of Atg proteins that have been annotated to be associated with this process . Recent studies indicate that autophagy may be an important pathogen control mechanism [24] , [25] , [26] . Atg5 plays roles in inhibiting the growth of pathogens , such as L . monocytogenes [30] , [31] , [32] and T . gondii [32] , [33] . In Drosophila , autophagy helps control L . monocytogenes infection of primary hemocytes [31] . In contrast to the potential protective role for autophagy against bacterial and parasitic infection , other studies suggest that several bacteria and parasites may subvert host autophagic processes to establish a successful infection; these pathogens include Staphylococcus aureus [73] , Legionella pneumophila [74] , B . melitensis [41] , C . burnetii [75] , Helicobacter pylori [76] , [77] , Trypanosoma cruzi [78] and T . gondii [79] . In addition , some pathogens , such as the HSV-1 virus , are known to exploit elegant mechanisms for inhibiting signaling processes that induce autophagy [80] , [81] . T . gondii infection induces host cell autophagy via a mechanism that depends on host Atg5 . In fact , the growth of this parasite is defective in Atg5-deficent cells [79] . Therefore , in addition to their roles in the canonical autophagy pathway , host Atg proteins may also contribute to modulating the level of replication of viral or intracellular pathogens . In this study , we demonstrated that Cn exploits host cell autophagy proteins to secure an intracellular replicative niche . First , we showed that replicative Cn cells reside in CnCVs that are tightly associated with the autophagosome marker LC3 and lysosomal marker LC3 . However , similar LC3 localization was not observed in pathogen-infected Atg5−/− MEFs . Second , we demonstrated that compounds that are known to disrupt the activities of proteins that regulate the host autophagy pathway ( BAF , LY294002 and 3-MA ) inhibit Cn uptake and replication . In an early stage of autophagosome formation , the Beclin 1-hVps34 ( class III PI3-kinase ) complex supplies phosphatidylinositol-3-phosphate ( PI[3]P ) to the pre-autophagosome membrane , a process that is essential for autophagosome membrane formation [82] . 3-MA and LY294002 are inhibitors of class III PI3-kinase and also block autophagy [54] , [55] . Our data demonstrate that host cells treated with 3-MA or LY294002 display reduced levels of Cn infection , and implicate PI3-kinase activity in supporting the intracellular lifestyle of the pathogen . Finally , we showed that RNAi-mediated depletion of host Atg proteins , including Atg2a , Atg5 , Atg9a , reduced Cn infection of host cells . RNAi-mediated depletion of Pi3K59F ( class III PI3-kinase ) also failed to support Cn infection in S2 cells . Taken together , our data demonstrate that Cn exploits host proteins that regulate the host autophagy machinery to facilitate intracellular replication . Interestingly , Cn employs autophagy as a survival mechanism and virulence-associated trait . Cn strains lacking the Vps34 PI3-kinase and ( vps34Delta ) and depletion of Atg8 markedly attenuated virulence in a mouse model of infection [83] . Therefore , autophagy pathways in both the host and pathogen play important roles in supporting the interaction between Cn and host cells . Based on our findings , we propose a model in which key elements of the autophagy pathway , including Atg5-Atg12 , Atg9a and LC3 , coordinate with other host factors identified in our RNAi screen to regulate the sequential intracellular replication and escape of Cn from host cells ( Figure 9 ) . The following step-wise events are envisioned . First , phagocytosis of Cn is mediated by several factors , including host actins and class III PI3-kinase activities . Upon entry into host cells , Cn traffics to EEA1-positive CnCVs , which then traffic along the endocytic pathway , interact with membranes containing late endosome markers and autophagosomal markers ( LC3 ) , and fuse with lysosomes to become late CnCVs that are decorated with markers associated with endosome/lysosome ( LAMP-1 ) and lysosome ( cathepsin D ) proteins . These compartments are permissive for Cn intracellular replication . Finally , Cn cells can escape from CnCvs at any stage of the maturation process ( Figure 9 ) . Intracellular and extracellular populations of Cn each contribute to disease progression in the infected host . For example , enlarged Cn cells that resist phagocytosis have been reported [84] , [85] , [86] , [87] . Strains that harbor mutations in signaling pathways of the pathogen that support giant cell formation display reduced virulence [85] . These observations are consistent with the conclusion that this extracellular population of Cn cells is important for the infection and dissemination of the pathogen . On the other hand , several lines of evidence support the conclusion that macrophage invasion and replication constitutes an important to disease progression in humans and whole animal model systems . For example , animals depleted of macrophages by clodronate treatment display reduced pathogen load in a variety of organs after infection with Cn [88] . Moreover , Cryptococcus gatti strains with enhanced abilities for intracellular replication display enhanced virulence in humans and murine models of infection [89] . The proposed model of intracellular trafficking dynamics put forth in this report will thus illuminate an important aspect of fungal pathogenesis .
The C . neoformans strains used in this study are listed in Table S1 . Yeast forms of Cn cells were grown on YPD ( Difco ) agar plates and maintained on the plates for no more than one week prior to experimentation . For infection , 2 . 4 ml of YPD broth was inoculated with a loop of Cn cells taken from a freshly streaked YPD plate . Cultures were then grown overnight with shaking at 37°C . Drosophila S2 cells were maintained at 25°C in Insectagro DS2 serum free medium ( DS2-SFM , Cellagro ) with or without 10% fetal bovine serum ( FBS ) . Murine J774 . A1 and RAW264 . 7 macrophages and Atg5-deficient ( Atg5−/− ) MEFs and the corresponding control ( Atg5+/+ ) [65] were routinely incubated at 37°C in a 5% CO2 atmosphere in DMEM supplemented with 10% FBS . Cells were seeded in 24 or 48-well plates and cultured overnight before infection . For CFU assays , 1 . 0×106 S2 and 2 . 5×105 mammalian cells were seeded in each well of a 24-well plate; for a 48-well plate , 5 . 0×105 S2 and 1 . 2×105 mammalian cells were seeded; for fluorescence microscopy assays , 6 . 0×105 S2 and 1 . 0×105 mammalian cells were seeded on 12-mm glass coverslips ( Fisherbrand ) that were placed on the bottom of 24-well plates . The seeded S2 and mammalian cells were cultured overnight before infection . rosophila S2 and mammalian cells were infected with Cn cells at a multiplicity of infection ( MOI ) of 5 and 10 , respectively , unless otherwise indicated . Infected cells were then incubated at 25°C ( S2 cells ) or 37°C in a 5% CO2 atmosphere ( mammalian cells ) . One or 3 h . p . i . , culture media was removed , and the infected cells were extensively ( 6 to 8 times ) washed with 1×phosphate buffered saline ( PBS ) ( S2 and J774 . A1cells ) or a mixture of 1×PBS and DMEM with 10% FBS ( V∶V = 1∶1 ) ( RAW264 . 7 cells and MEFs ) . Fresh media , supplemented with fluconazole ( 20 µg/ml ) , was added to each well containing infected host cells and the infected cells were continuously incubated in this antifungal agent for various lengths of time at the indicated conditions . At the indicated time points , culture medium ( containing escaped Cn cells ) and the corresponding Cn infected host cells were collected separately to measure the number of CFUs . Fluconazole inhibition of Cn replication in S2 and mammalian cell media was also measured by inoculating Cn in fresh or conditioned medium containing 20 µg/ml fluconazole and performing CFU assays . Cn intracellular replication or escape efficiencies for experiments ( in which infection was initiated at starting time point t0 and proceeded to time point tn ) were calculated as described in the Mathematical Formulae section ( see below ) . If no CFUs were recovered from the culture medium at t0 , then we could not calculate escape efficiency . In this case , the net increases in the number of CFUs recovered from the media for the experimental treatments were compared to the corresponding net increase of the control at tn . We used fluconazole protection assays to measure the number of intracellular Cn and also the number of Cn that had escaped from these same host cells . To measure the intracellular population of Cn cells , we infected host cells for 3 hrs , washed these cells extensively ( 6–8 times ) , and then added fresh medium containing fluconazole ( 20 µg/ml ) . The cells were then continuously incubated for various lengths of time . At the indicated time points , the media was removed . The cells were then lysed by the addition of 0 . 5% Tween-20 in sterile water ( 100 µl ) and incubation at room temperature ( S2 cells ) or at 37°C ( mammalian cells ) for ∼10 min . Additional 100 µl of YPD was added to the resultant whole cell lysates ( which contained the intracellular Cn population ) . Finally , 5 µl of cell lysates were plated onto solid YPD medium after serial dilution . CFUs were counted after 24 hrs of incubation at ∼30°C . To measure the escaped Cn cells , the removed media were transferred to 96-well plates ( if total volume <300 µl ) or 1 . 5 ml microcentrifuge tubes ( if total volume >300 µl ) and centrifuged at 1200×g for 12 min . The media with fluconazole were then removed and the Cn cells were re-suspended in YPD broth . The Cn suspension was serially diluted in YPD broth or 1×PBS buffer , and 5 µl of the diluted Cn suspension was plated onto solid YPD agar plates to determine CFUs as above . To measure internalized Cn cells , after extensive wash , the infected cells were lysed and performed CFU assay as described above . To measure the number of Cn cells in the culture media at the starting time point ( t0 , 3 h . p . i . ) , after extensive wash ( 6∼8 times ) , fresh media ( 200 µl ) without fluconazole or 1×PBS were added to the wells containing infected host cells . The supernatants ( which contained the extracellular Cn population ) were transferred to 96-well plates immediately after the plates containing infected cells were briefly shaken . After serial dilution , the supernatants were then performed CFU assay as described above . Host cells were coincubated with or without pharmacological compounds at the indicated concentration ( Table S2 ) for various lengths of time . Quantification of the viability of treated S2 and mammalian cells as well as the total cell number was performed as previously described [41] . To visualize intracellular populations of Cn cells , host cells ( e . g . , S2 or J774 . A1 cells ) were seeded on 12-mm coverslips that had been placed on the bottom of wells of 24-well plates . Next , these cells were infected with Cn strain AI100-dsRed or AI132-GFP . At 1 or 3 h . p . i . , the infected host cells were extensively washed and then continuously incubated in fresh medium supplemented with 20 µg/ml fluconazole . At different time points post infection , the coverslips with infected host cells were fixed and stained with Alexa 488-conjugated phalloidin ( for AI100-dsRed infected cells ) or rhodamine phalloidin ( for AI132-GFP infected cells ) to resolve the host cell actin cytoskeleton . The infected cells were then visualized by confocal fluorescence microscopy . Differential interference contrast ( DIC ) and fluorescence images were used to analyze phagocytosis of Cn cells by host cells . To elucidate Cryptococcus intracellular trafficking and interactions with host proteins , Drosophila S2 , J774 . A1 and RAW264 . 7 cells that were seeded onto 12-mm coverslips on the bottom of 24-well plates were infected with AI100-dsRed for various lengths of time . To analyze infections of less than 3 hrs , the infected cells were washed three times with 1×PBS or the mixture of PBS and DMEM with 10% FBS , and then the infected cells were fixed before IFMA . Otherwise , at 3 h . p . i . , the infected cells were washed three times with 1×PBS or a mixture of PBS and DMEM . In experiments in which the cross-reactivity of extracellular Cn cells with the indicated antibodies was assessed , the Cn infected host cells were washed only once . Next , fresh media supplemented with 20 µg/ml fluconazole were added to the wells containing Cn infected host cells . At various time points post infection , the infected cells were fixed and processed for IFMA as previously described [41] with the following modification: before incubation with antibody , the infected cells were blocked with 10% non-fat dry milk in PBSTT buffer ( 0 . 05% Triton X 100+0 . 05% Tween 20 in 1×PBS ) for ∼2 hrs at room temperature , and then the cell samples were incubated with antibody in PBSTT with 5% non-fat dry milk . The primary antibodies used were as follows: goat-anti-human EEA1; rabbit anti-human M6PR; rabbit anti-human LAMP-1; rabbit anti-human cathepsin D; rabbit anti-mouse LC3; rabbit anti-Calreticulin; purified Grasp65 from rabbit; mouse anti-human Atg5; mouse anti-Cn ( Meridian Life Science , Inc . , USA ) . Samples were stained with Alexa Fluor 488-conjugated and/or Alexa Fluor 594-conjugated secondary antibody ( Molecular Probes , 1∶1000 ) . Acquisition of confocal images , image processing and analyses were performed as previously described [41] . J774 . A1 or Drosophila S2 cells were seeded in 6- or 24-well plate with glass bottom . The cells were overnight cultured before Cn ( AI100-dsRed ) infection ( MOI = 3 ) . At 3 h . p . i . , Cn infected host were washed 3 times with 1×PBS . Fresh media with 20 µg/ml fluconazole were added to the infected cells , and the cells were then incubated in the small environmental chamber ( 28°C for S2 cells and 37°C , 5% CO2 for J774 . A1 macrophages ) mounted on the laser confocal microscopy ( Eclipse Ti , Nikon ) . Live cell images were automatically taken every 2 min for 24 ( J774 . A1 ) and 48 ( S2 cells ) hrs with NIS elements AR 3 . 0 software ( Nikon ) , which was initiated after phagocytosing . Images were processed with NIS elements AR 3 . 0 software . In 48 well plates , Drosophila S2 and murine macrophage J774 . A1 cells were coincubated with assorted pharmacological compounds including CytD , LY294002 , 3- 3-MA and BAF at the indicated concentrations ( Table S2 ) . Cells were treated with these drugs 1 hr before and during infection with Cn strain AI100-dsRed . The treated cells were incubated at 25°C ( S2 cells ) or at 37°C with 5% CO2 ( J774 . A1 macrophages ) . At 3 h . p . i . , the cells were extensively washed with 1×PBS . Fresh media supplemented with the same concentration of the drugs and 20 µg/ml fluconazole were then added to the infected cells . Fluconazole was used to inhibit the replication of extracellular Cn cells , if any . To evaluate phagocytosis of Cn cells , the infected cells ( 3 h . p . i . ) were lysed with 0 . 5% Tween 20 after extensive washing with 1×PBS . The cell lysate was used to perform CFU assays as described above . In addition , the media were separately collected for CFU assays which determined the initial number of Cn cells present in the media at the starting time point ( t0 ) . At 12 h . p . i . , the media and extracts derived from infected cells were also analyzed using CFU assays as described above . To evaluate the effects of drugs on Cn pathogenicity , the Cn cells were pre-treated with the indicated drugs for 3 hrs and then the drugs were washed out . These cells were then used to infect normal host cells . Phagocytosis , intracellular replication and escape of the drug pre-treated Cn cells were analyzed as described above . Generation of dsRNAs that target the knockdown in the expression of Drosophila proteins were performed as previously described [41] . The dsRNA products were stored at −80°C until use . As shown in Figure S7 , dsRNAs were added to 48-well plates at a final concentration of 15 µg/ml . dsRNAs were tested in duplicate in independent plates . S2 cells were then seeded in the plates at a density of 5 . 0×105 cells/well in 150 µl DS2-SFM medium . dsRNA-treated cells were incubated at 25°C for 1 day; 150 µl of fresh medium with 10% FBS was then added to each well to allow the cells to recover from dsRNA treatment . The dsRNA-treated cells were continuously incubated at 25°C for an additional three days to allow the knockdown of target gene expression . The efficiency of dsRNA mediated gene knock down was checked as previously described [41] . The treated cells ( in 100 µl ) together with 100 µl of fresh DS2-SFM medium were added into 48 well plates , and the treated cells were incubated overnight at 25°C before Cn infection . Cn strain H99 or AI100-dsRed was used to infect the dsRNA-treated S2 cells at an MOI of 5 . At 3 h . p . i . , the infected cells were washed 6∼8 times with 1×PBS and then 200 µl of fresh DS2-SFM medium supplemented with 30 µg/ml fluconazole was added to inhibit the replication of extracellular Cn cells , if any . To determine the number of Cn cells present in the medium at the starting time point ( 3 h . p . i . , t0 ) , the media were transferred to U-bottom 96-well plates ( MicroTEST ) . Cn cells were centrifuged ( 1 , 200×g , 12 min ) and resuspended in 200 µl of YPD broth . CFU assays were performed on the recovered Cn cells . To evaluate the effects of dsRNA-mediated knockdown of target genes in S2 cells on the phagocytosis of Cn cells , the infected cells were extensively washed and then lysed with 100 µl of 0 . 5% Tween 20 in sterile water for 15 min . 100 µl of YPD broth was added to the cell lysates . The material was thoroughly mixed and the cell lysate was transferred to a 96-well plate . CFU assays were performed as described above . To assess the effects of dsRNA treatment on Cn intracellular replication and escape at 24 h . p . i , media and cell lysates were first transferred to U-bottom 96-well plates . CFU assays were then performed on this material . Five µl of diluted cell lysate and Cn cell suspension were separately plated onto single well OmniTray ( Nunc ) plates containing YPD solid medium . CFUs were read after 24 hr of incubation at 30°C , and the infection index and RIF ( see the Mathematical Formulae section ) were calculated . dsRNAs that displayed a RIF that differed by more than 2-fold of SD from the untreated control in the screen were picked out for next round of screen . Candidates identified after two rounds of screening were selected for re-testing in triplicate . 2 . 0×104 RAW264 . 7 macrophages were seeded in each well of a 48-well plate one day before siRNA transfection . siRNA treatment was performed using Lipofectamine RNAiMAX or Lipofectamine 2000 transfection reagent ( Invitrogen ) in accordance with the manufacturer's instructions . All the siRNA constructs , including the scrambled siRNA and siRNAs targeting Atg2a , Atg5 , Atg9a , Atg12 and LC3 were purchased from Santa Cruz Biotech . , Inc . ( Santa Cruz , CA , USA ) or US Sigma-Aldrich . After two days of transfection , the medium was replaced with fresh DMED containing 10% FBS at least 2 hr before Cn ( strain H99 ) infection . The growth and viability of siRNA-treated cells was monitored using the trypan blue exclusion assay [41] . At the indicated time points , siRNA transfected RAW264 . 7 macrophages or Cn infected host cells were washed three times with cold 1×PBS . The washed cells were then lysed without disrupting the intact Cn cells . ∼15 µg of the recovered protein extracts were separated by 12% SDS-PAGE , transferred onto polyvinylidene difluoride membranes ( PVDM ) , and blotted with the indicated antibodies . The blots were detected using an enhanced chemiluminescence kit ( Pierce , Rockford , IL ) . Blots were stripped and reprobed for the protein level of GAPDH ( Glyceraldehyde 3-phosphate dehydrogenase , a constitutive housekeeping gene ) as a loading control . Primary antibodies , including anti-LC3 , anti-Atg5 , anti-Atg12 and anti- GAPDH were purchased from Santa Cruz Biotech . , Inc . Anti-Atg9a was purchased from Thermo Scientific . Densitometry of blots was performed using the ImageJ ( http://rsbweb . nih . gov/ij/ ) software package . For each blot , the value of Blot area×Mean gray value×Integrated density ( BaMI ) was calculated . The ratio of blot LC3-II/LC3-I at an indicated time points were calculated as shown in the Mathematical Formulae section ( see below ) . All quantitative data were derived from results obtained from triplicate experiments that were independently performed at least three times . To easily compare results from independent experiments , the data from controls were normalized as 1 or 100% . The significance of the data was assessed using the Student's t-test . The NCBI ( http://www . ncbi . nlm . nih . gov/ ) database accession numbers for the mouse genes and gene products discussed in this paper are Atg2a: 329015 , Atg5: 11793 , Atg9a: 245860 , Atg12: 67526 and LC3β: 67443 .
|
Cryptococcus neoformans is a facultative intracellular fungal pathogen that causes cryptococcosis in both immunocompromised and immunocompetent individuals worldwide . Initial infection usually occurs in the lungs , but the fungus can disseminate to other organs . The pathogen shows a predilection to the central nervous system ( CNS ) , which can result in life-threatening cryptococcal meningoencephalitis . Significant progress has been made in developing molecular genetic approaches to elucidate mechanisms of Cn pathogenesis . However , because of the unavailability of genetically tractable host cell systems , host factors that mediate the phagocytosis , intracellular replication and escape of Cn have remained obscure . Our data demonstrate that the combination of Drosophila S2 cells and RNA interference technology provides a powerful platform for identifying and characterizing host factors that mediate Cn infection . After screening over 400 genes that were annotated to be associated with host cell membrane trafficking and phagosome formation , we identified 57 evolutionarily conserved gene products that when depleted significantly altered the infection phenotype of the pathogen . Finally , we demonstrated that Cn infection of host cells requires autophagy proteins ( including Atg2a , Atg5 and Atg9a ) and class III PI3-kinase activities , thereby implicating host cell autophagy in supporting the intracellular lifestyle of the pathogen . Our work contributes to understanding host mechanisms that mediate the intracellular survival and dissemination of Cn .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"infectious",
"diseases"
] |
2011
|
Functional Analysis of Host Factors that Mediate the Intracellular Lifestyle of Cryptococcus neoformans
|
Genomic enhancers regulate spatio-temporal gene expression by recruiting specific combinations of transcription factors ( TFs ) . When TFs are bound to active regulatory regions , they displace canonical nucleosomes , making these regions biochemically detectable as nucleosome-depleted regions or accessible/open chromatin . Here we ask whether open chromatin profiling can be used to identify the entire repertoire of active promoters and enhancers underlying tissue-specific gene expression during normal development and oncogenesis in vivo . To this end , we first compare two different approaches to detect open chromatin in vivo using the Drosophila eye primordium as a model system: FAIRE-seq , based on physical separation of open versus closed chromatin; and ATAC-seq , based on preferential integration of a transposon into open chromatin . We find that both methods reproducibly capture the tissue-specific chromatin activity of regulatory regions , including promoters , enhancers , and insulators . Using both techniques , we screened for regulatory regions that become ectopically active during Ras-dependent oncogenesis , and identified 3778 regions that become ( over- ) activated during tumor development . Next , we applied motif discovery to search for candidate transcription factors that could bind these regions and identified AP-1 and Stat92E as key regulators . We validated the importance of Stat92E in the development of the tumors by introducing a loss of function Stat92E mutant , which was sufficient to rescue the tumor phenotype . Additionally we tested if the predicted Stat92E responsive regulatory regions are genuine , using ectopic induction of JAK/STAT signaling in developing eye discs , and observed that similar chromatin changes indeed occurred . Finally , we determine that these are functionally significant regulatory changes , as nearby target genes are up- or down-regulated . In conclusion , we show that FAIRE-seq and ATAC-seq based open chromatin profiling , combined with motif discovery , is a straightforward approach to identify functional genomic regulatory regions , master regulators , and gene regulatory networks controlling complex in vivo processes .
Gene expression in higher eukaryotes is tightly controlled by complex cis-regulatory systems consisting of multiple enhancers modulating the transcription levels of a target gene . Mapping all active promoters and enhancers in a cell type provides an entry point to reverse engineer the functional gene regulatory network and to decipher the genomic cis-regulatory logic underlying a cell type’s functional transcriptome . Understanding changes in regulatory landscapes between different cell types , for example during cellular differentiation , or between normal and diseased cellular states , can furthermore provide a bridge between the genomic sequence and emerging transcriptome changes . Recent advances in chromatin studies have uncovered several characteristic features of active and repressed chromatin within and around regulatory regions . A typical active promoter or enhancer in higher eukaryotes is depleted for nucleosomes over relatively large regions of up to hundreds of base pairs , often spanning the entire enhancer length [1] . In addition , nucleosomes flanking active regulatory regions usually carry histone modifications , such as H3K27Ac and H3K4me1 [2] , and in human and other vertebrates , active promoters and enhancers may have dynamic hyper- or hypomethylated CpG dinucleotides , denoting inactive and active enhancers respectively , whereby changes in methylation usually accompany changes in activity [3] . All of these features , to some extent , correlate significantly with the expression of the target gene ( s ) that they control [1 , 4] . Furthermore , these features can be exploited to identify regulatory regions at a genome-wide scale , for example , integrative methods have been developed based on Hidden Markov Models that accurately predict various types of regulatory regions using particular combinations of histone modifications [5] . Another , more recently discovered feature of active enhancers is that RNA , known as eRNA , is transcribed from them in a bidirectional way by RNApolII [4] . This property was used in the Fantom project and lead to the identification of ∼44 , 000 tissue and cell type specific enhancers in the human genome . However , the technique that has been primarily used to map regulatory landscapes across hundreds of human cell lines , notably in the ENCODE and Epigenomics Roadmap projects , is the detection of open chromatin using DNaseI hypersensitivity coupled with high throughput sequencing ( DNaseI-seq ) [1 , 6 , 7] . DNaseI-seq identifies regions by fragmenting chromatin with the DNase I enzyme , an endonuclease that randomly cleaves regions of accessible DNA , and for which cleavage is hindered by the presence of nucleosomes . This cleavage results in an increased number of cut sites in nucleosome-depleted regions , generating fragments with smaller sizes , allowing their enrichment based on size-selection , before high-throughput sequencing . As such , DNaseI-seq has mapped genome-wide regulatory element across many human and mouse cell types , as well as across developmental stages during Drosophila embryogenesis [8] . DNaseI-seq however , has some limitations as it requires large amounts of input material and has a complicated procedure . Consequently , it has been mainly applied to in vitro cell cultures and cancer cell lines and has seen limited applications to in vivo post-embryonic biological systems . For example , it is generally not possible with this approach to assay to what extent the open chromatin profile of cells within a tumor , or during tumor development , is altered . It is indeed conceivable that the joint interpretation of cancer genomes and cancer transcriptomes , both of which can be sequenced directly from tumor biopsies down to the single cell level , will require the intermediate layer of the “functional genome” to permit us to understand how the epigenomic changes drive changes in gene expression . Therefore , to profile this chromatin layer , alternative approaches are required to progress towards smaller cell populations . Two alternative methods for open chromatin profiling were recently developed which overcome the limitation of sample size , namely FAIRE-seq ( Formaldehyde Assisted Isolation of Regulatory Elements ) [9] , and ATAC-seq ( Assay for Transposase Accessible Chromatin ) [10] . We wanted to know whether these methods are both suitable to identify the functional regulatory elements operating in a wild type tissue or in a tumor . FAIRE-seq relies on the separation of open versus closed chromatin by phenol-chloroform extraction , whereby fragments with high nucleosome content are captured in the organic phase , and sequencing libraries are prepared from the aqueous phase [9] . FAIRE-seq derived genome-wide enhancer maps , although noisier than DNaseI-seq have been shown to be highly correlated with DNaseI-seq [11] . FAIRE-seq has recently been applied successfully to identify differential enhancer usage between multiple Drosophila tissues and developmental time points , finding thousands of enhancers that change activity during development [12] . The most recent method , ATAC-seq , uses a bacterial ( Tn5 ) transposase , an enzyme that inserts a short fragment of DNA ( a transposon ) into another molecule of DNA , or in this case , inserts two short fragments separate from each other [10] . As the transposase is unable to access DNA that is either bound by nucleosomes or strongly bound transcription factors , it integrates its transposons preferentially into accessible or open chromatin . In the case of ATAC-seq the transposase inserts two fragments of DNA which act as tags , and a mechanism to fragment the DNA ( as they are inserted ∼9 base pairs apart ) , a process known as tagmentation [13] . Buenrostro et al . recently showed that ATAC-seq was able to accurately identify the nucleosome free regions of a lymphoblastoid cell line ( GM12878 ) , and the authors were able to obtain profiles comparable to DNaseI-seq in signal to noise ratio and specificity from much lower quantities of cells ( 100–10000 fold less ) than generally used for DNaseI-seq [10] . This technique is thus promising to apply to dissected tissues , for example by micro-dissection of tumor samples , sorting of low populations of cells or other low-input samples . In this study we compare FAIRE-seq and ATAC-seq to discover active regulatory regions in a wild type tissue and during tumor development , and compare both approaches in terms of signal-to-noise ratio , accuracy of enhancer identification , resolution to recover TF footprints , and ability to identify changes during cellular state transitions , such as during oncogenesis . As a model system , we have used the developing Drosophila eye on the one hand , and a genetically induced tumor model in the developing Drosophila eye on the other hand . The tumor model is based on the over-expression of oncogenic RasV12 combined with the loss of the cell polarity gene scribble ( scrib ) , and is a well-established model to study Ras-dependent oncogenesis in vivo [14 , 15] . The combination of RasV12 and scrib-/- mutations cause differentiation defects , coupled with over-proliferation and invasion of the surrounding tissue [16] . Whether such a severe phenotypic change is driven by transcription factors operating within a stable , unchanged chromatin state , or whether changing gene expression profiles are accompanied by widespread changes in the regulatory landscape , has not yet been investigated . Given the similarity between this invasive cancer model and the mesenchymal state transitions occurring in human epithelial cancers , such as epithelial-to-mesenchymal transition , further insight into chromatin modulation in this model system may also be relevant to understand regulatory changes during human oncogenic processes . Our results suggest that open chromatin profiling can provide valuable and previously unobtainable information crucial for understanding how regulatory information is encoded in the genome and how regulatory programs determine phenotype and behavior in vivo .
The larval eye-antennal disc of Drosophila melanogaster is a widely used model system for the study of spatio-temporal gene regulation and cellular differentiation . This is because in one tissue both dimensions of "time and space" are present , with cells in many different states , from pluripotent cells and specified neuronal precursors , to the lineage committed sensory neurons and accessory cells [17–20] . To map the plethora of regulatory regions operating in all of these cell types , and to simultaneously assess the performance of different biochemical approaches to identify accessible regions , we applied both ATAC-seq and FAIRE-seq to several different genotypes of wild type eye-antennal discs ( see Materials and Methods ) ( Fig . 1 ) . When looking at well-known master regulators of eye development , such as Optix and sine oculis ( so ) , we found both methods to yield highly reproducible results whereby the open chromatin peaks mark previously known enhancers within these genes ( Fig . 1 ) . Peaks called by both methods highly overlap ( Fig . 2A ) , and the normalized peak heights between the methods are strongly correlated ( Fig . 2B-D ) . Both methods show a similar distribution of peaks within promoters , intronic , and intergenic regions , and a very strong depletion in coding exons , although ATAC-seq peaks are slightly more biased towards distal regions relative to promoter peaks than FAIRE-seq ( S1 Fig . ) . To investigate how accurately these methods can identify promoters , enhancers , and insulators , we looked in turn to each of these three classes of regulatory elements . Firstly , the accessibility of promoters and transcription start sites is overall highly enriched for both methods , although ATAC-seq shows slightly lower background levels , indicating a higher signal-to-noise ratio for ATAC-seq ( Fig . 2E ) . Secondly , to evaluate whether distal open chromatin peaks also identify functional enhancers , we made use of two collections of Drosophila enhancers , namely the REDFly database and the FlyLight [21] database . Out of 56 REDFly enhancers that are active in the eye , ATAC-seq and FAIRE-seq recover 28 and 24 respectively whilst combined they recover 34 . The recovery rates for eye enhancers are the highest of all tested enhancer categories from REDFly , which illustrates the specificity of these approaches at similar thresholds in the ranked list of all genomic regions ( S2 Fig . ) . Likewise , eye enhancers from the FlyLight database show increased chromatin activity ( Fig . 2F ) ; and the highest overlap between our open chromatin peaks and all FlyLight enhancers is found for eye enhancers ( 402 out of 576 eye enhancers have an open chromatin peak ) , much more than for adult brain enhancers and embryo enhancers [22] ( Fig . 2G ) . Although both approaches have a good performance for enhancer detection , ATAC-seq shows a higher recall of true enhancers than FAIRE-seq , detecting ∼18 . 75% more enhancers ( blue versus red bars in Fig . 2G ) , even at similar levels of specificity ( S2 Fig . ) . Thirdly , we also asked whether the open chromatin peaks overlap with one more regulatory genomic feature , namely functional insulators . To identify insulators , we performed ChIP-seq against CTCF in the wild type eye-antennal disc , under the same conditions as the ATAC-seq and FAIRE-seq ( see Materials and Methods ) . We identified 3682 CTCF peaks , which are significantly enriched for the CTCF motif ( PeakMotif adj . p-value = 2 . 63*10–15 ) ( Fig . 3A ) . Based on this motif , we selected 805 high-confidence CTCF binding sites , having both a significant peak and a significant motif . Both ATAC-seq and FAIRE-seq signals are strongly enriched in the regions around the CTCF binding sites ( Fig . 3A-C ) . Generally , CTCF peaks were identified as accessible regions , with 2244 of all the 3682 CTCF peaks overlapping ( minimum 40% ) with ATAC or FAIRE peaks . Of the 805 high-confidence CTCF sites , 279 and 278 are effectively ( minimum 40% overlap ) called as peaks by ATAC-seq and FAIRE-seq respectively , indicating a highly similar detection rate of both techniques . Thus , both approaches allow for efficient genome-wide detection of promoters and enhancers whilst also detecting insulator sites , starting from low input tissue samples . To further analyze the resolution of both open chromatin profiling methods we assessed whether ATAC-seq and FAIRE-seq can be used for transcription factor footprinting [23] . In the original publication of ATAC-seq [10] , it was shown that CTCF binding regions , determined by ChIP-seq , show increased ATAC-seq signals , while certain nucleotides within the actual CTCF binding site are protected from tagmentation , similar to protection from DNaseI cleavage in DNaseI-seq . To test whether this is also the case in Drosophila eye development , we centered the 805 high-confidence CTCF peaks on the best scoring occurrence of the CTCF motif ( we used the strongest enriched PWM ) to investigate open chromatin signals around CTCF binding sites ( Fig . 3A ) with background signal removed . Similarly to DNaseI-seq , for ATAC-seq the actual CTCF sites show a clear drop in cut sites . This indicates that , although the regions are generally open and therefore accessible by the transposase , the CTCF protein protects its binding site from being cut at specific nucleotides ( Fig . 3B-C ) . This further demonstrates the high degree of resolution obtained with ATAC-seq while identifying active genomic regulatory elements . A second structural feature of ATAC-seq , as shown in the original publication for human chromatin in vitro , is its ability to correlate distances between cut sites with nucleosomal positioning [24] . To assess whether such information could also be obtained from small in vivo samples in Drosophila , we sequenced two samples with paired-end sequencing at low coverage , and found a distribution of insert sizes that almost perfectly resembles that of human chromatin ( Fig . 3D-E ) . Interestingly , this analysis also shows that even by shallow sequencing , with as few as 0 . 5 million mapped reads , regulatory regions can be accurately identified across the entire genome due to the high signal-to-noise ratio of ATAC-seq ( S3 Fig . ) , and that both single-end and paired-end sequencing provide near identical results ( S4 Fig . ) . Taken together , these two experiments illustrate that ATAC-seq can identify nucleosome-depleted regions , and protected nucleotides at high resolution , even from low-input in vivo material from a heterogeneous tissue . Encouraged by the power of ATAC-seq and FAIRE-seq to identify active regulatory regions in heterogeneous wild type tissues , we combined both methods to map all the differentially activated regulatory regions during tumor development . To this end , we used a well-established cancer model in which the eye disc is transformed by over-expression of oncogenic Ras protein ( RasV12 ) in combination with a homozygous scrib-/- mutation ( Fig . 4A-E ) . The combination of these two perturbations in clones of cells in the eye disc has been shown to generate invasive tumors and to serve as a bona fide cancer model [14 , 25–29] , but there is only limited molecular and pathway characterization of these tumors . To identify which parts of the genome are differentially open in the tumors we performed ATAC-seq and FAIRE-seq both on early tumors ( RSE , Fig . 4C , D ) and late tumors ( RSL , Fig . 4E ) ( see Materials and Methods ) and compared their open chromatin landscapes to that of the wild type tissue . We found that the regulatory landscape in tumors is drastically different from the wild type , having both thousands of significantly increased peaks ( 4851 for ATAC-seq ) and thousands of significantly decreased peaks ( 4984 for ATAC-seq ) . Notably , the dynamic range of ATAC-seq seems to be greater than that of FAIRE-seq , as ATAC can detect both smaller and greater significant differences between normal and tumor states than FAIRE ( Fig . 4F ) . Interestingly , when we apply a statistical model allowing for the analysis of both ATAC-seq and FAIRE-seq signals together , the total amount of significantly opening and closing peaks increases to 11516 ( Fig . 4G and Materials and Methods ) ; these differential peaks have slightly different genomic distributions , although at present it is unclear which mechanism could be causing this ( S5 Fig . ) . Since a database with regulatory regions specific for these tumors does not exist , we needed an alternative approach to test whether our candidate regulatory regions are indeed functional regions affecting gene expression of their candidate target genes . To investigate this , we ranked all our regions according to their fold-change between wild type and tumor tissue , and linked them to their candidate target genes based on their location in the genome ( see Materials and Methods ) . Using publicly available gene expression data obtained under the same conditions as ours ( GEO accession: GSE42938 ) [30] we examined the enrichment of the differentially expressed genes in the RasV12; scrib-/- tumors relative to this ranked region-gene list . The enrichment plots in Fig . 4H ( by Gene Set Enrichment Analysis , see Materials and Methods ) show that the up-regulated genes strongly correlate ( p-value < 10–7 , and Normalized Enrichment Score ( NES ) = 2 . 4 ) with the differentially opening regions , while genes down-regulated in the tumor strongly correlate ( p-value < 10–7 , NES = -2 . 4 ) with the genomic regions showing a decrease in open chromatin ( differential gene-region pairs are available in S1–S2 Tables ) . The correlation still holds when we stratify our putative regulatory regions in two groups based on their distance from a TSS ( proximal versus distal ) or when we assign an alternative way to assign peaks to genes ( S6 Fig . ) . This indicates that the differentially opening/closing chromatin regions in the tumor tissue are overall functional and play a role in the perturbation of the transcriptome . An example of a tumor-specific regulatory region is a previously unknown putative enhancer within an intron of p53 for which the strongly increased chromatin opening likely points to the underlying activation of p53 in the tumor cells ( Fig . 4I ) . Next , we wanted to test whether the newly opened chromatin during tumor formation corresponds to functional regulatory regions that have an endogenous role in other tissues during development . To this end we took the 4111 differential peaks whose normalized number of reads was at least two fold increased in the tumor samples when compared to the wild type control , and clustered these peaks into 3778 unique candidate regulatory regions that gain activity in the tumor ( see Materials and Methods ) . We compared this set of differentially active regions against the entire collection of REDFly , FlyLight , and VDRC enhancer resources , covering a total of more than 16000 enhancers with at least one tissue of activity . The most significant overlap was found with sets of enhancers known to be active in the “leading edge of invading tissue” , and with “epidermis” and “midgut primordium” ( S3 Table ) . These relationships may indicate re-activation of endogenous invasive processes . Activated regulatory regions also overlap with genitalia enhancers , which may indicate a re-activation of germline expression typical for pluripotent stem cells , corresponding to previous studies [31] . An example of such a gene is fruitless ( fru ) , for which our data pinpoints a previously described genitalia enhancer [21] that may underlie the observed over-expression of fru during the oncogenic process in the RasV12; scrib-/- cancer model , and in another cancer model in the Drosophila brain [31] ( S7 Fig . ) . In conclusion , profound changes in the open chromatin landscape can be identified between wild type tissue and RasV12; scrib-/- tumor tissue using ATAC-seq and FAIRE-seq . In the tumor , many endogenous enhancers and promoters are ectopically activated , and their activity strongly correlates with changes in gene expression . Having identified the exact locations of activity gaining regulatory regions in the tumors downstream of oncogenic RasV12; scrib-/- provides us with a high-quality set of sequences that are likely regulated/bound by a shared set of transcription factors . To predict which transcription factors might be binding to these activity gaining tumor-induced regulatory regions , we used a recently developed motif discovery method , known as i-cisTarget [32] . From a total of 9713 candidate TF motifs ( as position weight matrices ( PWM ) ) , i-cisTarget yielded the AP-1 and Stat92E motifs as the two most enriched , AP-1 is ranked 1st with a normalized enrichment score ( NES ) of 8 . 73 and Stat92E is ranked 2nd with a NES of 5 . 1 . In the active , but unchanging , regulatory regions the Stat92E motif is not enriched and there is only a minor enrichment for the AP-1 motif ( NES = 2 . 7 ) , indicating that the enrichment is specifically in the activity gaining regulatory regions . ( see Materials and Methods for an explanation of the NES score ) . Remarkably , of the tumor-induced regulatory regions , AP-1 is predicted to target the majority ( 3065 of the 3778 , or 81% ) , pointing to its driving role in oncogenesis . The AP-1 complex is a homo- or heterodimer of bZIP proteins such as Jun and Fos , binds to highly similar DNA motifs , and is functionally activated downstream of phosphorylated Jun Kinases ( JNK ) . Interestingly , multiple labs have shown the importance of JNK signaling , and of the AP-1 complex , specifically to the development of RasV12; scrib-/- tumors . Particularly , either suppressing JNK signaling or knocking-down the AP-1 complex , is sufficient to block tumor development [14 , 29 , 33 , 34] . The Stat92E motif is enriched in a smaller part of the regulatory regions and shows a significant overlap with the predicted AP-1 responsive regions ( Fig . 5A , B ) . This is consistent with reports that show that these pathways can have synergistic effects on RasV12; scrib-/- tumorigenesis [16] . Besides AP-1 and Stat92E , a few additional motifs for other TFs are also significantly enriched , although with a lower representation ( Fig . 5A ) . One of them is the Scalloped ( Sd ) motif , a transcription factor that acts together with its coactivator Yorkie ( Yki ) to promote tissue overgrowth , as effectors of the Hippo signaling pathway [35] . It has been shown that knockdown of either Scalloped or Yorkie can rescue scrib-/- mutant tissue overgrowth and reduces RasV12; scrib-/- tumor size [36] . Another group are the zinc-finger protein motifs of Zelda ( Vfl ) that are known for their activating roles in early Drosophila development [37] , and for which a role during oncogenesis has not been described . Note that not all enriched motifs are necessarily involved in the regulatory oncogenic program , and some can be “bystander” motifs for the key regulators . An example of such a bystander motif could be the Zelda motif , which often co-occurs with Stat92E motifs in the same regulatory region because these two TFs cooperate during early embryonic development [38] , but in the eye tumor the zelda gene is not mis-regulated , while all 3 the ligands of the JAK/STAT signaling pathway are ( S1 Table ) . Finally , we recovered several nuclear receptor motifs ( e . g . , Ftz-f1 and Hr39 ) and motifs of the TFIIB-related factor ( Brf ) , which increases RNA polymerase III-mediated transcription , and its overexpression has been linked to several human cancer types [39] . Interestingly , this set of candidate Brf-regulated tRNA genes could be discovered by open chromatin profiling , but not by microarrays or mRNA-seq , showing another advantage , and complementarity , of using chromatin profiling besides classical gene expression profiling . Next we asked whether the activated regulatory regions in the early tumors become even more activated in the late tumors , or if different regulatory regions become activated during tumor progression . In the early tumor , about half of the eye disc still consists of normal tissue , while in the late phase the tumor tissue has overtaken the entire eye disc and invades into the optic lobes of the brain and the ventral nerve cord to a greater extent . By comparing wild type versus early tumors , and early tumors versus late tumors , we found that the majority of changes are gradual , showing an increase in the height of the open chromatin peak between wild type and early tumor , and a further increase between the early and the late tumor . Such a gradual increase of the peak height on a regulatory region most likely indicates that this region is becoming accessible in a higher fraction of cells in the dissected tumor tissue , which is likely the consequence of a lower percentage of normal cells in the late tumor tissue . On the other hand , we found a subset of regulatory regions that are more open in the early tumor versus wild type , but do not show an increasing signal in the late tumor ( determined by Fisher’s omnibus , see Materials and Methods ) . Interestingly , the motif enrichment scores for AP-1 and Stat92E are very different between the gradually and stably opening regulatory regions . More specifically , the stably open regions are mainly enriched for Stat92E ( ranked 1st , NES = 7 . 83 ) , while the enrichment for AP-1 motifs is reduced ( ranked 3rd , NES = 4 . 74 ) . On the other hand , the gradually open regions are strongly enriched for AP-1 motifs ( ranked 1st , NES = 12 . 4 ) , while the Stat92E motif is no longer enriched in this set ( NES < 2 . 5 ) ( Fig . 5C ) . This finding may indicate that , either Stat92E targets are activated earlier than AP-1 targets , or that a relatively small proportion of cells in the invasive tumor retain Stat92E activity . It could also indicate that Stat92E is active in both tumor and non-tumor cells , and that the secretion of Unpaired ligands from the tumor cells can cause non-autonomous activation of the JAK/STAT pathway in surrounding non-tumor tissue [40] . In conclusion , motif inference on differentially open chromatin peaks during tumor development provides valuable hypotheses about the identity of the transcription factors that are driving the oncogenic regulatory program . Our data confirms previous observations of an important role for AP-1 and Stat92E downstream of RasV12; scrib-/- induced oncogenesis , and now suggests that these two regulators can explain a very large fraction of the changing chromatin landscape . One of the predicted transcription factors involved in changing the open chromatin landscape in the tumors is Stat92E , the effector of the JAK/STAT signaling pathway [41] . All three the ligands of this pathway are present in our list of significantly up-regulated gene-peak pairs ( S1 Table ) , strongly supporting the predicted role of JAK/STAT signaling in the tumor . Indeed , previous reports have shown that blocking the activity of the JAK/STAT pathway , for example by using a dominant negative form of the receptor Domeless , reduced the tumor phenotype [16] . However , the direct involvement of Stat92E in this process has not been explicitly tested . We therefore incorporated a null mutation of Stat92E in the RasV12; scrib-/- tumors ( Stat92E[85c9] , see Materials and Methods ) to determine if Stat activity is required for the tumor phenotype . We observed that tumor growth was severely reduced ( Fig . 6A-C ) . In addition , in the Stat92E loss-of-function , a fraction of larvae with reduced tumors now also reach pupation stages , which is not observed in RasV12; scrib-/- larvae . Based on our motif predictions ( see above ) , Stat92E may activate more than three hundred regulatory regions and thereby play a role in regulating the nearby target genes , downstream of RasV12; scrib-/- during oncogenesis . If these promoters and putative enhancers depend on Stat92E for their opening , the same regions may be activated and opened by Stat92E alone , independently of the RasV12; scrib-/- induced oncogenesis . To test this we hyper-activated the JAK/STAT pathway in the wild type eye disc , and thereby the downstream Stat92E activity , by overexpressing one of the ligands that is up-regulated in the tumor , namely Unpaired ( Upd , encoded by os ) . ATAC-seq on these discs could demonstrate that the tumor-induced changes at Stat92E predicted target sites are recapitulated by JAK/STAT activation alone . The changes are significant and mainly in the same direction ( Fig . 6D ) , but as expected the changes in open chromatin caused by the Upd overexpression are quantitatively more subtle compared to those in the tumors . For example , an intronic regulatory region of Imp shows activated chromatin by Stat92E ( log2 fold change = 1 . 23 ) but this activation is stronger during RasV12; scrib-/- tumor formation ( log2 fold change = 3 . 44 ) ( Fig . 6E ) . In the RasV12; scrib-/- tumors we had identified 356 candidate target regulatory regions of Stat92E , of these we found that 72% have a positive fold change in the Upd over-expressing discs ( Fig . 6D ) , indicating that this group is Stat92E responsive ( p-value 0 . 0097 ) . However , this still means that 28% of the candidate regions did not respond to the activation of Stat92E by overexpressing Upd . We analyzed these responsive and non-responsive regions and found that the AP-1 motif is only enriched in the regions that are not responsive to Stat92E , indicating that for those regions AP-1 might be the main input and the Stat92E motif might be of less importance here . Next , we asked whether these changes in cis-regulatory activity directly affect gene expression of target genes . To test this , we linked the 254 Stat92E responsive regulatory regions to nearby ( <5kb upstream or intronic ) candidate target genes and examined their expression in Upd over-expressing discs using publicly available gene expression data ( GEO accession: GSE15868 ) [42] . After applying two filtering steps , requiring the regulatory region to go significantly open both in tumor and UPD-overexpressing tissue , combined with a significant differential expression of the assigned gene , we end up with a final list of 28 potential direct Stat92E target genes ( Fig . 6F ) . Among these are at least seven well-known transcriptional targets of the JAK/STAT signaling pathway including: dorsal , pointed , lama , chinmo and trol , as well as two key negative regulators of this pathway: Socs36E and Ptp61F [42–49] . Interestingly , not all genes that have an opened Stat92E regulatory region are up-regulated . The JAK/STAT pathway can also repress transcription of target genes and is known to block the Wnt/Wingless signaling pathway in the eye imaginal disc [50] . In our list we recover two genes that are known to be involved in the Wingless signaling pathway , wingless itself and sulfateless , an essential enzyme for this pathway [51] . This may indicate that Stat92E can function directly in the repression of genes , through an as yet unidentified mechanism . An interesting future challenge will be to understand the cis-regulatory mechanisms and possible co-factors that determine the mechanisms by which Stat92E can act as activator or repressor . In conclusion , we demonstrate that Stat92E is a significant transcriptional regulator and required for the growth of our tumor model . We identify known and newly predicted Stat92E targets in the tumor and are able to independently recover >70% of those targets using a Stat92E activation model in the normal eye . This second analysis also illustrates the feasibility of an integrated approach of ATAC-seq and motif discovery to capture and annotate modest , yet functional , changes in the chromatin landscape . We conclude that Stat92E induced chromatin-opening correlates with a change in the transcription of nearby genes , and that Stat92E may function as a key regulator of the tumor transcriptome .
Genome-wide characterization of all promoters and enhancers controlling a particular gene expression profile and/or phenotype is a key challenge for understanding the regulatory underpinnings of any biological process in vivo . Our study first compares , and subsequently combines , two recent methods for open chromatin profiling to obtain genome-wide regulatory landscapes in the developing Drosophila eye as model system . By applying FAIRE-seq and ATAC-seq to the normal eye-antennal imaginal disc , alongside anti-CTCF ChIP-seq , we found both methods to be highly robust in identifying accessible or open regions , with few differences between strains with different genetic backgrounds . The main advantages of ATAC-seq , for our application , are besides its undemanding experimental procedure , ( 1 ) its higher signal-to-noise ratio , with low background signal and sharper peaks; ( 2 ) its ability to identify TF footprints , as binding sites are protected from transposon insertion , similar to their protection from DNaseI cleavage [23]; and ( 3 ) its ability to determine nucleosome positioning when using paired-end sequencing . Although these assets of ATAC-seq may be important for some studies , overall both FAIRE and ATAC allow identification of promoters and candidate enhancers , and here we use them as independent “replicate” measurements ( taking batch effects into account ) to examine the open chromatin status of a tumor model . We observe that the cis-regulatory landscape of active promoters and enhancers changes dramatically in RasV12; scrib-/- eye tumors , while more moderate changes were observed in tissues with JAK/STAT induced hyperplasia . A possible explanation for the marked differences in the intensity of change may be found in the cellular composition of the samples . Open chromatin signals represent the average signal across all cells within a sample , and since each regulatory region , in each cell can yield only two open alleles , this observed activity mainly reflects the number of cells in which the region is active , rather than the quantitative activity of the region . Once the repertoire of tumor-induced regulatory regions was identified by open chromatin profiling , we reasoned that if many of these functional regions are activated by a small set of master regulators , then the motifs of these TFs should be enriched within the sequences of these regions . While motif inference on gene sets is challenging due to the large intergenic and intronic regulatory space around genes , motif inference on enhancer-size regulatory sequences often gives highly accurate results [32 , 52–54] . We used the tool i-cisTarget , which is optimized for Drosophila genomes , and found not only AP-1 and Stat92E , but also Zelda , Scalloped ( the Hippo pathway effector ) , Brf , and Ftz-f1 as candidate regulators of the oncogenic , Ras-dependent program . All of these TFs ( except perhaps Zelda ) can be directly linked to cancer-related processes that are conserved to human [35 , 39] , and AP-1 , Stat , and Scalloped have each been previously linked to the RasV12; scrib-/- program specifically [16 , 29 , 36 , 55 , 56] . We could confirm that the RasV12; scrib-/- tumor phenotype depends on Stat92E , and furthermore reveal that JAK/STAT signaling causes specific chromatin changes at Stat92E-responsive regulatory regions . Finally , once the regulatory regions and their candidate regulators are identified , an important next step is to examine which target genes are now differently regulated , as a consequence of the activation of these promoters and enhancers . Previous work has found that open chromatin peaks obtained across cell lines are correlated with gene expression changes of “nearby” target genes [4] . We also found evidence that a high degree of chromatin changes are concordant with transcriptional changes of nearby target genes , both in the tumor and in the overgrown tissue with hyper-activated JAK/STAT , confirming that the chromatin state of these regions is not only altered , but that they are also functionally activated . Interestingly , not all putative target genes with an increased open chromatin peak are up-regulated , but a small subset is also down-regulated , which indicates that regulatory regions can be “activated” by TF binding and nucleosome depletion , but that the consequence of this activation can also be gene repression ( e . g . , if the bound TF act as a repressor ) . Overall , our integrated approach reveals a large cistrome of changing activity at promoters and enhancers during tumor development , which is mainly operated by AP-1 and Stat92E; and illustrates how integrative open chromatin profiling , motif detection , and gene expression analyses have great potential to unravel tissue and cell type specific regulatory programs in vivo , in health and disease .
The following fly stocks were used during this investigation: y , w , eyFlp; act>y+>Gal4 , UAS-GFP; FRT82 tub Gal80 and y , w; UAS-RasV12; FRT82 scrib2 , e/ TM6B , yw; UAS-Rasv12; FRT82 Stat92E85c9 , scrib2 , e/ FRT82 tub-Gal80 , Optix-GFP [57] and UAS-GFP:Upd ( Courtesy of Fernando Casares ) , GMR-Gal4 , isogenic wild type DGRP-208 ( Bloomington stock 25174 ) , y , w; FRT82 and CantonS wild types . Crosses were raised at 25°C on a yeast based medium . Wild type , RasV12; scrib-/- early and Upd overexpression eye-antennal discs were dissected from wandering third instar larvae ( day 6 ) in 1xPBS . RasV12; scrib-/- late discs were collected three days after larvae began wandering ( day 9 ) ; this is possible because RasV12; scrib-/- do not pupate , but can persist more than one week in a prolonged larval stage . Immunohistochemistry was performed as previously described in [58] . Confocal images were taken on an Olympus FV1000 or FV1200 microscope . Images were processed using ImageJ and Adobe Photoshop software . The previously described ATAC-seq protocol was adapted for working with Drosophila rather than human cells [10] . Ten eye antennal imaginal discs ( or three RasV12; scrib-/- late total tumors ) were immediately placed in 50 μl ice cold ATAC lysis buffer ( 10 mM Tris-HCl , pH 7 . 4 , 10mM NaCl , 3mM MgCl2 , 0 . 1% IGEPAL CA-630 ) . Lysed discs were then centrifuged at 800 xg for 10 minutes at 4’C and the supernatant was discarded . The rest of the ATAC-seq protocol was performed as described previously [10] using the following primers: Fwd:- ‘AATGATACGGCGACCACCGAGATCTA CACTCGTCGGCAGCGTCAGATGTG’ and Rev:- ‘CAAGCAGAAGACGGCATACGAGATXXX XXXGTCTCGTGGGCTCGGAGATGT’ ( where X indicates barcode nucleotides ) . The final library was purified using a Qiagen MinElute kit ( Qiagen ) and Ampure XP beads ( Ampure ) ( 1:1 . 2 ratio ) were used to remove remaining adapters . The final library was first checked on an Agilent Bioanalyzer 2000 for the average fragment size . Resulting successful libraries were sequenced with 50bp , single end reads on the Illumina HiSeq 2000 platform . Single end sequencing was chosen for this study because we were not interested in the fragment contents ( i . e . , how many nucleosomes are placed between two insertion sites ) , rather just the profile of insertion sites , which also made comparisons with pre-existing FAIRE-seq data easier . Methodology adapted from [12] . In short , 150 head complexes were dissected from wandering third instar larvae , these were fixed for 10 min with 4% formaldehyde , . The formaldehyde was then replaced with 750 μl quenching buffer ( 125 mM Glycine 0 . 01% Triton X-100 in PBS ) was added and incubated at room temperature for 10 minutes . Quenching buffer was replaced with buffer A ( 10 mM HEPES-KOH pH8 . 0 , 20 mM EDTA pH8 . 0 , 1mM EGTA pH8 . 0 , 0 . 25% Triton X-100 , 1mM PMSF ) and 200 eye-antennal discs were then dissected in buffer A and kept on ice , these were centrifuged at 6000rpm , 4’C to pellet the discs and lysed in lysis buffer 1 ( 50mM HEPES-KOH , pH 7 . 5 , 140mM NaCl , 1mM EDTA , 10% glycerol , 0 . 5% NP-40 , 0 . 25% Triton X-100 ) rocking at 4’C for 10 minutes , centrifuged and the supernatant removed . Next lysis buffer 2 ( 10mM Tris-HCl , pH 8 . 0 , 200 mM NaCl , 1mM EDTA , 0 . 5 mM . EGTA ) was added and the sample was rocked at room temperature for 10 minutes . Finally lysis buffer 3 ( 10 mM Tris-HCL , pH 8 . 0 , 100mM NaCl , 1mM EDTA , 0 . 5mM EGTA , 0 . 1% Na-deoxycholate , 0 . 5% N-lauroylsarcosine ) was added and samples were sonicated ( Bioruptor UCD-200 ( Diagenode ) at 4’C , 8 Cycles set to pulse high 30 seconds , rest 30 seconds ) immediately . Double phenol/chloroform extraction was performed with a final chloroform extraction . DNA was precipitated using Sodium acetate ( 0 . 3 M , pH 5 . 2 ) , 20mg Glycogen and 100% ethanol . DNA was washed with 500 μl ice cold 70% Ethanol . Supernatant was removed and the pellet air-dried . The dried pellet was re-suspended in 50 μl TE buffer ( 10 mM Tris pH 8 . 0 , 1 mM EDTA in MilliQ water ) and incubated at 65’C overnight to reverse crosslinks . Finally 1 μl 10mg/ml RNAseA was added and incubated at 37’C for 1 hour , samples were cleaned using the QiaQuick MinElute kit ( Qiagen ) and DNA was measure using a Qubit analyzer . Final libraries were prepared as per standard Illumina protocols . Head complexes were dissected from wandering third instar larvae ( 500 , in batches of 100 ) and fixed in 1ml crosslinking solution ( 1 . 8% formaldehyde , 50 mM HEPES pH 7 . 9 , 1 mM EDTA , 0 . 5 mM EGTA , 100 mM NaCl in MilliQ water ) for 25 minutes at room temperature while rotating . Crosslinking solution was replaced 5 times after 5 minutes each time . Crosslinking was stopped by adding 1ml stop solution ( 0 . 01% Triton X-100 , 125 mM Glycine in PBS ) and incubating at room temperature for 10 minutes while rotating , this was repeated for 3 more washes . Complexes were washed in 1 ml wash A ( 10 mM HEPES pH7 . 9 , 10 mM EDTA , 0 . 5 mM EGTA , 0 . 25% Triton-X100 in MilliQ water ) for 10 minutes at room temperature while rotating . Complexes were washed in 1 ml wash B ( 10 mM HEPES pH7 . 9 , 200 mM NaCl , 1 mM EDTA , 0 . 5 mM EGTA , 0 . 01% Triton X-100 in MilliQ water ) for 10 minutes at room temperature while rotating , this was repeated 3 more times . Eye-antennal discs ( 200 ) were dissected from head complexes in wash B and collected in a tube containing 100 μl sonication buffer ( 10 mM HEPES pH7 . 9 , 1mM EDTA , 0 . 5 mM EGTA in MilliQ water ) , samples were sonicated ( Bioruptor UCD-200 at 4’C , 8 Cycles set to pulse high 30 seconds , rest 30 seconds ) immediately . Samples were centrifuged at 21000 xg for 10 minutes at 4’C . For each immunoprecipitation , the pellet from 100 μl extract was re-suspended in 1 ml RIPA buffer ( 140 mM NaCl , 10 mM Tris-HCl pH8 . 0 , 1 mM EDTA , 1% Triton X-100 , 0 . 1% SDS , 0 . 1% Na-deoxychoate , 1% PMSF in MilliQ water ) , an extra sample ( 10 μl ) was kept aside as input . Immunoprecipitation was performed by adding 20 μl protein A/G magnetic beads ( Millipore ) and incubating for 1 hour at 4’C; samples were centrifuged at 3000 rpm for 2 minutes and supernatant kept . Anti-CTCF antibody ( 10 μl crude rabbit serum—A kind gift from Dr . Ranier Renkawitz ) was added to the supernatant and rotated at 4’C overnight . Immunocomplexes were recovered by adding 20 μl protein A/G magnetic beads to sample and incubating at 4’C for 3 hours while rotating . Magnetic beads were separated using a magnetic stand and supernatant discarded . Beads were re-suspended and washed for 5 minutes on a rotating platform , with 1 ml of the following buffers in order:- Low salt immune complex wash buffer ( Millipore ) , High salt immune complex wash buffer ( Millipore ) , LiCl immune complex wash buffer ( Millipore ) , TE buffer . ChIP elution buffer ( Millipore ) was warmed to 38’C and 100 μl was added to the beads , 1μl RNaseA was also added , samples were incubated at 36’C for 30 minutes , shaking at 950rpm . Both immunoprecipitated and control samples had 1ul proteinase K added and were incubated at 62’C for 2 hours , shaking at 950rpm and then 95’C for 10 minutes . Samples were allowed to cool to room temperature and beads were separated with a magnetic stand . Final DNA was purified according to the manufacturers guidelines ( MagnaChIP , Millipore ) Final libraries were prepared as per standard Illumina protocols . Raw reads were first cleaned for adapter sequences using fastq-mcf using default parameters and an adapter file containing common Illumina adapter sequences . Cleaned samples were mapped to the Drosophila melanogaster FlyBase release r5 . 53 genome [59] using bowtie2 [60] ( default parameters ) and reads with a mapping quality of less than 4 were discarded . In several samples we discovered reads mapping to the Wolbachia genome ( 60–80% ) , these were also discarded . Both ATAC-seq and DNaseI-seq reads ( publicly available data ) were adjusted to better represent the open chromatin by centering reads on the cut-sites and extending this by 5bp on either side , samples were finally sorted and indexed using Samtools [61] . Peaks were called on ATAC-seq , FAIRE-seq and DNaseI-seq samples using the MACS2 software suite [62] with the added parameters “-g dm –nomodel –shiftsize 50 –q 0 . 01” . For comparison between samples , all peaks from each sample were merged to provide one set of combined peaks . Peaks on ChIP-seq data were called also using the MACS2 software suite with the parameters “-g dm –q 0 . 01” also using input as a control . For quantification of peaks , bed files of combined peaks were converted into a GFF3 format and then the number of reads per peak , per sample were counted using htseq-count [63] . Finally bigWig files were created from bam files for each sample using genomeCoverageBed [64] ( using the –scale option ) and bedGraphToBigWig [65] . Scales were determined by the ‘sizeFactors ( ) ’ command from DESeq2 [66] on a matrix of all samples , counted on all combined peaks . Genomic locations of peaks was determined by the CEAS software package using default parameters and a prebuilt dm3 gene annotation file [67] . For the CTCF protection analysis motif scanning was performed with Cluster Buster to find occurrences of the JASPAR-MA0139 . 1 motif in CTCF peaks using the parameters ‘-m7 –c0’ , each peak was then re-centered on the center of the best scoring motif present . The cut sites from each read was determined and plotted 500bp around the re-centered CTCF peaks . Signal from the same motif at random regions of DNA was subtracted to remove background noise . To determine overlaps between peaks , the tool intersectBed was used , with all CTCF peaks or high confidence regions as file A and the ATAC-seq/FAIRE-seq peaks as file B , the option ‘-f 0 . 4’ was also used to enforce a 40% overlap of the CTCF peak by the ATAC-seq/FAIRE-seq peak . To identify differential peaks between conditions we used DESeq2 [66] with a p-adjusted cutoff of 0 . 01 . This cutoff was supported by the leading edge of a Gene Set Enrichment Analysis ( GSEA [68] ) analysis whereby all genes ranked by their most significant peak ( tumor vs wild type ) are compared to a Ras/Scrib gene signature . To link peaks to genes we assigned a peak within any intron or in the 5kb upstream region of the TSS to the gene . To rank genes according to peak heights ( for GSEA analysis ) we used the peak with the most significant adjusted p-value . When FAIRE-seq and ATAC-seq are used as replicates , we took batch effects into account in DESeq2 . Rankings for the recovery curves seen in S2 Fig . were generated by scoring peaks ( ATAC-seq wild type and FAIRE-seq wild type , merged ) based on the number of reads falling within the region for the appropriate samples . ATAC-seq and FAIRE-seq were ranked individually and their recoveries overlaid . We acquired two Affymetrix Drosophila Genome 2 . 0 Array data sets from GEO [30] , one comprised of 3 wild type and 3 RasV12; scrib-/- biological replicates ( GEO accession: GSE42938 ) [69] , and the second comprised of 5 wild type and 5 Upd overexpression biological replicates ( GEO accession: GSE15868 ) [42] . We discarded three low-quality samples from our analyses ( GSM398336 , GSM398339 , GSM398341 ) . Differential expression analysis was carried out in R using the Bioconductor packages affy , limma , Biobase and GEOquery , applying a standard limma protocol [70] . After obtaining differential values we associated each probe to their respective gene; for genes with more than one associated probe , we decided to use the probe with the most significant adjusted p-value . The differential peaks ( wild type control vs RasV12; scrib-/- tumors ) were assigned to genes ( <5kb upstream or intronic ) and for each gene the most significantly differential peak was kept . The genes were ordered based on the relative openness of their assigned peak ( opening on top and closing on bottom ) , to obtain the ranked gene list ( x-axis of GSEA plots ) . Two groups of differentially expressed genes were determined based on the publically available micro array data ( GEO accession: GSE42938 ) . One group contained the significantly upregulated and the other the significantly downregulated genes in the tumors . We used GSEA ( with 100000 perturbations ) to determine if these groups of differentially regulated genes were significant enriched on either side of the ranked gene list . The globally opening regions in the RasV12; scrib-/- tumors were determined by differential peak calling between all 5 wild type controls and all 4 RasV12; scrib-/- samples ( combining early and late ) . Of the significantly differential peaks , only those with a logFC greater than 1 were selected , ending up with 4111 globally opening regions . To determine the early versus late RasV12; scrib-/- regulatory regions , differential peaks were called between the 5 wild type controls and the 2 RasV12; scrib-/- early samples and between the 2 RasV12; scrib-/- early samples and the 2 RasV12; scrib-/- late samples . We selected a subset of regulatory regions that were opening from WT to RSE ( with a logFC > 1 ) and that remained at similar levels between RSE and RSL ( -0 . 2 < logFC < 0 . 2 ) ; we defined these regulatory regions as ‘stably opening’ . For the regulatory regions defined as ‘gradually opening’ , we selected the regions that are becoming more open between WT and RSE ( logFC > 0 ) and that further open between RSE and RSL ( logFC > 0 . 5 ) . Using a Fisher's Omnibus test we combined the p-values for each regulatory region ( one from WT vs RSE and one from RSE vs RSL ) and obtain a new chi-squared p-value . The regulatory regions opening in Upd-overexpression were determined by differential peak calling between the 3 ATAC-seq wild type controls and the 2 ATAC-seq Upd-overexpressing samples . We took the top 250 opening regions ( ordered on singed P Value ) to perform motif enrichment analysis . For motif enrichment we used i-cisTarget [32] , a tool developed in our lab to discover motifs significantly enriched in our four regulatory region groups ( the stably , globally and gradually opening in RasV12; scrib-/- and opening in Upd-overexpression ) . We ran i-cisTarget via the command line with the rank threshold = 10000 , enrichment score threshold = 2 and a collection of 9713 motifs . The enrichment of each motif in the input set is calculated as an area under the recovery curve ( AUC ) , whereby recovery is observed over a genome-wide ranking of 136K a priori defined candidate regulatory regions [32] . The AUC score is normalized by subtracting the mean of all AUCs over all motifs , and dividing it by the standard deviation , to obtain a Normalized Enrichment Score ( NES ) . We use a cutoff of NES > 2 . 5 to select significantly enriched motifs . Relationships between NES and False Discovery Rates can be found in [71] . For each factor of interest with multiple enriched motifs , we selected the motif with the highest NES score . ATAC-seq , ChIP-seq and FAIRE-seq data for all conditions are available in GEO ( http://www . ncbi . nlm . nih . gov/geo/ ) , with accession number GSE59078 . Genome browser tracks for all data , and called peaks for wild type and cancer-related regulatory regions are all available within a UCSC Genome Browser hub from this URL: http://genome . ucsc . edu/cgi-bin/hgTracks ? db=dm3&hubUrl=http://ucsctracks . aertslab . org/ATAC-paper/hub . txt
|
The functional expression of all genes is regulated by proteins , namely transcription factors that bind to specific areas of DNA known as regulatory regions . Whereas most DNA in our genome is normally bound by other proteins ( histones ) and packaged into units called nucleosomes , a specific subset of tissue-specific regulatory regions is responsible for tissue-specific gene expression; these active regions are nucleosome-depleted and bound by transcription factors . We use two techniques to identify these open chromatin regions , in a normal tissue and a RasV12 induced cancer tissue . We discovered a remarkable change in the accessible regulatory landscape between these two tissues , with several thousand regions becoming more accessible in the cancer tissue . We identified two transcription factors known to be involved in cancer ( AP-1 and Stat92E ) controlling these newly accessible regulatory regions . Finally , we introduced a mutation resulting in Stat92E becoming non-functional in the cancer tissue , which decreased the severity of the tumor . Our study shows that open chromatin profiling can be used to identify complex in vivo processes , and we shed new light on Ras dependent cancer development .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Discovery of Transcription Factors and Regulatory Regions Driving In Vivo Tumor Development by ATAC-seq and FAIRE-seq Open Chromatin Profiling
|
By following the evolution of populations that are initially genetically homogeneous , much can be learned about core biological principles . For example , it allows for detailed studies of the rate of emergence of de novo mutations and their change in frequency due to drift and selection . Unfortunately , in multicellular organisms with generation times of months or years , it is difficult to set up and carry out such experiments over many generations . An alternative is provided by “natural evolution experiments” that started from colonizations or invasions of new habitats by selfing lineages . With limited or missing gene flow from other lineages , new mutations and their effects can be easily detected . North America has been colonized in historic times by the plant Arabidopsis thaliana , and although multiple intercrossing lineages are found today , many of the individuals belong to a single lineage , HPG1 . To determine in this lineage the rate of substitutions—the subset of mutations that survived natural selection and drift– , we have sequenced genomes from plants collected between 1863 and 2006 . We identified 73 modern and 27 herbarium specimens that belonged to HPG1 . Using the estimated substitution rate , we infer that the last common HPG1 ancestor lived in the early 17th century , when it was most likely introduced by chance from Europe . Mutations in coding regions are depleted in frequency compared to those in other portions of the genome , consistent with purifying selection . Nevertheless , a handful of mutations is found at high frequency in present-day populations . We link these to detectable phenotypic variance in traits of known ecological importance , life history and growth , which could reflect their adaptive value . Our work showcases how , by applying genomics methods to a combination of modern and historic samples from colonizing lineages , we can directly study new mutations and their potential evolutionary relevance .
Colonizing or invasive populations sampled through time [1 , 2] constitute “natural experiments” where it is possible to study evolutionary processes in action [3] . Colonizations , which are dramatically increasing in number [4 , 5] , sometimes are characterized by strong bottlenecks and genetic isolation [6 , 7] , and thus greatly facilitate the observation of new mutations and potentially their effects under natural population dynamics and selection [8] . Colonizations thus offer a complementary approach to other studies of new mutations , which often minimize natural selection , for example in laboratory mutation accumulation experiments [9] and parent-offspring comparisons [10] . The study of colonizations is also complementary to the investigation of genetic divergence over long time scales , e . g . , between distant species [11] , where the results are largely independent of short-term demographic fluctuations . There is broad interest in understanding how genetic diversity is generated [12] , and how new mutations can provide a path for rapid adaptive evolution [13–15] . Additionally , accurate evolutionary rates permit dating historic population splits , which is fundamental to the study of population history [16] . The analysis of colonizing populations can also contribute to resolving the “genetic paradox of invasion” [17] . This paradox comes from the observation that colonizing populations can be surprisingly successful and spread very widely and in multiple even when strongly bottlenecked , suggesting some level of adaptation to new environments that goes beyond the exploitation of unoccupied ecological niches [17] . Much of the work in plant ecology and evolution has focused on evidence that populations can rapidly adapt from standing variation [18] . In invasive lineages , initial standing variation may originate from incomplete bottlenecks , multiple introductions , or admixture with local relatives [19] . Much less work has been done with respect to the role of de novo mutations as a solution to the genetic paradox of invasion , although this has been proposed as an alternative explanation for rapid adaptation by colonizing lineages [3 , 17 , 20] . The self-fertilizing plant Arabidopsis thaliana is native to Africa and Eurasia [21 , 22] but has recently colonized N . America , where it likely experienced a strong founder effect [23] . At nearly half of N . American sites sampled during the 1990s and early 2000s , more than 80% of plants belong to a single haplogroup , HPG1 , as inferred from genotyping with 149 intermediate-frequency markers evenly spread throughout the genome [23] . The HPG1 lineage has been reported from many sites along the East Coast and in the Midwest as well as at a few sites in the West [23] ( Fig 1 , S1 Table ) . The great ubiquity of HPG1 in comparison to any other haplogroup could be due to either some adaptive advantage , or , more parsimoniously , be the result of HPG1 being derived from one of the first arrivals of A . thaliana in the continent . Here , we focus on 100 HPG1 individuals that do not show any evidence of outcrossing with other lineages . We combine genomes from herbarium specimens and live individuals , collectively covering the time span from 1863 to 2006 , to infer mutation rates , to date the birth of the HPG1 lineage , and to investigate the evolutionary forces that shape genetic diversity and potentially adaptive trait variation . Our analyses of this lineage serves as a model for future studies of similar colonizing or otherwise recently bottlenecked plant populations , in order to better understand how diversity is generated and to which extent it contributes to adaptation in nature .
In a self-fertilizing species , a single individual can give rise to an entire lineage of millions of offspring , which then diversify through new mutations and eventually intra-lineage recombination . If self-fertilization is much more common than outcrossing , the founder is likely to have been homozygous throughout almost the entire genome . Because it is so wide spread , HPG1 presents an opportunity to sample many natural populations that have been potentially derived from a common , very recent ancestor with such characteristics . In the best possible case , this would allow for new mutations to be directly observed through time . To test these assumptions and to better understand the evolution of HPG1 , we sequenced two different groups of plants . The first group were live descendants of 87 plants that had been collected between 1993 and 2006 ( Fig 1; S1 Table ) , and which had been identified as likely members of the HPG1 lineage with 149 genome-wide markers spaced at roughly 1-Mb-intervals [23] . We aimed for broad geographic representation , with at least two accessions per collection site , where available . The second group comprised 36 herbarium specimens , collected between 1863 and 1993 , for which we had no a priori information whether they may or may not belong to the HPG1 lineage , but which were selected from the herbarium records to cover the full historical geographic range and overlap with modern samples when possible ( Fig 1 ) . The DNA from the herbarium specimens showed biochemical features typical of ancient DNA ( aDNA ) from plants , which we have previously described in detail [24] . Such DNA damage included a median fragment length of 60 bp , an excess of C-to-T substitutions of about 2 . 5% at the first base of sequencing reads and a 1 . 5 to 1 . 8 fold enrichment of purines at DNA breakpoints ( S1 Fig , S2 Text ) . To remove aDNA associated damage and produce high-quality genomes , chemically-repaired libraries ( see Methods ) were later sequenced . These reads were mapped against an HPG1 pseudo-reference genome [25] , focusing on single nucleotide polymorphisms ( SNPs ) because the short sequence reads of herbarium samples preclude accurate calling of structural variants . Genome sequences were of high quality , with herbarium samples covering 96 . 8–107 . 2 Mb of the 119 Mb reference , and modern samples covering 108 . 0–108 . 3 Mb ( S1 Table ) . We visualized the relationships between the sequenced historic and modern plants building a neighbor joining tree of all 123 samples and confirmed that the majority fell within an almost-identical clade , the HPG1 ( Fig 2A ) [23] . Because any degree of introgression from other non-HPG1 lineages would confound the discovery of new mutations downstream , we removed all divergent samples and built a neighbour joining tree ( n = 103 samples ) , which revealed that the HPG1 samples were very similar to each other , with very little within-population structure ( Fig 2B ) . A parsimony network was used to detect recombinant genomes within this HPG1 clade ( Fig 2C ) , which led us to remove three potential intra-lineage recombinants . Repeating the parsimony network cleared all previously inferred reticulations due to recombinations ( Fig 2D ) . After such stringent filtering , we kept 27 of the 35 herbarium samples , and 73 of the 87 modern samples ( S1 Table ) . These constitute a set of non-admixed , non-recombined and quasi-identical HPG1 individuals . Pairs of HPG1 herbarium genomes differed by 28–207 SNPs genome-wide , pairs of HPG1 modern genomes by 2–259 SNPs , and pairs of historic-modern HPG1 genomes by 56–244 SNPs . That is , whole-genome identity was at least 99 . 9997% in any pairwise comparison . Of the approximately five to six thousand segregating SNPs in the HPG1 population , the vast majority , about 95% ( Supplementary Text 3 ) , have not been reported outside of this lineage [21] . Importantly , the density of SNPs along the genome was low and evenly distributed ( typically fewer than 20 SNPs / 100 kb ) with no peaks of much higher frequency , which makes us confident that chunks of introgressions from other lineages do not exist in this putatively pure HPG1 set ( Fig 4 ) . For comparison , random pairs of A . thaliana accessions from the native range or pairs of non-HPG1 typically differ by about 500 SNPs / 100 kb [21] ( see scale in Fig 2A ) . There were no SNPs in mitochondrial nor chloroplast genomes , which already suggested a recent common origin , and genome-wide nuclear diversity ( π = 0 . 000002 , θW = 0 . 00001 , with 5 , 013 full informative segregating sites ) was two orders of magnitude lower than in the native range of the species ( θW = 0 . 007 ) [21] ( S1 Table ) ( Supplementary Text 6 ) . The population recombination parameter was also four orders of magnitude lower ( 4Ner = ρ = 3 . 0x10-6 cM bp-1 ) than in the native range ( ρ = 7 . 5x10-2 cM bp-1 ) [26] ( Supplementary Text 6 ) . While recombination occurs in every generation , regardless of self-fertilization or outcrossing , it is only observable after outcrossing between genetically non-identical individuals . We must stress that because A . thaliana can outcross at rates of several percent per generation [23 , 27] , but because the HPG1 population is genetically so homogeneous , we are mostly “blind” to the consequences of outcrossing in this special case . The lack of “observable recombination” in the genome is important , as it allows for the use of straightforward phylogenetic methods to calculate a mutation rate . The enrichment of low frequency variants in the site frequency spectrum ( Tajima’s D = -2 . 84; species mean = -2 . 04 , [21] ) and low levels of polymorphism are consistent with a recent bottleneck followed by population expansion , which usually generates star-like phylogenies ( Figs 2 and 3 ) . The obvious explanation is that the strong bottleneck corresponds to a colonization founder event , likely by few closely related individuals or perhaps even a single plant . Altogether these patterns indicate that the collection of HPG1 plants we investigated constitute a quasi-clonal and quasi-identical set of individual genomes , mostly devoid of observable recombination and population structure , and thus eminently suited for the study of naturally arising de novo mutations . It is important to distinguish between the mutation rate , which is the rate at which genomes change due to DNA damage , faulty repair , gene conversion and replication errors , and substitution rate , which is the rate at which mutations survive and accumulate under the influence of demographic processes and natural selection [28 , 29] . Under neutral evolution , mutation and substitution rates should be equal [29] . The simple evolutionary history of the HPG1 population enables direct estimates of substitution rates , and the comparison of theses between different genome annotations , as well as with mutation rates from controlled conditions experiments , could reveal the role played by both demographic and selective forces . To estimate the substitution rate in the HPG1 lineage , we used distance- and phylogeny-based methods that take advantage of the known collection dates ( Supplementary Text 7 ) . The distance method is independent of recombination and has been previously applied to viruses [30] and humans [31] . The substitution rate is calculated from correlation between differences in collection time in historic-modern sample pairs , and the number of nucleotide differences between those pairs relative to a reference ( Fig 3C ) , scaled to the size of the genome accessible to Illumina sequencing . This method resulted in an estimated rate of 2 . 11x10-9 substitutions site-1 year-1 ( 95% bootstrap Confidence Interval [CI]: 1 . 88–2 . 33x10-9 ) using rigorous SNP calling quality thresholds . Relaxing the thresholds for base calling and minimum genotyped rate affects both the number of called SNPs and the length of the interrogated reference sequence [32] . These largely cancelled each other out , and the adjusted estimates were relatively stable , between 2 . 1–3 . 2x10-9 substitutions site-1 year-1 ( S3 Table , Supplementary Text 3 ) . The second method , a Bayesian phylogenetic approach , uses the collection years for tip-calibration and assumes a relaxed molecular clock . It summarizes thousands of plausible coalescent trees , and it has been extensively used to calculate evolutionary rates in various organisms [33–35] . This method yielded a substitution rate of 4 . 0x10-9 , with confidence ranges overlapping the above estimates ( 95% Highest Posterior Probability Density [HPPD]: 3 . 2–4 . 7x10-9 ) . Based on the similar results obtained with two very different methods , we can confidently say that the substitution rate in the wild populations of HPG1 is between 2 and 5 x10-9 site-1 year-1 . To date the colonization of N . America by HPG1 A . thaliana and to improve the description of intra-HPG1 relationships compared to that from a NJ tree , we further used a Bayesian phylogeny . At first sight , the 73 modern samples appeared separated from the herbarium samples ( Fig 3B ) , but the superimposition of thousands of possible trees showed that the apparent separation of samples was less clear near the root ( Fig 3A ) . Long terminal branches reflected that the majority of the variants are singletons , typical of populations that expand after bottlenecks . The mean estimate of the last common HPG1 ancestor , the average tree root , was the year 1597 ( HPPD 95%: 1519–1660 ) ( Fig 3A and 3B ) , and an alternative non-phylogenetic method gave a similar estimate , 1625 . Both estimates are older than a previously suggested date in the 19th century , using a laboratory mutation rate estimate and having no information from herbarium samples [25] . Because HPG1 appears to have been the most abundant lineage in N . America since the 1860s , we believe it could have been one of the first , if not the first A . thaliana colonizer that could establish itself in N . America . If that is true , the time of coalescence of the HPG1 diversity could be close to the time of HPG1 introduction to N . America . During the colonial period , many European immigrants settled on the East coast , consistent with N . American A . thaliana lineages being genetically closest to British and coastal West European populations [21] . Coincidently , the oldest herbarium samples ( 12 out of the 27 ) were HPG1 and came from the East Coast , and we found a significant correlation between collection date and both latitude and longitude ( Fig 1C ) . This could indicate that after the colonization they moved from the East Coast to the Midwest—the other main area of the distribution that experienced an agricultural expansion in the 19th century [36] . Still , these conclusions need to be treated with caution , since regardless of the robustness of the results and our attempts to sample evenly from available collections , there could be unknown biases in the 19th century herbaria . Although for dating divergence events a substitution rate expressed in years is ideal , in order to compare substitution and mutation rates , both need to be expressed per generation . While A . thaliana is an annual plant , seed bank dynamics generate a delay of average generation time at the population scale . A comprehensive study of multiple A . thaliana populations in Scandinavia found that dormant seeds could wait for longer than a year in the seed bank , generating overlapping generations and an delayed average generation time of 1 . 3 years [37] with a notable variance across populations . Multiplication by the mean generation time led to an adjusted rate of 2 . 7x10-9 substitutions site-1 generation-1 ( 95% CI 2 . 4–3 . 0x10-9 ) ( Fig 3E ) . To be able to compare this rate with a reference , we also re-sequenced mutation accumulation ( MA ) lines in the Col-0 reference background grown under controlled conditions in the greenhouse that had been analyzed before with less advanced short read sequencing technology [38] . From the new re-sequencing data , we obtained an updated rate of 7 . 1x10-9 mutations site-1 generation-1 ( 95% CI 6 . 3–7 . 9x10-9 ) ( S2 and S3 Tables Supplementary Text 4 and 7 ) . This mutation rate is two- to three-fold higher than the per-generation substitution rate estimate in the wild , but within the same order of magnitude . The same holds for rates in different genome annotations , i . e . genic , intronic and intergenic regions , but the confidence intervals overlapped in many cases ( S3 Table ) . Differences in per-generation rates between laboratory and wild populations could stem from both methodological as well as biological causes . For instance , if the true average generation time was actually over 3 years / generation , the differences would cancel out ( Fig 3E ) . Limitations in mapping structural variation in non-reference samples could lower the substitution rate , which may explain why we calculated an atypically low substitution rate in regions with transposable elements ( see Supplementary Text 7 . 2 . 1 ) . Environmentally-driven effects that are not yet well understood , such as variable methylation status of cytosines , account for much of the variation in local substitution rates [39] , and could increase or decrease the rate ( see Supplementary Text 7 . 2 . 3 , S4 Fig ) . An alternative evolutionary explanation to the aforementioned laboratory and wild populations’ rates differences is that purifying selection in the wild would slow down the accumulation of mutations by removing deleterious mutations ( Fig 3E ) . This has been observed before and is one of the accepted causes of the discrepancy between the so called long- and short-term substitution rates in a range of organisms [40] . In order to provide evidence for negative purifying selection acting in the wild , we performed three types of analyses involving comparisons across genomic annotations within the HPG1 dataset . Firstly , by calculating contingency tables and computing a Fisher’s exact test , we compared the deviation of expected and observed SNPs between coding regions ( more likely under purifying selection ) , with intergenic regions , intronic regions , and all non-coding regions of genome . All three pairwise comparisons showed a depletion of coding SNPs and an enrichment of intergenic , intronic and non-coding SNPs ( odds ratio>2 , p<10−16 ) . An obvious explanation is that in genome annotations where a mutation is more likely to be deleterious , i . e . coding regions , the number of observed variants should be lower due to selection having removed them from the population before we could sequence them . Secondly , we studied the Site Frequency Spectrum ( SFS ) of genetic variants . The rationale was that because purifying natural selection is more efficient at removing intermediate-frequency variants , variants that tend to be deleterious or slightly deleterious should be found at lower frequency than those that only suffer neutral drift [41] . We built contingency tables of coding , intergenic , intronic and non-coding variants segregating above and below the conventional frequency cutoff of 5% to separate low- and intermediate-frequency variants [42] . We found that SNPs in coding regions were more likely to be at low frequency than those in intergenic ( odds ratio = 2 . 34 , p = 3 . 09x10-11 ) , intronic ( odds ratio = 1 . 48 , p = 0 . 02 ) , and all non-coding regions ( odds ratio = 2 . 05 , p = 1 . 29x10-8 ) . We carried out the same analysis using nonsynonymous and synonymous SNPs , which are easily interpretable in terms of the selection regimes under which they evolve . We did not find an enrichment ( p = 0 . 67 ) , perhaps due to an insufficient number of testable mutations ( S3 Table ) . Thirdly , to verify that the full frequency spectrum of coding SNPs was shifted to lower frequencies ( i . e . the results were not dependent on the arbitrary 5% frequency cutoff ) , we used the nonparametric Kolmogorov-Smirnov test for two samples . We found that the cumulative distribution of the site frequency spectrum ( CDSFS ) of coding regions is above ( i . e . , the frequency distribution is overall skewed to lower values ) both the intergenic CDSFS ( p = 3 . 25x10-6 ) and the non-coding regions CDSFS ( p = 0 . 001 ) , but not the intronic CDSFS ( p = 0 . 60 ) ( S5 Fig ) . As in our previous analysis , the comparison between the nonsynonymous and synonymous CDSFS yielded , likely for similar reasons , no differences ( p = 0 . 53 ) . All in all , these results support that purifying selection is a force shaping to some degree the diversity across the HPG1 genome and might therefore as well contribute to the differences between HPG1 and MA rates . Finally , having discovered over 5 , 000 de novo mutations in the HPG1 lineage , we wondered whether there is any evidence for an adaptive role of these de novo mutations in the colonization of N . America by HPG1 . We noted that some new mutations had risen to intermediate or even high frequencies in the HPG1 samples . This might have been the consequence of drift from stochastic demographic processes , or it could have been caused by positive natural selection . To find direct evidence for the latter , we grew the modern accessions in a common garden and studied phenotypes of known importance in ecology of invasions [43] , namely flowering time and root traits ( see Supplementary Text 8 ) . Using linear mixed models , we calculated the proportion of variance explained ( also called narrow sense heritability , h2 ) with a kinship matrix of all SNPs that had become common ( >5% , n = 391 ) . We found significant heritable variation for multiple traits including the growth rate in length ( h2 = 0 . 64 ) and the average root gravitropic direction ( h2 = 0 . 54 ) . As in our study mutations are the main source of genetic variants , these mutations—or mutations linked to them—should be responsible for significant quantitative variation in several traits ( S4 Table , Supplementary Text 10 ) . The existence of mutation-driven phenotypic variation at least indicates that natural selection could have acted upon such phenotypic variation . Although linkage disequilibrium ( LD ) among SNPs is high , the fact that HPG1 genomes differ in very few SNPs greatly reduces the list of candidate loci that might generate the observed phenotypic variation ( S7 Fig ) [44] . With this reasoning in mind and understanding the limitations imposed by LD , we carried out a genome-wide association ( GWA ) analysis and found 79 SNPs associated with one or more root traits , mostly growth and directionality ( Fig 4 ) . Twelve SNPs were in coding regions and seven resulted in nonsynonymous changes—some producing non-conservative amino-acid changes and thus likely to affect protein structure and/or function ( Table 1 , based on transition scores from [45] ) . Due to the aforementioned LD , in some cases the results of associations could not be confidently assigned to a specific SNP and thus we report the number of other associated mutations with r2 > 0 . 5 ( Table 1 , S7 Fig ) . We note that linked genetic variation that has gone undetected ( e . g . , structural variation ) could be causal rather than the identified SNPs . For some cases , however , we were able to pinpoint clear candidates that were not in LD with other SNPs and whose functional annotation had a strong connection to the phenotype ( Table 1 , S7 Fig ) . For example , one SNP associated with root gravitropism was not linked to any other SNP hit and it was found at 40% frequency ( top 3% percentile ) . This SNP produces a cysteine to tryptophan change in AT5G19330 , which is involved in abscisic acid response , strongly expressed in growing roots , and confers salt tolerance when overexpressed [46] . Another nonsynonymous SNP associated with root growth is located in AT2G38910 , which encodes a calcium-dependent kinase that is a factor regulating root hydraulic conductivity and phytohormone response in vitro [47 , 48] . Nineteen other SNPs were associated with climate variables after correction for latitude and longitude ( www . worldclim . org , S4 Table ) , and generally tended to coincide with top root-associated SNPs ( odds ratio = 3 . 9 , Fisher’s Exact test p = 0 . 002; Fig 4 , and S5 Table ) . Specifically , this means that alleles increasing root length and gravitropic growth were present in areas with lower precipitation , and vice versa ( Pearson’s correlation r = 0 . 85 , p = 0 . 003 ) . This indicates that phenotypic variation generated by mutations coincides with environmental ( and not geographic ) gradients along the colonized areas . Compared to other mutations with matched allele frequencies , root-associated mutations are first found in older herbarium samples nearer to Lake Michigan ( S6 Fig ) , the area in N . America that seems to be most densely populated by A . thaliana [21] . A more densely spaced time series of samples would be needed to confirm the older age of specific mutations , but our observation could be explained by spatially varying selection across N . America , which may maintain antagonistic pleiotropic mutations for longer time than neutral mutations . The association of putatively adaptive mutation with climate variables could also be explained by such a phenomenon . Nevertheless , to confirm hypotheses of local adaptation by de novo mutations , it will be necessary to grow collections of divergent HPG1 individuals in multiple contrasting N . American locations over several years . Ideally , one would revive historical specimens to compare their performance to modern populations [49] . All in all , our results are compatible with natural positive selection having already acted on root morphology variation that was generated by de novo mutations in this colonizing lineage . In summary , we have exploited whole-genome information from historic and contemporary collections of a herbaceous plant to empirically characterize evolutionary forces during a recent colonization . With this natural time series experiment we could directly estimate the nuclear substitution rate in wild A . thaliana populations—a parameter difficult to characterize experimentally [9] . This allowed us to date the colonization time and spread of HPG1 in N . America . We provide evidence that purifying selection has already changed the site frequency spectrum in the course of just a few centuries . Finally , we discovered that a small number of de novo mutations that rose to intermediate frequency can together explain quantitative variation in root traits across environments . This strengthens the hypothesis that some de novo variation could have had an adaptive value during the colonization and expansion process , a hypothesis that has been put forward as one of the possible solutions to the genetic paradox of invasion in plants [17] . This process might be more relevant in self-fertilizing plants , which typically have less diversity than outcrossing ones [50] , but have higher growth rates [43] and account for the majority of successful plant colonizers [5] . While A . thaliana HPG1 is not an invasive , harmful species , it can teach us about fundamental evolutionary processes behind successful colonizations and adaptation to new environments . Our work should encourage others to search for similar natural experiments and to unlock the potential of herbarium specimens to study “evolution in action” .
Modern A . thaliana accessions were from the collection described by Platt and colleagues [23] , who identified HPG1 candidates based on 149 genome-wide SNPs ( S1 Table , S1 Text ) . Herbarium specimens were directly sampled by Max Planck colleagues Jane Devos and Gautam Shirsekar , or sent to us by collection curators from various herbaria ( S1 Table , S1 Text ) . Among the substantial number of specimens in the herbaria of the University of Connecticut , the Chicago Field Museum and the New York Botanical Garden , we selected herbarium specimens spaced in time so there was at least one sample per decade starting from the oldest record ( 1863 ) . The differences in geographic biases of herbarium and modern collections are difficult to know [2] , thus we did choose both historic and modern samples that were as regularly distributed in space as possible , and sample overlapping locations wherever possible . DNA from herbarium specimens was extracted as described [51] in a clean room facility at the University of Tübingen . Two sequencing libraries with sample-specific barcodes were prepared following established protocols , with and without repair of deaminated sites using uracil-DNA glycosylase and endonuclease VIII ( refs . [52–54] ) ( S2 Text ) . The reads of repaired libraries are available at https://www . ebi . ac . uk/ena/data/view/PRJEB24619 . We also investigated patterns of DNA fragmentation and damage typical of ancient DNA [24] ( S2 Text ) . DNA from modern individuals was extracted from pools of eight siblings using the DNeasy plant mini kit ( Qiagen , Hilgendorf , Germany ) . Genomic DNA libraries were prepared using the TruSeq DNA Sample or TruSeq Nano DNA sample prep kits ( Illumina , San Diego , CA ) , and sequenced on Illumina HiSeq 2000 , HiSeq 2500 or MiSeq instruments . Paired-end reads from modern samples were trimmed and quality filtered before mapping using the SHORE pipeline v0 . 9 . 0 [25 , 55] . Because ancient DNA fragments are short ( S1 Fig ) we merged forward and reverse reads for herbarium samples after trimming , requiring a minimum of 11 bp overlap [51] , and treated the resulting as single-end reads . Reads were mapped with GenomeMapper v0 . 4 . 5s [56] against an HPG1 pseudo-reference genome [25] , and against the Col-0 reference genome , and SNPs were called with SHORE for the HPG1 pseudo-reference genome mappings [25 , 57] using different thresholds ( Supplementary Text 3 ) . Average coverage depth , number of covered genome positions , and number of SNPs identified per accession relative to HPG1 are reported in S1 Table . We also re-sequenced the genomes of twelve Col-0 MA lines [57 , 58] ( S2 Table ) ( Supplementary text 4 ) to recalculate and update the laboratory mutation rate from Ossowski et al . [38] with the newer sequencing technologies . We used the Pegas , Ape and Adegenet packages in R [59–61] to manipulate and visualize the genetic distances of all samples as well as the HPG1 subset ( Supplementary Text 7 ) . We constructed parsimony networks using SplitsTree v . 4 . 12 . 3 [62] , with confidence values calculated with 1 , 000 bootstrap iterations . We built Maximum Clade Credibility Trees using the Bayesian phylogenetic tools implemented in BEAST v . 1 . 8 [63] ( see below ) . Transforming the variant sites into a FASTA format , we estimated genetic diversity as Watterson´s θ [64] and nucleotide diversity π , and the difference between these two statistics as Tajimas’s D [65] using DnaSP v5 [66] . Then we re-scaled the estimates using the sequencing-accessible genome sizes ( S3 Table ) . We estimated pairwise linkage disequilibrium ( LD ) between all possible combinations of informative sites , ignoring singletons , by computing r2 , D and D’ statistics using DnaSP v5 [66] . For the modern individuals , we calculated the recombination parameter rho ( 4Ner ) also using DnaSP v5 [66] . Similarly as in Fu et al . [67] , we used genome-wide nuclear SNPs to calculate pairwise “net” genetic distances using the equation D'ij = Dic-Djc , where D'ij is the net distance between a modern sample i and a herbarium sample j; Dic the distance between the modern sample i and the reference genome c; and Djc is the distance between a modern sample ( j ) and the reference genome ( c ) . We calculated a pairwise time distance in years between the collection times , T'ij , and calculated the linear regression: D' = a+bT' . The slope coefficient b describes the number of substitution changes per year . We used either all SNPs or subsets of SNPs at different annotations ( genic , intergenic etc . ) appropriately scaled by accessible genome length . Because the points used to calculate the regression are non-independent , a bootstrap has been recommended to overcome to a certain extent the anti-conservative confidence intervals [30] ( Supplementary Text 7 and S3 Fig ) . To fully account for the non-independence of points , we need to work with phylogenies . The Bayesian phylogenetics approach we used is implemented in BEAST v1 . 8 [63] and is called tip-calibration , and calculates a substitution rate along the phylogeny . Our analysis optimized simultaneously and in an iterative fashion using a Monte Carlo Markov Chain ( MCMC ) a tree topology , branch length , substitution rate , and a demographic Skygrid model ( Supplementary Text 7 ) . The demographic model is a Bayesian nonparametric one that is optimized for multiple loci and that allows for complex demographic trajectories by estimating population sizes in time bins across the tree based on the number of coalescent—branching—events per bin [68] . We also performed a second analysis run using a fixed prior for substitution rate of 3x10-9 substitutions site-1 year-1 based on our previous net distance estimate to confirm that the MCMC had the same parameter convergence , e . g . tree topology , as in the first “estimate-all-parameters” run . Having a substitution rate per year we can estimate the time to the most common recent ancestor L solving d = 2L x μ where d is the average pairwise genetic distance between our samples and μ is the calculated substitution rate from the distance method . This yielded 363 years , which subtracted to the average collection date of the samples , produced a point estimate of 1615 . We compare this estimate with the inferred phylogeny root from the BEAST analysis . We separately analyzed sequences at different annotations , since as they might be under different selection regimes ( i . e . evolutionary constraints ) . We computed , using the HPG1 dataset , one-tailed Fisher’s exact test using the base stats package in R [69] on contingency tables of the total number of base pairs against the number of SNPs , and those separated by positions being annotated as a coding against non-coding ( intergenic , intronic , all other noncoding ) . The test returned whether coding regions have a lower number of SNPs than other reference annotation ( intronic , interenic , all non-coding regions ) , as expected by the total number of positions in the genome annotated as such . We also constructed contingency tables to test whether SNPs annotated as coding compared to those annotated as non-coding were more likely to be found at low ( <5% ) or intermediate ( 5≥% ) frequency . Finally , we calculated the unfolded Site Frequency Spectrum ( SFS ) based on the order of appearance of genetic variants in the herbarium dataset . We then used the Kolmogorov–Smirnov two-samples test and 10 , 000 bootstrap resampling using the R package Matching v . 4 . 9–2 ( ref . [70] ) to calculate whether the frequency spectrum was lower for coding SNPs than for other SNPs . Additionally , we also repeated these analyses comparing nonsynonymous and synonymous mutations instead of coding and non-coding regions . We collected flowering , seed and root morphology phenotypes for 63 accessions ( Supplementary Text 8 ) . For associations with climate parameters , we followed a similar rationale as previously described [71] . We extracted information from the bioclim database ( http://www . worldclim . org/bioclim ) at a 2 . 5 degrees resolution raster and intersected it with geographic locations of HPG1 samples ( n = 100 ) . We performed association analyses under several models and p-value corrections using the R package GeneABEL [72] ( Supplementary Text 8 . 2 ) . To calculate the variance of the trait explained by all genetic variants , we used a linear mixed model: y = Zu + ε; where y is the phenotype or climate variable , Z is the design matrix of genome identities , u is the random genome background effect informed by the kinship matrix and distributed as MVN ( 0 , σgA ) , and ε is the random error term . The ratio of σg / σT is commonly called narrow sense heritability , “chip” heritability , or proportion of variance explained by genotype [73] . Only SNPs with MAF>5% ( n = 391 ) were used to build a kinship or relationship matrix A . Note that the differences between any two genotypes were of the order of one or few dozens of SNPs . While this approach is appropriate to calculate a chip heritability , it would not be very useful to detect significant SNP , as the random factor accumulates all the available variation ( S4 Table ) . We therefore run a regular GWA model without kinship matrix: y = Xb + ε; where X corresponds to the genotype states at a given SNP , and b is the fixed phenotypic effect of the SNP . To evaluate significance , we generated a p-value empirical null distribution based on running such model over 1 , 000 permuted datasets , which lead to conservative associations ( S7 Fig , Data Appendix S1 ) . The p-values from running the association in the real data that were below the 5% tail in the empirical distribution could be considered significant . However , we also established a conservative “double” Bonferroni correction , where the significant threshold was lowered to 0 . 01% ( = 5% / [number of SNPs + number of phenotypes tested] ) . All significant SNPs are shown in S5 Table , and a subset in Table 1 . Although many phenotypic traits did not have significant SNPs , we show all the QQ plots in the S2 Text .
|
A consequence of an increasingly interconnected world is the spread of species outside their native range—a phenomenon with potentially dramatic impacts on ecosystem services . Using population genomics , we can robustly infer dynamics of colonization and successful population establishment . We have compared hundred genomes of a single Arabidopsis thaliana lineage in North America , including genomes of contemporary individuals as well as 19th century herbarium specimens . These differ by an average of about 200 mutations , and calculation of the nuclear evolutionary rate enabled the dating of the initial colonization event to about 400 years ago . We also found mutations associated with differences in traits among modern individuals , suggesting a role of new mutations in recent adaptive evolution .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Methods"
] |
[
"biotechnology",
"taxonomy",
"population",
"genetics",
"brassica",
"mutation",
"phylogenetics",
"data",
"management",
"model",
"organisms",
"substitution",
"mutation",
"plant",
"science",
"methylation",
"phylogenetic",
"analysis",
"experimental",
"organism",
"systems",
"genome",
"analysis",
"plant",
"genomics",
"genome",
"annotation",
"population",
"biology",
"plants",
"arabidopsis",
"thaliana",
"research",
"and",
"analysis",
"methods",
"genomic",
"libraries",
"computer",
"and",
"information",
"sciences",
"plant",
"genetics",
"chemistry",
"evolutionary",
"systematics",
"eukaryota",
"plant",
"and",
"algal",
"models",
"genetics",
"biology",
"and",
"life",
"sciences",
"chemical",
"reactions",
"physical",
"sciences",
"genomics",
"evolutionary",
"biology",
"plant",
"biotechnology",
"computational",
"biology",
"organisms"
] |
2018
|
The rate and potential relevance of new mutations in a colonizing plant lineage
|
Prions propagate as multiple strains in a wide variety of mammalian species . The detection of all such strains by a single ultrasensitive assay such as Real Time Quaking-induced Conversion ( RT-QuIC ) would facilitate prion disease diagnosis , surveillance and research . Previous studies have shown that bank voles , and transgenic mice expressing bank vole prion protein , are susceptible to most , if not all , types of prions . Here we show that bacterially expressed recombinant bank vole prion protein ( residues 23-230 ) is an effective substrate for the sensitive RT-QuIC detection of all of the different prion types that we have tested so far – a total of 28 from humans , cattle , sheep , cervids and rodents , including several that have previously been undetectable by RT-QuIC or Protein Misfolding Cyclic Amplification . Furthermore , comparison of the relative abilities of different prions to seed positive RT-QuIC reactions with bank vole and not other recombinant prion proteins allowed discrimination of prion strains such as classical and atypical L-type bovine spongiform encephalopathy , classical and atypical Nor98 scrapie in sheep , and sporadic and variant Creutzfeldt-Jakob disease in humans . Comparison of protease-resistant RT-QuIC conversion products also aided strain discrimination and suggested the existence of several distinct classes of prion templates among the many strains tested .
Prion diseases , or transmissible spongiform encephalopathies , are neurodegenerative disorders that include Creutzfeldt-Jakob disease ( CJD ) , Gerstmann-Straussler-Scheincker syndrome ( GSS ) , fatal familial insomnia ( FFI ) and sporadic fatal insomnia ( sFI ) in humans , bovine spongiform encephalopathy ( BSE ) in cattle , scrapie in sheep , and chronic wasting disease ( CWD ) in cervids . The origin of prion diseases can be infectious , genetic or sporadic . Many prion diseases also have subtypes or strains that can be distinguished based on the PRNP ( prion protein ) genotype , transmission characteristics , clinical manifestations , neuropathological lesion profiles and/or biochemical properties of the disease-associated forms of prion protein ( PrPD ) [1–9] . PrPD , or a subset thereof , is the predominant molecular component of the infectious agent , or prion , which propagates itself by inducing misfolding of the hosts’ normal protease-sensitive prion protein , PrPC or PrPSen , into additional PrPD . This propagation mechanism appears to involve seeded , or templated , polymerization in which the given PrPD conformation is imposed upon normally monomeric PrPSen molecules as they are recruited into growing PrPD multimers [3 , 4] . PrPD usually includes forms called PrPRes that , unlike the normal PrPSen , are partially resistant to digestion by proteinase K ( PK ) . The banding pattern of PrPRes in immunoblots can vary distinctively depending on the prion strain , host species and/or PRNP genotype . With most prion diseases the predominant 21-32-kDa variably glycosylated PrPRes fragments observed on immunoblots extend from ragged N-termini between residues ~80–96 to the GPI-anchored C-terminus ( e . g . , at residue 231 in humans ) . In contrast , the PrPRes associated with sheep Nor98 scrapie and human GSS linked to the P102L , F198S , A117V and H187R PRNP mutations include much smaller 6–14 kDa bands [10–12] . These bands are internal fragments with ragged N-and C-termini within residues ~80-~160 [13] . In cases of P102L GSS , brain tissue from some individuals can also give 21–32 kDa PrPRes bands with the 7–8 kDa bands , while others give the 21–32 kDa PrPRes bands but lack the 7–8 kDa bands . Hereafter , we will refer to the former cases as GSS P102L* and the latter as GSS P102L . A major challenge for the prion disease field is the development of sufficiently practical and sensitive tests for routine prion disease detection and strain discrimination in medicine , agriculture , wildlife management and research . The Real Time Quaking Induced Conversion ( RT-QuIC ) assay , which is based on prion-seeded fibrillization of recombinant prion protein ( rPrPSen ) , is known to be highly specific and ultra-sensitive for detection of multiple human and animal prion diseases [14–20] . Moreover , like the amyloid seeding assay [21] , RT-QuIC is more practical than comparably ultra-sensitive assays by being relatively rapid and based on a 96-well plate format with fluorescence readout [14 , 16] . Appropriate combinations of prion type and rPrPSen substrate have been important in the performance of various RT-QuIC assays [14 , 18–20 , 22–25] . For several types of prion disease , however , no effective rPrPSen substrate has been identified; these types include human GSS arising from P102L* , F198S , A117V and H187R PRNP mutations and the atypical sheep scrapie strain Nor98 . Moreover , no single substrate has yet been shown to detect all of the different prion variants of humans , cattle , sheep , cervids and rodents . One potential rPrPSen substrate that has not been described for RT-QuIC assays is based on the bank vole sequence . Bank voles [26] , and transgenic mice that express bank vole ( BV ) PrPSen [27] , are susceptible to an unusually wide range of prion strains from different species . Furthermore , PrPSen in bank vole brain tissue homogenates is a broadly reactive , but not universal , substrate for the highly sensitive protein misfolding cyclic amplification ( PMCA ) assay for prions [28] . Here we have tested the suitability of recombinant bank vole PrPSen ( BV rPrPSen ) , when expressed in E . coli and purified , as an RT-QuIC substrate . We have found so far that BV rPrPSen is a universally effective substrate for multiple prion strains from multiple species , and , most notably , for prions for which no effective substrate has been available . Furthermore , we have found that BV rPrPSen-based RT-QuIC reactions give strain-dependent PK-resistant products in a manner that should further facilitate prion strain discrimination .
Most mammalian PrPRes types with predominant 21–32 kDa PrPRes bands can seed Thioflavin T-positive ( ThT ) amyloid formation in RT-QuIC reactions [17 , 22 , 29–32] using at least one of the following substrates: Syrian golden hamster rPrPSen residues 90–231 [14 , 30 , 31] , Syrian golden hamster rPrPSen 23–231 [18] , human rPrPSen 23–231 [16] , murine rPrPSen 23–231 [24] or hamster-sheep chimeric rPrPSen 23–231 [19 , 22] . For example , detection of human P102L GSS brain tissue using hamster rPrPSen 90–231 is shown in Fig 1 . However , to date , no detection of RT-QuIC seeding activity has been reported using these rPrPSen substrates with cases of human GSS or sheep scrapie that give prominent low molecular weight PrPRes fragments in immunoblots , despite extensive efforts . Specifically , these cases include human GSS with the F198S , A117V or H187R mutations and sheep Nor98 scrapie types giving 6–14 kDa PrPRes fragments [10–13] , and the human P102L-GSS with an ~8 kDa PrPRes fragment ( P102L* ) [10] . Our inability to detect these prion types is exemplified in Fig 1 using 10-3 brain tissue dilutions of human GSS F198S and P102L* and sheep Nor98 scrapie with the hamster rPrPSen 90–231 . In contrast , P102L GSS ( without the ~8-kDa fragment ) gave positive reactions with 1 , 000 , 000-fold smaller amounts of brain tissue . We then tested bank vole rPrPSen residues 23–230 ( BV rPrPSen ) as a substrate to detect seeding activity of two human prion subtypes that have not been detectable previously by RT-QuIC , namely F198S- and A117V-GSS . Concurrently , we varied two parameters that we have shown to be influential , namely the concentrations of NaCl [14] and Sodium Dodecyl Sulfate ( SDS ) [30] . Each reaction was seeded with 10-4 dilutions of frontal cortex brain tissue from confirmed GSS cases carrying either the F198S or A117V mutation of the prion protein gene . We found that our standard concentrations of SDS ( 0 . 002% ) in combination with either 130 or 300mM NaCl failed to allow a distinction in lag phase between prion positive and uninfected brain homogenate ( BH ) seeded reactions ( Fig 2A and 2B ) . Lowering the SDS concentration to 0 . 001% with either 130 or 300mM NaCl improved this distinction between prion positive and uninfected BH seeded reactions ( Fig 2C and 2D ) . However , using final concentrations of 300 mM NaCl and 0 . 001% SDS , provided much shorter lag phases in reactions seeded with the two GSS subtypes than with the cerebral ischemia negative control ( Fig 2D ) . These results indicated that under these latter RT-QuIC conditions BV rPrPSen detected seeding activity associated with PrPD conformers that had not otherwise been detectable by RT-QuIC or PMCA prion seed amplification techniques [33] . Next , we assessed the sensitivity of this new RT-QuIC assay for detecting GSS-associated prion seeding activity . Reactions were seeded with 10-4 to 10-9 dilutions of brain tissue from GSS patients carrying the P102L , P102L* , A117V , F198S and H187R mutation of the prion gene ( Fig 3A–3E ) . A reaction time cutoff of 50h was chosen because in more than 20 independent RT-QuIC experiments seeded with negative control Alzheimer’s disease ( AD ) or cerebral ischemia brain homogenates , no positive RT-QuIC reactions were observed until after 55h ( in rare wells ) . We detected GSS P102L , P102L* , A117V and F198S and H187R prion seeding activity in as little as 10-9 , 10-4 , 10-8 , 10-8 and 10-6 dilutions of brain ( frontal cortex ) tissue dilutions , respectively ( Fig 3 ) . After finding that BV rPrPSen supported RT-QuIC detection of prion seeding activity from previously undetectable types of GSS , we tested whether BV rPrPSen could be used to detect other types of prion diseases . We tested 28 different types of prions in brain tissue from humans , sheep , mouse , hamster , cattle , elk , and deer ( Tables 1 and 2 ) and found that all of them gave faster and stronger positive ThT fluorescence responses than a variety of uninfected negative control brain specimens ( Fig 4 ) . Among the 28 were the five prion types that have not been detectable by RT-QuIC under other conditions , namely human GSS F198S , A117V , H187R , and P102L* and sheep Nor98 scrapie ( Fig 4 , red traces ) . These results indicated that under these conditions , BV rPrPSen is the most broadly prion-seeded RT-QuIC substrate described to date . Prion strain-dependence has not been observed previously in the immunoblot banding profile of PK-treated recombinant PrPRes ( rPrPRes ) products of RT-QuIC reactions . However , using BV rPrPSen we observed consistently distinct products of RT-QuIC reactions seeded with different types of human prions ( Fig 5 and Table 1 ) . The observed banding patterns could be grouped based on the type of seed: GSS cases ( F198S , A117V , H187R ) with the ~8–14 Kda protease-resistant bands and sFI gave 2 bands: a major ~10kDa band and a ~6-9kDa band; the GSS ( P102L ) , gCJD ( E200K , V210I , six octarepeat insertion ) , and the iatrogenic CJD ( iCJD ) cases with ~21–32 kDa PrPRes bands gave multiple bands with a major ~12 Kda band and multiple minor bands between ~6–10 kDa; variant CJD , GSS ( P102L* ) and FFI ( D178N ) cases gave a single predominant band at ~10 kDa; and sporadic CJD in some cases gave two bands between ~10–12 kDa , while in other cases gave a predominant band at ~ 10 kDa . Repeated analyses ( >4 ) of individual sCJD cases indicated that they consistently seeded the formation of only one or the other of the latter two rPrPRes products . This observation provided evidence that the different sCJD-seeded rPrPRes products were dictated by differential templating activity in the tissue samples rather than stochastic events during the RT-QuIC reaction . Additionally , because these immunoblots used an antiserum to the C-terminus of PrP , the fragments likely differed primarily at their N-termini . We further compared the BV rPrPRes products of reactions seeded with different rodent , bovine , cervine and ovine prion strains ( Table 2 ) . As with the human prion seeds , we observed distinct strain-dependent BV rPrPRes banding profiles from reactions seeded with different prion types . Mouse 22L scrapie-seeded BV rPrPRes products consistently showed a ~10 and ~12 kDa PK-resistant band , whereas BV rPrPRes products from reactions seeded with Chandler , ME7 , 87V and anchorless 22L ( 22L GPI- ) scrapie displayed a predominant ~10 kDa band ( Fig 6A ) . The lack of the GPI anchor in the 22L GPI- scrapie seed resulted in an RT-QuIC product that was distinct from the wild-type GPI-anchored 22L scrapie . Additionally , closely related hamster prion strains ( Hyper and 263K; Fig 6B ) showed similar BV rPrPRes banding profiles ( ~10 and ~12 kDa PK-resistant bands ) which were distinct from the Drowsy-seeded BV rPrPRes products ( primarily a ~10 kDa band; Fig 6B ) . Deer and elk CWD-seeded reactions each gave ~8 , 9 , 10 , and 12 kDa bands , but differed in the relative intensities of the top two bands between the two ( Fig 6C ) . Furthermore , distinct strain-dependent BV rPrPRes banding profiles were observed between classical ( C-BSE ) and atypical ( L-BSE ) ( ~10 kDa vs . ~9 , 10 , and 12 kDa bands , respectively; Fig 6C ) , as well as between classical and atypical Nor98 sheep scrapie ( ~10 kDa vs . ~9 , 10 , and 12 kDa bands , respectively; Fig 6D ) . Collectively , these immunoblotting results suggested that certain human and animal prion diseases can be discriminated in part based on analysis of the rPrPRes products of BV rPrPSen-based RT-QuIC reactions . We previously reported that classical and atypical L-type BSE strains can be discriminated on the basis of relative RT-QuIC reactivities with hamster rPrPSen 90–231 and hamster-sheep chimeric rPrPSen 23–231 substrates [22] . Here we found that BV rPrPSen can similarly detect both classical and L-type BSE , providing an alternative substrate for discrimination between the two bovine strains . Specifically , detection of seeding activity with BV rPrPSen ( Fig 7 ) but not with three other rPrPSen substrates that detected only L-type BSE , namely human 23–231 , hamster 23–231 or hamster 90–231 [22] , can be used to differentiate these two bovine prion types . Having detected Nor98 sheep scrapie with BV rPrPSen , ( Fig 4 ) we tested whether a strategy similar to the one described above for C- vs . L-type BSE using different rPrPSen substrates would allow discrimination of Nor98 and classical sheep scrapie . Brain tissue from eight sheep with classical scrapie [ARQ/ARQ ( n = 6 ) , VRQ/VRQ ( n = 2 ) PrP genotypes , Table 2] were readily detected using the hamster-sheep chimeric rPrPSen 23–231 within ~40hs ( Fig 8A ) . However , brain tissue from eight cases of Nor98 scrapie [ARR/AHQ ( n = 1 ) , ARQ/ARQ ( n = 4 ) , ARQ/AHQ ( n = 2 ) and ARR/ARR ( n = 1 ) genotypes , Table 2] gave no positive responses using the same substrate ( Fig 8B ) . In contrast , consistent with the results in Fig 4 , seven of these cases gave positive responses when BV rPrPSen was used in reactions seeded with 10-4 brain tissue dilutions ( Fig 8A–8E , orange lines ) and those that were weaker or not detected were positive when seeded with 10-3 dilutions ( Fig 8A–8E , red lines ) . To compare the sensitivities of the assay for detection of classical and atypical scrapie using these two substrates , we diluted representative brain homogenates from classical and Nor98 scrapie positive sheep ( Fig 8G–8J ) and tested them using both BV and Ha-S rPrPSen . We detected classical scrapie down to 10-8 dilutions using Ha-S rPrPSen and down to 10-6 using BV rPrPSen . Consistent with the data in Fig 8 , no fluorescence increases were seen in reactions seeded with the same dilutions of a Nor98 atypical scrapie sample when using Ha-S rPrPSen . In contrast , parallel reactions with BV rPrPSen gave positive reactions when seeded with Nor98 brain dilutions down to 10-6–10-7 , indicating that BV rPrPSen is ~1 , 000-fold more sensitive at detecting Nor98 scrapie than is Ha-S rPrPSen . Collectively , these results suggest that if an ovine brain sample gives a positive RT-QuIC response with BV rPrPSen , it should give a stronger positive reaction with Ha-S rPrPSen if it contains classical scrapie , but a negative , or at least much weaker , reaction if it contains Nor98 scrapie . To investigate the discrimination of two non-genetic human prion strains , we tested 10-4 brain tissue dilutions from two confirmed cases of Type 1 sCJD ( Fig 9A and 9B , Cases a and b , green lines ) and two cases of vCJD ( Fig 9A and 9B , Cases c and d , orange lines ) . We used previously described SDS conditions ( 0 . 002% final concentration of SDS; [23] ) with hamster 23–231 rPrPSen , and 0 . 001% SDS with BV rPrPSen , both in the presence of 300mM NaCl . We observed rapid amplification of prion seeding activity in the two Type 1 sCJD samples when using either hamster 23–231 or BV rPrPSen ( Fig 9A and 9B ) . Our detection of the sCJD samples with the hamster 23–231 substrate was consistent with previous demonstrations that all sCJD subtypes are detectable with this substrate [23 , 30 , 34] . No increase in ThT fluorescence was seen in vCJD-seeded hamster 23–231 rPrPSen RT-QuIC reactions ( Fig 9A ) . However , in accordance with the results shown in Fig 4 , seeding activity was detected in both vCJD samples using BV rPrPSen ( Fig 9B ) . Thus , sporadic and variant CJD sample were discriminated by differential reactivities with the BV and hamster 23–231 rPrPSen substrates . Next we compared the RT-QuIC sensitivities for detection of sCJD and vCJD brain homogenates using hamster and BV rPrPSen . We performed end-point dilution RT-QuIC analysis of brain tissue from sCJD ( Case a ) and vCJD ( Case c ) ( Fig 9C–9F , green and orange lines , respectively ) . We detected sCJD down to 10-8/10-9 with hamster 23–231 rPrPSen , ( Fig 9C ) and 10-8 with BV rPrPSen ( Fig 9D ) . Although sCJD gave slightly slower amplification kinetics with hamster 23–231 rPrPSen ( Fig 9C ) compared to BV rPrPSen ( Fig 9D ) , the overall sensitivities using the two substrates were comparable . In contrast , markedly different sensitivities were observed with the two substrates in the vCJD-seeded reactions . Specifically , only weak seeding activity was occasionally detected in 10-4 or 10-5 brain dilutions with hamster 23–231 rPrPSen ( representative data in Fig 9E ) , but fast and sensitive detection of vCJD seeding activity down to 10-7 brain tissue dilution was observed using BV rPrPSen ( Fig 9F ) . These results suggest that BV rPrPSen is 100–10 , 000-fold more sensitive than hamster 23–231 rPrPSen in detecting vCJD brain derived prion seeding activity . Collectively , these findings further support the potential broad applicability of a BV rPrPSen prion discrimination strategy to a variety of prion types .
The lack of practical and cost-effective tests that are sensitive enough to detect the lowest infectious levels of prions has long been a major impediment in coping with prion diseases . Rapid commercially available immunoassays have allowed post-mortem detection of prion infections in high-titered tissues such as brain or lymphoid tissues , but diagnostic specimens that are most readily accessible in living hosts , such as blood , CSF and nasal brushings , have much lower prion titers that are undetectable with these assays . In contrast , RT-QuIC assays have been highly effective in detecting prion seeding activity in such low-titered specimens , and are being widely implemented as state-of-the-art diagnostic tests for humans and animals [16 , 18–20 , 25 , 34–37] . Moreover , recent improvements have increased the speed and sensitivity of RT-QuIC assays such that sCJD testing based on human CSF samples can now be performed in a matter of hours rather than days [30] . In our experience , the most demanding and costly requirement for RT-QuIC testing is the availability of suitable rPrPSen substrates . Prior to the present study , testing facilities would typically have to produce or procure multiple rPrPSen sequences to be able to test for multiple prion types . However , we have now shown that all of the prion diseases that we have tested so far from humans and other mammals can be detected sensitively by using BV rPrPSen ( Fig 4 ) . This provides a useful platform for broad-based prion detection and strain discrimination . Thus , we envision that most initial screening for the presence of a wide variety of prions could be performed using BV rPrPSen . Once a prion-infected sample from a given host species is identified , one could then often discriminate between strains by targeted use of another rPrPSen substrate that is known to be differentially sensitive to seeding by prion strains of that host species ( Figs 7–9 ) and/or by performing immunoblots of the PK-resistant RT-QuIC products of the reactions ( Figs 5 and 6 ) . Although we have demonstrated detection of a wide variety of prion types , the relative sensitivities of BV rPrPSen-based RT-QuIC for brain homogenates of hosts with different prion diseases is presumably dependent on the concentrations of PrPD in the tissue samples . Clearly PrPD concentrations may vary markedly between individuals and different regions of the brain as a function of strain . Furthermore , because PrPD can vary markedly in its properties , e . g . amyloid vs . non-amyloid , protease-sensitive vs . resistant , small vs . large particles , infectious vs . non-infectious , it is probable that the RT-QuIC seeding activity will vary per unit PrPD between different prion strains and tissue sources . Thus , although we have shown the potential for BV rPrPSen-based RT-QuIC to detect and help discriminate prion strains , much additional work with each type of prion and sample type will be required to better establish the quantitative relationships between RT-QuIC seeding activity and the levels of various types of PrPD in different tissues of diagnostic or scientific interest . Since the inception of prion-seeded cell-free PrP conversion reactions [38] , striking sequence- and strain-specificities have been observed that appeared to correlate , at least largely , with transmission barriers and strain phenotypes of prion diseases in vivo [3 , 39–41] . Indeed , sequence differences of as little as a single residue between the PrPD seed and PrPSen substrate can block PrPRes formation in such cell-free reactions [42] , as it can in scrapie-infected cells [43] and in vivo [44] . However , RT-QuIC assays have tended to be less constrained by such sequence differences [14] . We reason that this is due in part to the fact that in RT-QuIC reactions , it is only the C-terminal residues ~160–231 of the substrate molecules that must refold into the PK-resistant amyloid core [45] to give a positive reaction , i . e . , an increase in ThT fluorescence . In contrast , earlier cell-free conversion [38 , 46 , 47] and PMCA reactions [48] have used the immunoblot-based detection of much larger PK-resistant cores , typically comprised of residues ~90–231 , as a positive readout . Thus , much more extensive packing of more N-proximal residues is required in the latter reactions , as it is in vivo , giving more opportunities for sequence differences between seed and substrate to influence conversion . Nonetheless , despite the lower sequence specificity of RT-QuIC reactions , we and others have observed multiple examples of rPrPSen substrates that can be converted by some types of prion seeds and not others [22 , 23] . Therefore , we were surprised to find that BV rPrPSen can be induced to convert to ThT-positive amyloid by every type of prion-associated seed that we have tried so far ( n = 28 ) , including several that had never before been detected by RT-QuIC or PMCA . We also did not anticipate that different PK-resistant BV rPrPRes products of RT-QuIC reactions would be seeded with different prion strains from a single host species , because we had never seen such distinct templating with the many other rPrPSen substrates that we have tested . These findings suggest that BV rPrPSen-based RT-QuIC reactions may provide a new means of probing the strain-dependent heterogeneity of prion seeding activities and conformational templates . However , overall , the RT-QuIC technology has been established largely for the practical purposes of rapid , sensitive prion disease-associated seed detection rather than the in vitro recapitulation of prion propagation . As such , the RT-QuIC tests have not been developed to reflect prion transmission barriers or strain-specificities . In any case , the availability of BV rPrPSen as an apparently universal RT-QuIC substrate may markedly improve the practicality , efficiency and cost-effectiveness of detecting and discriminating prions .
Brain tissue from scrapie-infected mice and hamsters ( Table 2 ) were collected under Protocols 2013–030 and 2010–045 , respectively , that were approved by the Rocky Mountain Laboratories Animal Care and Use Committee . Human brain tissues ( Table 1 ) were obtained from the National Prion Disease Pathology Surveillance Center ( USA ) . Brain tissue from humans with vCJD ( Table 1 ) was obtained from the National Institute for Biological Standards and Controls ( UK ) repository . No human samples were collected expressly for this study , but were instead obtained from the existing collections noted above with approval , as needed , under exemption #1197 from the NIH Office of Human Subjects Research . All human samples were , and remain , anonymized to the investigators at Rocky Mountain Laboratories where the RT-QuIC testing was performed . Recombinant prion protein ( rPrPSen ) substrates were purified as previously described [49] . Briefly , PrP DNA sequences encoding for Syrian golden hamster ( residues 23 to 231; accession no . K02234; or residues 90–231 ) , Bank Vole ( residues 23 to 230; Methionine at residue 109; accession no . AF367624 ) or hamster-sheep chimera ( Syrian hamster residues 23 to 137 followed by sheep residues 141 to 234 of the R154Q171 polymorph [accession no . AY907689] ) prion protein genes were ligated into the pET41 vector ( EMD Biosciences ) . Vectors were transformed into Rosetta ( DE3 ) Escherichia coli and were grown in Luria broth medium in the presence of kanamycin and chloramphenicol . Protein expression was induced using the autoinduction system [50 , 51] and was purified from inclusion bodies under denaturing conditions using Ni-nitrilotriacetic acid ( NTA ) superflow resin ( Qiagen ) with an ÄKTA fast protein liquid chromatographer ( GE Healthcare Life Sciences ) . The protein was refolded on the column using a guanidine HCl reduction gradient and eluted using an imidazole gradient as described [49] . The eluted protein was extensively dialyzed into 10 mM sodium phosphate buffer ( pH 5 . 8 ) , filtered ( 0 . 22-μm syringe filter [Fisher] ) and stored at -80°C . Protein concentration was determined by measuring absorbance at 280 nm . Brain homogenates ( BH; 10% w/v , Tables 1 and 2 ) were prepared as previously described [14] and stored at -80°C . For RT-QuIC analysis BHs were serially diluted in 0 . 1% SDS ( sodium dodecyl sulfate , Sigma ) /N2 ( Gibco ) /PBS as previously reported ( 25 ) , or where indicated the last dilutions were performed to a final concentration of 0 . 05% SDS/N2/PBS . RT-QuIC reactions were performed as previously described [14] . Reaction mix was composed of 10 mM phosphate buffer ( pH 7 . 4 ) , 300 or 130 mM NaCl , 0 . 1 mg/ml rPrPSen , 10 μM thioflavin T ( ThT ) , 1 mM ethylenediaminetetraacetic acid tetrasodium salt ( EDTA ) , and 0 . 002% or 0 . 001% SDS . NaCl and SDS concentrations were varied where indicated . Aliquots of the reaction mix ( 98 μL ) were loaded into each well of a black 96-well plate with a clear bottom ( Nunc ) and seeded with 2 μL of indicated BH dilutions . The plate was then sealed with a plate sealer film ( Nalgene Nunc International ) and incubated at 42°C in a BMG FLUOstar Omega plate reader with cycles of 1 min shaking ( 700 rpm double orbital ) and 1 min rest throughout the indicated incubation time . ThT fluorescence measurements ( 450 +/-10 nm excitation and 480 +/-10 nm emission; bottom read ) were taken every 45 min . To compensate for minor differences in baselines between fluorescent plate readers and across multiple experiments , data sets were normalized to a percentage of the maximal fluorescence response ( 260 , 000 rfu ) of the plate readers after subtraction of the baseline , as described [34] , and plotted versus reaction time . Reactions were classified as RT-QuIC positive base on criteria similar to those previously described for RT-QuIC analyses of brain specimens [14 , 34] . RT-QuIC reaction products were collected from the plates by extensive scraping and pipetting and treated with 10 μg/ml Proteinase K ( PK ) for 1 hour at 37°C with 400 rpm orbital shaking . Equal volumes of PK-treated reactions were run on 12% Bis-Tris NuPAGE gels ( Invitrogen ) . Proteins were transferred to an Immobilon P membrane ( Millipore ) using the iBlot Gel Transfer System ( Invitrogen ) . Membranes were probed with R20 primary antiserum ( hamster epitope: residues 218–231 ) [52] diluted 1:15 , 000 and visualized with the Attophos AP fluorescent substrate system ( Promega ) according to the manufacturer's recommendations .
|
Prion diseases are neurodegenerative disorders that propagate as multiple strains in a variety of mammalian species . The detection of all such prion types by a single ultrasensitive assay , such as the Real Time Quaking-induced Conversion ( RT-QuIC ) assay , would facilitate prion disease diagnosis , surveillance , and research . Here we show detection of minute amounts of 28 different prion types from humans , cattle , sheep , cervids and rodents , some of which were previously undetectable , using a single recombinant bank vole prion protein substrate . We also demonstrate the generation of prion type-dependent RT-QuIC conversion products which may help with prion strain discrimination and the characterization of distinct classes of prion templates . Finally , we describe a practical strategy for prion strain discrimination , e . g . classical and atypical L-type bovine spongiform encephalopathy; classical and atypical Nor98 sheep scrapie; and human sporadic and variant Creutzfeldt-Jakob disease . Thus , our study provides a basis for wide-ranging prion detection and strain discrimination .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Bank Vole Prion Protein As an Apparently Universal Substrate for RT-QuIC-Based Detection and Discrimination of Prion Strains
|
Synaptic interactions between neurons of the human cerebral cortex were not directly studied to date . We recorded the first dataset , to our knowledge , on the synaptic effect of identified human pyramidal cells on various types of postsynaptic neurons and reveal complex events triggered by individual action potentials in the human neocortical network . Brain slices were prepared from nonpathological samples of cortex that had to be removed for the surgical treatment of brain areas beneath association cortices of 58 patients aged 18 to 73 y . Simultaneous triple and quadruple whole-cell patch clamp recordings were performed testing mono- and polysynaptic potentials in target neurons following a single action potential fired by layer 2/3 pyramidal cells , and the temporal structure of events and underlying mechanisms were analyzed . In addition to monosynaptic postsynaptic potentials , individual action potentials in presynaptic pyramidal cells initiated long-lasting ( 37 ± 17 ms ) sequences of events in the network lasting an order of magnitude longer than detected previously in other species . These event series were composed of specifically alternating glutamatergic and GABAergic postsynaptic potentials and required selective spike-to-spike coupling from pyramidal cells to GABAergic interneurons producing concomitant inhibitory as well as excitatory feed-forward action of GABA . Single action potentials of human neurons are sufficient to recruit Hebbian-like neuronal assemblies that are proposed to participate in cognitive processes .
Functional characterization of microcircuits of the cerebral cortex of rodents , carnivores , and to some extent , monkeys has been propelled by simultaneous multiple recordings from synaptically connected neurons combined with anatomical and molecular analysis of the recorded cells providing direct experimental analysis of neural connections [1–5] . In the human cortical microcircuit , however , only single cells were characterized to date; interactions between identified neurons were not studied [6–8] . Recent in vivo experiments in rodents showed that individual neurons of the cerebral cortex can effectively initiate movements [9] and modulate behavioral tasks [10] . This suggests that the activity of a single cell is sufficient for driving a relatively widespread functional assembly of neurons . However , mechanisms at the level of microcircuits are not clear in producing single-neuron–triggered events requiring the activation of neural assemblies originally postulated to be important in higher order brain functions by Hebb [11] . Feed-forward excitatory and inhibitory connections are required for wiring spike and signal propagation in neural circuits and were proposed to participate in sculpting the pattern of firing traveling through the network [11–20] . Experiments testing the effectiveness of individual pyramidal neurons of the cortex showed excitatory , but usually subthreshold , postsynaptic potentials and occasional triggering of postsynaptic spikes in interneurons , leading to temporally limited ( <3 ms ) microcircuit events terminated by disynaptic inhibitory responses [2 , 18 , 21–23] . Thus , it is considered that single presynaptic spikes in pyramidal cells are not sufficient for initiating postsynaptic firing in glutamatergic neurons [24] , and effective triggering of subsequent multistep event sequences characteristic to functional neuronal assemblies requires concomitant activation of several convergent inputs or repeated firing of single presynaptic cells . We set out to record the first interactions between identified human pyramidal cells and their postsynaptic target neurons in order to characterize the saliency of single cells and the contribution of unitary signals to the triggering of compound network events . Our recordings reveal that single spikes of pyramidal neurons are followed , not only by monosynaptic excitatory postsynaptic potentials ( EPSPs ) , but also by complex event sequences with a stereotyped series of polysynaptic potentials . We then show some of the underlying network mechanisms that sequentially combine the pathway-specific effectiveness of glutamatergic excitation followed by a concomitant and bimodal GABAergic wave of events composed of inhibitory and excitatory effects .
We prepared brain slices from small blocks of nonpathological samples that had to be removed for the surgical treatment of brain areas beneath temporal ( n = 30 ) , frontal ( n = 22 ) , and parietal ( n = 6 ) cortices of 58 nonepileptic patients aged 18 to 73 y [25] . Applying simultaneous dual , triple , and quadruple whole-cell recordings from cell groups ( n = 429 ) in layer 2/3 of slices ( n = 383 ) , we elicited presynaptic action potentials in 681 pyramidal cells and searched for mono- and polysynaptic potentials in possible target neurons of which 252 were pyramidal cells and 481 were interneurons , reflecting a search bias towards interneurons during the experiments . As expected , single action potentials of pyramidal cells ( n = 121 ) triggered monosynaptic EPSPs in simultaneously recorded pyramidal cells ( n = 38 ) and interneurons ( n = 93 ) with latencies of 0 . 91 ± 0 . 46 ms and 0 . 86 ± 0 . 32 ms ( Figure 1A ) . Interestingly , single action potentials of pyramidal cells ( n = 177 ) also triggered polysynaptic postsynaptic potentials in pyramidal cells ( n = 144; latency , 10 . 02 ± 6 . 83 ms ) and in interneurons ( n = 82; latency 8 . 54 ± 6 . 19 ms ) with latencies significantly longer than that of monosynaptic events ( p < 0 . 001 for both; Figure 1A–1C ) . Postsynaptic pyramidal cells and interneurons were targeted by polysynaptic events with inhibitory ( n = 195 and n = 96 , respectively ) and excitatory ( n = 12 and n = 5 , respectively ) polarity , and several polysynaptic events could arrive at the same postsynaptic pyramidal cell ( n = 22 ) or interneuron ( n = 13; Figure 1A–1C ) . The frequency of single-spike–initiated synaptic potentials was increased for 37 ± 17 ms ( range , 5–65 ms ) compared to the control period that was measured 0–10 ms prior to the action potential ( Figure 1A–1C ) . Thus , single spikes of individual pyramidal cells initiated a series of synaptic potentials with latencies approximately ten times longer than detected previously [2 , 18 , 21–23] . The standard deviation of the latency of polysynaptic connections showed a double logarithmic correlation with the latency ( R = 0 . 90 , p < 0 . 001 , Figure 1D ) , but the probability of polysynaptic events was not correlated with their latency . Polysynaptic events in response to single spikes were detected in 105 out of 383 slices ( 27% ) of samples taken from 54 out of 58 ( 93% ) patients , showing no correlation with the cortical area from which slices were cut or with the age , diagnosis , and medication of the patients . This suggests that detected polysynaptic events are based on properties intrinsic to the human microcircuit that are conserved across brain areas . Mono- and polysynaptic potentials in the network followed the trigger spike in a sequence of waves as observed on scattergrams representing the temporal pattern of events following unitary activation ( Figure 1C ) . We searched for temporal correlations between synaptic events in single-neuron–activated networks , i . e . , whether the latency of a polysynaptic postsynaptic potential would move together with that of a different event detected in the same or in a distinct cell relative to the trigger spike . Some events showed no significant correlation with preceding or following synaptic responses , presumably due to the aforementioned increase in the jitter of onset accompanying longer latencies ( Figure 2A ) . However , we also found highly correlated ( R > 0 . 84 , p < 0 . 05 ) pairs of synaptic events composed of polysynaptic inhibitory postsynaptic potential ( IPSP ) —polysynaptic IPSP ( n = 7 ) and polysynaptic IPSP—polysynaptic EPSP ( n = 7 ) sequences that occurred up to 35 ms after the action potential ( Figure 2B ) . Accordingly , the latency of some polysynaptic potentials moves together relative to the trigger spike and could be temporally referenced , not only to the initial action potential , but also to intermediate synaptic events , indicating firing of other neurons . This suggests that a single unitary signal is sufficient for activating a predefined sequence of action potentials traveling through a subset of neurons that do not necessarily receive direct input from the trigger cell and thus correspond to cell assemblies postulated by Hebb [11] . We applied pharmacological tools in search of mechanisms underlying unitary-spike–initiated network events . Application of the GABAA receptor antagonist gabazine ( 10 μM , n = 5 ) or the AMPA receptor antagonist NBQX ( 10 μM , n = 4 ) alone or consecutively ( n = 6 ) effectively blocked pyramidal cell–triggered polysynaptic IPSPs ( Figure 3A–3C ) . The dual sensitivity to glutamatergic and GABAergic antagonists suggests effective spike propagation from pyramidal cells to postsynaptic GABAergic neurons producing disynaptic IPSPs as shown earlier [2 , 18 , 21–23] . To date , however , pyramidal neurons were found to trigger disynaptic IPSPs only; further polysynaptic events in general , and EPSPs in particular , could not be initiated by individual pyramidal neurons . We found that human pyramidal neurons triggered depolarizing , polysynaptic EPSPs ( n = 17 ) with latencies ( 11 . 88 ± 8 . 66 ms ) distinct from that of monosynaptic EPSPs ( Figure 3D and 3E , p < 0 . 001 ) . Surprisingly , polysynaptic excitatory events were blocked by gabazine ( n = 6 ) similar to polysynaptic inhibitory events ( Figure 3D ) , indicating that a GABAergic step was required before the activation of downstream pyramidal neurons eliciting the polysynaptic excitation . Moreover , the temporal distribution of synaptic events recorded in experiments with single-spike–triggered polysynaptic EPSPs showed that monosynaptic EPSPs were followed exclusively by disynaptic IPSPs , and thus polysynaptic EPSPs had minimally trisynaptic latencies ( Figure 3E ) . These experiments suggest that human pyramidal cells initially recruit postsynaptic spikes only in GABAergic neurons , then network mechanisms produce downstream IPSPs and EPSPs . As an initial test of this hypothesis , we performed simultaneous paired recordings and measured the amplitude of unitary EPSPs elicited by local pyramidal cells targeting pyramidal cells ( n = 38 ) and fast-spiking interneurons ( n = 65 ) recorded at resting membrane potentials of −72 ± 3 and −62 ± 4 mV , respectively ( Figure 4 ) . The firing behavior alone does not define a type of interneuron , thus we classified fast-spiking interneurons as basket ( n = 56 ) and axo-axonic cells ( n = 9 ) , based on light microscopically identified axonal branches forming perisomatic baskets or characteristic axonal cartridges or candles , respectively ( Figures 5B and 6B ) [25–27] . According to similar intrinsic electrophysiological properties and EPSP characteristics , we pooled the data from basket and axo-axonic cells . The amplitude distribution of unitary glutamatergic inputs targeting pyramidal cells and fast-spiking interneurons differed significantly due to unitary EPSPs of high amplitude found in interneurons ( p < 0 . 001 , Kolmogorov-Smirnov test , Figure 4 ) . Similar to what was found in other species [2 , 5 , 21 , 23 , 28] , the input resistance of postsynaptic fast-spiking interneurons ( 134 ± 64 MΩ ) was higher ( p < 0 . 042 ) than that of pyramidal cells ( 102 ± 51 MΩ ) receiving unitary EPSPs . However , the approximately 30% difference in input resistances could not account for the finding of unitary EPSPs with amplitudes bigger than approximately 6 mV in fast-spiking cells , and , furthermore , there was no correlation between the amplitude of unitary EPSPs and postsynaptic input resistance in pyramidal cells ( p < 0 . 209; R2 = 0 . 025 ) and in interneurons ( p < 0 . 552; R2 = 0 . 017 ) . This suggests that a subset of connections targeting fast-spiking interneurons is selectively strengthened in the human cortical circuit . We also compared the amplitude distribution of spontaneous EPSPs arriving at the presynaptic pyramidal cells and postsynaptic fast-spiking cells ( Figure 4 ) . The distribution of spontaneous EPSPs normalized to input resistances was shifted towards higher amplitudes in interneurons relative to pyramidal cells ( p < 0 . 001 , Kolmogorov-Smirnov test ) , corroborating our data on unitary EPSPs . In line with our hypothesis , spike transmission between pyramidal cells was not detected , but unitary EPSPs could mediate direct spike-to-spike coupling ( n = 14 ) with a probability of 46 ± 27% in individual connections from pyramidal cells to fast spiking interneurons held at resting membrane potential of −59 ± 3 mV ( Figures 3D and 4A ) . The latency of postsynaptic action potentials in fast-spiking cells ( 3 . 34 ± 1 . 76 ms ) was in the range of the latency of pyramidal cell–triggered disynaptic IPSPs . Subsequent light microscopic reconstruction of the connected cells and electron microscopic analysis of all potential sites of interaction in two of such connections showed that a relatively small number of ultrastructurally identified synaptic junctions ( n = 3 and 6 ) scattered over the dendritic surface 29 ± 19 μm and 94 ± 34 μm from the soma of the postsynaptic basket cells was sufficient for mediating the spike-to-spike coupling ( Figures 3F and S1 ) . Single pyramidal spike–driven suprathreshold responses in inhibitory interneurons explain the occurrence of disynaptic IPSPs following action potentials of pyramidal cells [2 , 18 , 21–23] but do not provide the mechanism for the recruitment of gabazine-sensitive polysynaptic excitatory events at trisynaptic latencies . However , postsynaptic action potentials were triggered , not only in 20% of GABAergic basket cells known to elicit hyperpolarizing IPSPs ( n = 11 , Figure 5 ) , but also in 33% of axo-axonic or chandelier cells ( n = 3 , Figure 6 ) capable of triggering depolarizing IPSPs and postsynaptic action potentials in neocortical pyramidal cells [25 , 29] . Axo-axonic cells were shown to be effective in the selective recruitment of postsynaptic glutamatergic neurons and disynaptic EPSPs in the cortex and amygdala [25 , 30] . Thus , spike transmission from pyramidal to axo-axonic cells and then from axo-axonic cells to pyramidal neurons could form the mechanism underlying single pyramidal cell–triggered , trisynaptic , and gabazine-sensitive EPSPs suggested also by experiments detecting single action potential–initiated polysynaptic recruitment of downstream spikes ( n = 4 , Figure 7 ) . Furthermore , all human axo-axonic cells in our sample ( n = 9 ) triggered polysynaptic series of events with latencies corresponding to monosynaptic GABAergic , disynaptic glutamatergic , and/or trisynaptic GABAergic postsynaptic potentials ( Figure 6A and 6C ) . The resulting alternating sequence of inhibitory and excitatory waves of responses were similar to brief , high-frequency oscillations ( 175 ± 28 Hz , measured between the peaks of inhibitory components taken as troughs , Figure 6C ) .
Our results show that a single spike of a pyramidal cell in the human cortical microcircuit is capable of activating complex sequences of postsynaptic potentials lasting an order of magnitude longer than detected previously [2 , 18 , 21–23] . The initiation and internally precise temporal structure of these event series appears to follow a stereotyped mechanism of spike-to-spike transmission traveling through a subset of synaptically connected neurons that seems to be conserved across several brain regions of the human cerebral cortex . The flow of downstream activation that follows the first-order spike of the trigger pyramidal cell is directionally controlled at two consecutive synaptic steps . Second-order spikes are triggered exclusively in GABAergic interneurons and not in pyramidal cells due to interneuron-selective EPSPs of enormous amplitude . In turn , second-order spikes in axo-axonic cells give rise to third-order spikes detected only in pyramidal cells , resulting in trisynaptic EPSPs in the network because axo-axonic cells do not innervate other GABAergic cells [26 , 27] . Synchronized to the spikes in axo-axonic cells , second-order spikes in basket cells and possibly in other types of interneuron [18 , 23] elicit hyperpolarizing effects reported here as disynaptic IPSPs . Conversely , the IPSPs hyperpolarizing pyramidal cells could enhance the excitatory effect of axo-axonic cells by pushing the postsynaptic membrane potential further from the relatively depolarized axonal reversal potential for GABA [25 , 29] , especially during the second spike of spike doublets recorded frequently in cortical and hippocampal axo-axonic cells [25 , 31] . Moreover , hyperpolarizing IPSPs suppress the activity of pyramidal cells not recruited by axo-axonic cells , decreasing the frequency of EPSPs not routed to the initial spike . Thus , the actual pattern of single-cell–initiated network activity , especially during the first two to three synapses , could be shaped by the convergence and divergence relationship of inhibitory and excitatory GABAergic synapses in addition to glutamatergic recruitment . The clockwork precision of these mono- , di- , and trisynaptic events of alternating glutamatergic and GABAergic transmission resembles sharp , wave-like oscillation hotspots in the microcircuit [27 , 32–35] . At later stages of the polysynaptic chain , these mechanisms are likely to be combined with summation of glutamatergic EPSPs leading to suprathreshold responses in pyramidal cells in addition to interneurons and with rebound excitation following IPSPs at a relatively longer delay [16] . Synapses arriving from a single human pyramidal cell are not capable of driving the postsynaptic pyramidal cells to fire . Accordingly , there are two alternatives for eliciting suprathreshold responses in a pyramidal cell from local sources: ( 1 ) single axo-axonic cell-to-pyramid connections [25] and ( 2 ) synchronous inputs from several pyramidal/glutamatergic cells converging onto the same postsynaptic pyramidal cell [24] . Importantly , axo-axonic cells recruit the firing of several postsynaptic pyramidal cells in synchrony [25] , and thus they can contribute to pyramid–pyramid spike-to-spike coupling ( Figure 7 ) . The finding that polysynaptic EPSPs are relatively rarely found could be due to the limited number of axo-axonic cells relative to basket cells or to the limited percentage of pyramidal cells responding with a spike to axo-axonic input [25] . This , however , could be crucial in preventing overexcitation of the microcircuit and can effectively dampen the detonator effect of axo-axonic cells . Our data suggest that axo-axonic cells are crucial in the distribution of local excitation originating from sporadic spikes in pyramidal cells , but pyramid–pyramid connections are highly convergent in the cortex [36] , thus sufficient synchronization of inputs could occur during various cortical operations . The dataset presented here does not provide sufficient evidence to decide whether single-EPSP–triggered , long-lasting event series are specific to the human cerebral cortex . Interestingly , previous studies testing the effect of identified pyramidal cells in the monkey cortex did not report recruitment of polysynaptic activity [5] . When searching for similar , single pyramidal cell–triggered polysynaptic postsynaptic potentials in our library of recordings performed in the somatosensory and prefrontal cortex of the rat ( n > 4 , 500 unitary connections ) , we could only detect sporadic triggering of polysynaptic events that were IPSPs only ( n = 8 ) apart from the unitary EPSPs ( n = 724 ) elicited by layer 2/3 pyramidal neurons . Although the ratio of triggering poly- versus monosynaptic postsynaptic potentials was 0 . 01 in the rat and 1 . 73 in the human in our hands , it should be emphasized that the human patients were treated differently during anesthesia and surgery , and the excitability of human neurons might be different in the external solution also used for rat experiments . Applying similar constraints , unitary EPSPs of high amplitude arriving to human interneurons were not reported in other species to date . It is apparent that the potential amplitude difference in unitary EPSPs between human and nonhuman interneurons is not attributable to interspecies differences in input resistances [2 , 5 , 21 , 23 , 28] . Quantal parameters of synaptic transmission underlying gigantic EPSPs on subsets of human GABAergic interneurons are not known and are being tested as part of a separate study . The function of single-spike–initiated event series is not clear , although they fit into the framework of the cell assembly concept proposed by Hebb [11] . The present study identifies that Hebbian-like cell assemblies can be recruited by a unitary signal of a single pyramidal cell and suggests a canonical order of activation of identified cell populations assigning the numbers representing the sequence of active connections in the original illustration of Hebb [11] . Our data support the idea of specialized and strong pathways in the cortex among the many weaker pathways [11 , 20 , 33 , 35 , 37] and identifies a novel cell population , the axo-axonic cells , selectively targeted by strong inputs from individual pyramidal neurons . Mechanisms producing the powerful glutamatergic synapses on axo-axonic and basket cells are not fully understood . Interestingly , it has been suggested that a special form of LTP may occur in interneurons that are silent during periods of intense pyramidal cell firing , such as sharp waves [38] , and axo-axonic cells fire in vivo before , but not during sharp waves [34] . Feed-forward inhibition provides a framework for efficient propagation and selection of excitatory events [18 , 22] especially when synchronized to the feed-forward GABAergic excitation and recruitment of a subset of glutamatergic neurons leading to the correlated action of neurons several synapses downstream in the activated network . The increased signal-to-noise ratio in the network provided by hyperpolarizing GABAergic synapses is further amplified by the coincident action of chandelier cells , resulting in a sparse and potentially task-selective activation of pyramidal neurons . Thus , the human microcircuit appears to be tuned for unitary-EPSP–activated Hebbian-like functional cell assemblies [11 , 37 , 39] that were proposed as building blocks of higher-order cortical operations and could contribute to single cortical cell–initiated movements [9] and behavioral responses [10] .
All procedures were performed according to the Declaration of Helsinki with the approval of the University of Szeged Ethical Committee . Human slices were derived from material that had to be removed to gain access for the surgical treatment of deep-brain tumors from the left and right frontal , temporal , and parietal regions with written informed consent of the patients ( aged 18–73 y ) prior to surgery over the last 4 y . Anesthesia was induced with intravenous midazolam and fentanyl ( 0 . 03 mg/kg , 1–2 μg/kg , respectively ) . A bolus dose of propofol ( 1–2 mg/kg ) was administered intravenously . To facilitate endotracheal intubation , the patient received 0 . 5 mg/kg rocuronium . After 120 s , the trachea was intubated and the patient was ventilated with a mixture of O2-N2O at a ratio of 1:2 . Anesthesia was maintained with sevoflurane at monitored anesthesia care ( MAC ) volume of 1 . 2–1 . 5 . Blocks of tissue were immersed in ice-cold solution containing ( in mM ) 130 NaCl , 3 . 5 KCl , 1 NaH2PO4 , 24 NaHCO3 , 1 CaCl2 , 3 MgSO4 , 10 d ( + ) -glucose , saturated with 95% O2 and 5% CO2 in the operating theatre , sliced at a thickness of 350 μm with a vibrating blade microtome ( Microm HM 650 V ) and were incubated at room temperature for 1 h in the same solution . The solution used during recordings differed only in that it contained 2 mM CaCl2 and 1 . 5 mM MgSO4 . Recordings were obtained at approximately 36 °C from up to four concomitantly recorded cells visualized in layer 2/3 by infrared differential interference contrast videomicroscopy at depths 60–130 μm from the surface of the slice . Signals were filtered at 8 kHz , digitized at 16 kHz , and analyzed with PULSE software . Micropipettes ( 5–7 MΩ ) were filled with a low [Cl]i solution for discriminating GABAergic and glutamatergic events containing ( in mM ) 126 K-gluconate , 4 KCl , 4 ATP-Mg , 0 . 3 GTP-NA2 , 10 HEPES , 10 phosphocreatine , and 8 biocytin ( pH 7 . 20; 300 mOsm ) . Presynaptic cells were stimulated with brief ( 2–10 ms ) suprathreshold pulses delivered at >7-s intervals , to minimize intertrial variability . Membrane properties of human neurons or polysynaptic events did not show significant changes for up to 20 h after slicing , but recordings included in the analysis were arbitrarily terminated 15 h after slice preparation . Traces shown are single sweeps or averages of 50–100 consecutive episodes . Data are given as mean ± standard deviation ( S . D . ) , Mann-Whitney U-test , paired t-test ( pharmacology ) , and Kolmogorov-Smirnov test was used to compare datasets; differences were accepted as significant if p < 0 . 05 . Visualization of biocytin and correlated light and electron microscopy was performed as described [25 , 40] . Three-dimensional light microscopic reconstructions were carried out using Neurolucida with 100× objective .
|
We recorded the first connections , to our knowledge , between human nerve cells and reveal that a subset of interactions is so strong that some presynaptic cells are capable of eliciting action potentials in the postsynaptic target neurons . Interestingly , these strong connections selectively link pyramidal cells using the neurotransmitter glutamate to neurons releasing gamma aminobutyric acid ( GABA ) . Moreover , the GABAergic neurons receiving the strong connections include different types: basket cells , which inhibit several target cell populations , and another type called the chandelier cells , which can be excitatory and target pyramidal cells only . Thus , the activation originating from a single pyramidal cell propagates to synchronously working inhibitory and excitatory GABAergic neurons . Inhibition then arrives to various neuron classes , but excitation finds only pyramidal cells , which in turn , can propagate excitation even further in the network of neurons . This chain of events revealed here leads to network activation approximately an order of magnitude longer than detected previously in response to a single action potential in a single neuron . Individual-neuron–activated groups of neurons resemble the so-called functional assemblies that were proposed as building blocks of higher order cognitive representations .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"neuroscience",
"physiology"
] |
2008
|
Complex Events Initiated by Individual Spikes in the Human Cerebral Cortex
|
Salmonella enterica serovar Typhimurium ( S . Tm ) is a cause of food poisoning accompanied with gut inflammation . Although mucosal inflammation is generally thought to be protective against bacterial infection , S . Tm exploits the inflammation to compete with commensal microbiota , thereby growing up to high densities in the gut lumen and colonizing the gut continuously at high levels . However , the molecular mechanisms underlying the beneficial effect of gut inflammation on S . Tm competitive growth are poorly understood . Notably , the twin-arginine translocation ( Tat ) system , which enables the transport of folded proteins outside bacterial cytoplasm , is well conserved among many bacterial pathogens , with Tat substrates including virulence factors and virulence-associated proteins . Here , we show that Tat and Tat-exported peptidoglycan amidase , AmiA- and AmiC-dependent cell division contributes to S . Tm competitive fitness advantage in the inflamed gut . S . Tm tatC or amiA amiC mutants feature a gut colonization defect , wherein they display a chain form of cells . The chains are attributable to a cell division defect of these mutants and occur in inflamed but not in normal gut . We demonstrate that attenuated resistance to bile acids confers the colonization defect on the S . Tm amiA amiC mutant . In particular , S . Tm cell chains are highly sensitive to bile acids as compared to single or paired cells . Furthermore , we show that growth media containing high concentrations of NaCl and sublethal concentrations of antimicrobial peptides induce the S . Tm amiA amiC mutant chain form , suggesting that gut luminal conditions such as high osmolarity and the presence of antimicrobial peptides impose AmiA- and AmiC-dependent cell division on S . Tm . Together , our data indicate that Tat and the Tat-exported amidases , AmiA and AmiC , are required for S . Tm luminal fitness in the inflamed gut , suggesting that these proteins might comprise effective targets for novel antibacterial agents against infectious diarrhea .
Protein translocation constitutes an essential cell function in all types of prokaryotic or eukaryotic cells . Accordingly , bacteria have evolved several sophisticated translocation systems to transport proteins into , or across , the cytoplasmic membrane . In particular , for bacterial pathogens , this process is suspected to contribute to pathogenesis and interbacterial competition because the substrates include virulence factors in certain cases . Most general protein transport occurs via the Sec system , which predominantly transports unfolded proteins across the cytoplasmic membrane [1] . In addition , other additional transport systems have also evolved to facilitate the transport of different types of protein . The twin-arginine translocation ( Tat ) system is essential for the appropriate localization of a number of folded proteins that function outside the cytoplasm [2] . In most Gram-negative bacteria including Salmonella , the Tat transport machinery is composed of the integral membrane proteins TatA , TatB , and TatC , which are inserted in the inner membrane [3 , 4] . These proteins form their respective homo-oligomers in the membrane , with TatB and TatC subsequently associating as a complex ( TatBC ) [5] whereas TatA remains separate . During the initiation of protein export , a folded substrate protein docks at the TatBC complex through binding of its twin arginine signal peptide [6] . TatA tetramer is then recruited to the substrate-bound TatBC complex to form the translocation channel [7 , 8] . Finally , the substrate is transported by crossing the membrane via the polymerized TatA component , whereupon its signal sequence is cleaved and the TatA proteins disassociate [9–11] . Notably , the Tat system is conserved among many bacterial pathogens [3 , 12] , with Tat substrates including both virulence factors and virulence-associated proteins [13–24] . Hence , it is well recognized that the Tat system is implicated in bacterial virulence [4 , 25] , leading to the hypothesis that the Tat system may constitute a novel therapeutic target against infection with bacterial pathogens . Salmonella enterica serovar Typhimurium ( S . Tm ) is a common cause of diarrhea worldwide [26] . Following oral infection and upon reaching the gut lumen , S . Tm relies on flagella-based motility to gain access to the epithelial surface of the intestine [27 , 28] . During its interaction with the host mucosa , genomic islands , termed SPI-1 and SPI-2 , encoding type III secretion systems ( ttss-1 and ttss-2 ) allow the bacterium to invade the epithelial cells and elicit mucosal inflammation [29 , 30] . This inflammation depends upon activation of the NAIP/NLRC4 inflammasome , resulting in secretion of the inflammatory cytokines IL-1β and IL-18 [31–33] . During the initial stages ( the first 12–18 hours post-infection ) , gut inflammation acts as an innate defense by reducing S . Tm loads in the infected tissue [32] . However , at later stages , S . Tm can instead benefit from gut inflammation for its own competitive growth advantage in the gut lumen [34] . Accumulating evidence has recently supported several underlying mechanisms for explaining the inflammation-mediated outgrowth blooming of S . Tm ( reviewed in [28 , 35 , 36] ) . For example , gut inflammation provides electron acceptors , which promote anaerobic respiration of S . Tm [37] . Inflammation-based transmigration of polymorphonuclear leukocytes into the gut lumen leads to the generation of tetrathionate ( S4O62– ) via reactive oxygen species . S . Tm , but not gut commensal bacteria , can use the tetrathionate as a terminal electron acceptor for anaerobic respiration , and thereby out-grow the competing microbiota in the inflamed gut . Although these findings provide an initial understanding of the beneficial competitive growth of S . Tm in the inflamed gut , a complete comprehension of the molecular mechanisms underlying this bloom remains to be elucidated . Previous studies have shown that the Tat system is involved in Salmonella virulence [38–40] . For example , an S . Enteritidis mutant strain lacking a functional Tat system exhibits pleiotropic defects in virulence , with attenuated invasiveness into cultured cells , colonization of the cecum , and systemic infection in chickens [38] . Furthermore , the tatC mutant of S . Tm has been shown to be attenuated in a mouse model for typhoid fever in humans [39 , 40] . This is also suspected to be attributable to multiple defects including attenuated invasion of macrophages , resistance to antimicrobials , motility , and expression of ttss-1 and ttss-2 . However , the classical mouse model for typhoid fever was employed to evaluate virulence in these previous studies [39 , 40] , in which orally infected S . Tm causes systemic infection such as typhoid fever without gastrointestinal disease . The absence of enterocolitis is explained by the fact that the S . Tm strains cannot colonize the gut continuously owing to the competing gut microbiota . Therefore , it is completely unknown whether the Tat system is involved in Salmonella gut infection . Alternatively , in the present study we have utilized the streptomycin mouse model for gut infection to study the role of the Tat system in Salmonella-induced colitis . In particular , the treatment with streptomycin can transiently reduce the normal gut microbiota , allowing the elicitation of gut inflammation and gut colonization by S . Tm [41 , 42] . Using this infection model , we demonstrated that the Tat system is involved in S . Tm gut infection by contributing to sustained gut colonization . Furthermore , we showed that Tat substrate-dependent cell division is required for sustained colonization in the inflamed gut , as mutations in AmiA and AmiC cause S . Tm cells to form chains by a cell division defect , resulting in their enhanced susceptibility to bile acids . Our findings provide novel insight into S . Tm colonization in the inflamed gut and suggest that the Tat system and the control of bacterial cell shape might comprise promising therapeutic targets against gut infection by enteropathogenic bacteria .
To reveal the role of Tat in Salmonella enterocolitis , we first investigated whether Tat is required for the disease using the streptomycin mouse model . Two groups of streptomycin-pretreated C57BL/6 mice were infected intragastrically with 5 × 107 colony forming units ( CFU ) of S . Tm wild-type strain SL1344 or tatC mutant . At day 1 and 3 post infection ( p . i . ) , bacterial loads in feces were enumerated by dilute-plating on selective media . At day 1 p . i . , higher loads of tatC mutant were detected in feces compared with those of SL1344 , whereas both strains at day 3 p . i . displayed high levels of colonization ( Fig 1A ) . The bacterial loads of tatC mutant at day 3 p . i . in cecal content , mesenteric lymph node ( mLN ) , and spleen were significantly lower than those in SL1344 ( Fig 1B–1D ) . Histopathological analysis of the cecal mucosa showed that each mouse infected with SL1344 or tatC mutant featured gut inflammation ( Fig 1E ) . Notably , however , the mice infected with tatC mutant displayed lower levels of gut inflammation in comparison with those infected with SL1344 . In addition , the levels of gut inflammation were further verified by determining the amount of fecal lipocalin-2 , an inflammatory marker ( Fig 1F ) . To clarify the role of Tat in gut colonization , we performed a competitive infection ( CI ) experiment . Streptomycin-pretreated C57BL/6 mice were infected with a 1:1 mixture of SL1344 and tatC mutant by gavage ( 5 × 107 CFU in total ) . In the feces , the bacterial loads of T321 at day 1–4 p . i . were significantly lower than those of SL1344 ( Fig 1G ) . Similar colonization defects of tatC mutant at day 4 p . i . were found in the cecum lumen , mLN , and spleen ( Fig 1H–1J ) . Collectively , these results suggest that S . Tm Tat is involved in the Salmonella enterocolitis . To further investigate the role of Tat in gut colonization , we used an attenuated S . Tm strain harboring a mutation of the ssaV gene , which encodes a component of ttss-2 , in further experiments . The attenuated strain allowed us to evaluate sustained gut colonization in the streptomycin mouse model [43 , 44] . In Escherichia coli , S . Enteritidis , and S . Tm , tatC mutant cells have been shown to form long chains resulting from the failure to separate following cell division [38 , 40 , 45] . Furthermore , it has been reported that E . coli and S . Tm strains lacking certain Tat substrates such as Ami amidases display a cell division defect . [40 , 46] . As it is possible that the impaired colonization in the S . Tm tatC mutant strain might be due to a defect in the export of Tat substrate ( s ) , we focused on two amidases , AmiA and AmiC , as determinants responsible for the Tat-dependent gut colonization . First , we investigated whether the ssaV tatC mutant features a cell division defect . In the current experiment , test strain cells harbored a GFP-expressing plasmid ( pACYC-gfp ) , allowing the resulting bacterial strains grown in LB broth to be analyzed by fluorescence microscopy . Approximately 40–60% of the cells of ssaV mutant were single or paired , but no chains were observed ( S1A and S1B Fig ) . In contrast , in all cell configurations , the proportion of single cells of ssaV tatC mutant was reduced in comparison with that of the ssaV mutant , whereas increased proportion of chains was observed ( S1A and S1B Fig ) . Next , we performed the same analysis with amiA , amiC , or amiA amiC mutants . Similar to the ssaV tatC mutant , the proportions of single cells of ssaV amiC or ssaV amiA amiC mutants were reduced , whereas increased chains were observed . ( S1A and S1B Fig ) . In contrast , no difference in ssaV amiA mutant was observed ( S1A and S1B Fig ) . Furthermore , the cell division defect of the ssaV amiA amiC mutant was restored by introduction of a plasmid encoding the wild-type amiC , but not amiA gene ( S1C Fig ) . Finally , scanning electron microscopy analysis showed that the ssaV amiA amiC mutant strain results in chains , whereas the ssaV mutant strain displayed single or paired cells ( S1D Fig ) . These results confirmed that S . Tm Tat and its substrates AmiA and AmiC contribute to cell division and that AmiC , but not AmiA , is more likely to be required for the cell division . Thus , to investigate the role of Tat in sustained gut colonization , we performed a CI experiment using ssaV mutant and ssaV tatC mutants . The ssaV tatC mutant exhibited a substantial colonization defect in the feces at day 1 , 3 , 6 , and 8 p . i . ( Fig 2A ) . The colonization defect was also found in the cecal content and mLN at day 8 p . i . ( Fig 2B and 2C ) . Similarly , we tested ssaV amiA or ssaV amiC or ssaV amiA amiC mutants . In feces , each the ssaV amiC and the ssaV amiA amiC strains featured a significant colonization defect at day 3 , 6 , and 8 p . i . ( Fig 2G and 2J ) . Colonization defects in the cecum lumen and mLN by these mutants were also found ( Fig 2H , 2I , 2K and 2L ) . In contrast , in feces , the ssaV amiA strain displayed attenuated colonization at day 8 p . i . , whereas no colonization defects were found in the cecum lumen or mLN ( Fig 2D–2F ) . In addition , we performed a complementation CI experiment using ssaV amiC mutant harboring a plasmid encoding amiC gene . Introduction of amiC gene rescued the colonization defect of the ssaV amiC mutant at day 6 p . i . ( S2 Fig ) . These results suggest that AmiC , and to lesser extent AmiA contribute to gut colonization . Furthermore , to examine the contribution of AmiA and AmiC to the gut colonization in more detail , we repeated the CI experiments using the complemented ssaV amiA amiC strains with amiA or amiC genes . Introduction with amiC gene , but not amiA , complemented the colonization defect of the ssaV amiA amiC mutant ( S3 Fig ) . The results clearly indicated that AmiC mainly contributes to the gut colonization . The AmiA- and AmiC-dependent colonization defect was confirmed by comparing the mice infected with the ssaV or the ssaV amiA amiC mutants ( S4A–S4C Fig ) . Streptomycin-pretreated mice were infected with the ssaV or ssaV amiA amiC mutant , and gut colonization was monitored for the next 8 days p . i . . At day 1 and 3 p . i . , both strains colonized the gut at similar levels ( S4A Fig ) . However , at day 8 p . i . , the ssaV amiA amiC mutant loads were significantly decreased ( S4A Fig ) . Similar colonization defect of the ssaV amiA amiC mutant was found in the cecal lumen ( S4B Fig ) . Histopathological analysis of cecal mucosa showed that mice infected with the ssaV mutant exhibited moderate mucosal inflammation whereas the ssaV amiA amiC mutant-infected murine cecum displayed only slight inflammation ( S4C Fig ) . Collectively , these results suggested that the attenuated virulence of the S . Tm tatC mutant in the enterocolitis model could be explained by an export defect of AmiA and AmiC . To assess whether the S . Tm amiA amiC mutant strain displays long chains in the gut , streptomycin-pretreated C57BL/6 mice were infected with GFP-expressing S . Tm strains ssaV or ssaV amiA amiC by gavage . For the next 3 days , an obtained fecal pellet was suspended by pipetting gently and analyzed using fluorescence microscopy . Live microscopy analysis revealed that at day 3 p . i . , the majority of ssaV mutant were present in single or paired cells , whereas the ssaV amiA amiC cells exhibited chain formation ( Fig 3A ) . Quantitative analysis of the experiments revealed that approximate 20% of the S . Tm cells in the gut were presented in chains at day 3 p . i . ( Fig 3B ) . These results suggest that the S . Tm amiA amiC mutant forms chain-shaped structures in the gut and may imply that the chain form underlies the attenuated gut colonization observed with this mutant strain . S . Tm exploits host inflammatory responses to colonize in the gut [34] , which suggests that gut inflammation is required for the competitive fitness advantage of S . Tm . This prompted us to hypothesize that gut inflammation is also involved in the AmiA- and AmiC-dependent competitive fitness . To this end , we performed a CI experiment using the S . Tm invG ssaV avirulent strain , which is incapable of inducing the gut inflammation [47] . We first confirmed that a mutation of the invG gene had no effect on the cell division in vitro ( S5 Fig ) . Streptomycin-pretreated C57BL/6 mice were co-infected with invG ssaV and invG ssaV amiA amiC strains at a 1:1 mixture by gavage ( 5 × 107 CFU in total ) , and S . Tm loads in feces were monitored at day 1 , 3 , and 6 p . i . . Both strains displayed similar colonization levels at all monitored days ( Fig 4A ) . The absence of gut inflammation in the mice was verified by measuring fecal Lcn-2 levels ( Fig 4B ) . Next , we examined whether S . Tm amiA amiC mutant cells could form chains via the cell division defect in the normal gut using the GFP-expressing S . Tm . Live microscopy and quantitative analyses revealed that all of the invG ssaV and invG ssaV amiA amiC cells were present as single or paired cells in the gut , whereas no chains were observed ( Fig 4C and 4D ) . The results demonstrated that the cell division defect of the amiA amiC mutant strain does not occur in the normal gut . Next , to analyze the relevance of gut inflammation in supporting S . Tm colonization , we further investigated whether dextran sulfate sodium ( DSS ) could confer a colonization defect on the invG ssaV amiA amiC strain in the CI experiment . C57BL/6 mice with DSS-induced colitis were subjected to the same CI experiment as shown in Fig 4A . Effect of the DSS treatment was verified by measuring body weight of the mice , suggesting that DSS treatment for 7 days appears to induce colitis ( S6 Fig ) . In addition , the DSS-treated mice were administered orally with streptomycin prior to infection in order to reduce competitive microbiota . Colonization levels at day 3 p . i . of the invG ssaV amiA amiC strain in feces and cecal content were significantly reduced ( Fig 4E ) . Gut inflammation in the mice was confirmed by elevated fecal Lcn-2 levels ( Fig 4F ) . Finally , we investigated whether the invG ssaV amiA amiC strain cells could form chains in the DSS-induced inflamed gut . As shown in Fig 4G , the invG ssaV amiA amiC strain formed chains when used to infect the DSS-treated mouse . The results suggested that the invG ssaV amiA amiC mutant was incapable of fully dividing cells in certain conditions such as in the inflamed gut . Collectively , these results indicated that host inflammatory responses may be involved in the AmiA- and AmiC-dependent competitive fitness . To decipher the molecular mechanism underlying the AmiA- and AmiC-dependent competitive fitness advantage , we noted phenotypes that were modified following mutation of the amiA and amiC genes . Previously , motility has been shown to be attenuated by deletions of the tatC gene or both amiA and amiC genes in S . Tm strain 14028 [39 , 40] . Furthermore , motility was the first phenotype identified that allowed S . Tm to capitalize on mucosal inflammation , resulting in sustained gut colonization [27] . Thus , we investigated role of motility in the AmiA- and AmiC-dependent competitive fitness . First , we examined the ability of ssaV , ssaV tatC , or ssaV amiA amiC mutants to move on semi-agar ( 0 . 3% agar ) medium . The ssaV mutant were motile in comparison with ssaV flhA mutant as the negative control , which lacked an inner membrane protein of flagella ( Fig 5A ) . In contrast , the motility of ssaV tatC and ssaV amiA amiC mutant cells was significantly attenuated , albeit not diminished ( Fig 5A ) . This impaired motility in the ssaV amiA amiC mutant was complemented by introduction of a plasmid encoding the amiC gene , but not amiA ( Fig 5B and 5C ) . In addition , we examined swarming motility using swarming agar ( 0 . 5% agar ) medium . The ssaV mutant cells were motile on the swarming agar , whereas ssaV cheY mutant cells swarmed poorly ( S7A Fig ) . In contrast , ssaV tatC and ssaV amiA amiC strain cells were motile on the swarming agar , and the plasmid-based expression of AmiA or AmiC displayed no effect on the swarming ( S7B Fig ) . These results are in line with the previous reports [39 , 40] . To clarify the causal link between the impaired motility of the ssaV amiA amiC strain and the colonization defect , we first confirmed whether impaired motility confers colonization defect using the ssaV flhA mutant . In line with previous studies [27 , 48] , in a CI experiment , the ssaV flhA mutant displayed colonization defect ( Fig 5D ) . ELISA with fecal pellet showed that the mice in this CI experiment feature slight inflammation ( Fig 5E ) . Thus , we next performed a CI assay using two S . Tm strains ssaV flhA and ssaV flhA amiA amiC . At day 1 p . i . , both strains displayed similar colonization levels ( Fig 5F ) . However , at day 4 p . i . , the ssaV flhA amiA amiC strain showed impairment in competitive colonization ( Fig 5F ) . Accordingly , measurement of fecal Lcn-2 levels suggested that the infected mice did not develop gut inflammation ( Fig 5G ) . Furthermore , to reveal the involvement of directional movement , termed chemotaxis , in the gut colonization by S . Tm , we performed a similar CI experiment by ssaV cheY and ssaV cheY amiA amiC mutants . At day 1 and 4 p . i . , colonization levels of the ssaV cheY amiA amiC mutant were significantly reduced ( S7C Fig ) . Fecal Lcn-2 measurement of the infected mice suggested that the infected mice demonstrated only minimal inflammation ( S7D Fig ) . These results showed that the non-motile or non-chemotactic S . Tm strains still exhibit the AmiA- and AmiC-related colonization defect in the noninflamed gut , suggesting that impaired motility alone is not sufficient to explain the attenuated colonization of the S . Tm amiA amiC mutant strain . Because mice in the experiments of Fig 5F and 5G exhibited little gut inflammation , we could not exclude the possibility that bacterial motility is involved in the AmiA- and AmiC-dependent competitive fitness in the inflamed gut . Therefore , to overcome this limitation , we applied the DSS colitis model in the same CI experiment as shown in Fig 5F and 5G . In DSS-induced colitis mice , the ssaV flhA amiA amiC mutant was impaired in gut colonization ( Fig 5H ) . Lcn-2 ELISA indicated that DSS treatment elicits inflammation , and that the mice in this CI experiment have gut inflammation during the S . Tm infection ( Fig 5I ) . The results indicate that flagella-based motility may be not involved in the AmiA- and AmiC-dependent competitive fitness in the inflamed gut . Antimicrobial molecules such as antimicrobial peptides are believed to play a crucial role in the competitive bacterial fitness of the intestinal tract [49] . Recently , we have shown that S . Tm gut colonization requires a robust outer membrane to confer resistance to α-helical antimicrobial peptide [44] . Thus , to clarify the involvement of resistance to antimicrobial peptide in the AmiA- and AmiC-dependent competitive fitness , we next investigated the S . Tm amiA amiC mutant strain for sensitivity to magainin 2 , an α-helical antimicrobial peptide , by determining the minimal inhibitory concentrations ( MICs ) of magainin 2 towards S . Tm strains ( Table 1 ) . S . Tm ssaV phoP mutant featuring an attenuated outer membrane barrier was used as a control . As expected , the MICs of the ssaV phoP mutant were reduced in comparison to the SL1344 wild-type strain and ssaV mutant . In contrast , the MICs towards the ssaV tatC and ssaV amiA amiC strains were identical for the ssaV mutant . Collectively , these data suggest that the amiA amiC mutant strain is resistant to α-helical antimicrobial peptide . Earlier studies have demonstrated that tatC or amiA and amiC mutants in E . coli or S . Tm are hypersensitive to detergents such as sodium dodecyl sulfate and bile acids [39 , 40 , 45 , 50] . Furthermore , resistance to bile acids has been shown to confer a competitive fitness advantage on S . Tm [51] . Thus , we next tested for sensitivity to deoxycholate , a component of bile acid detergents . We determined the MICs of deoxycholate towards S . Tm strains ( Table 1 ) . Compared to SL1344 and ssaV mutant , the MICs of deoxycholate for ssaV phoP or ssaV tatC or ssaV amiA amiC mutants were reduced . In contrast , the MICs of complemented S . Tm strains with amiC , but not amiA , were restored partially . Collectively , these data demonstrated that S . Tm strains harboring mutations of tatC or amiC genes are sensitive to bile acid such as deoxycholate . Furthermore , it is notable that AmiC , but not AmiA , contributes to the resistance of S . Tm to bile acids . Since resistance of S . Tm to bile acids generally depends upon the robust outer membrane [52] , we next assessed bacterial outer membrane barrier by using the ethidium bromide ( EtBr ) influx assay [44] . If the outer membrane barrier is attenuated , EtBr can pass through the outer membrane easily , and subsequently reach the cytosol by traversing the cytoplasmic membrane . The EtBr in the bacterial cytosol leads to an increase in fluorescence signal by binding to intracellular nucleic acids . Compared to SL1344 , S . Tm phoP mutant cells displayed the increased fluorescence intensity ( Fig 6 ) . In contrast , the intensity of tatC or amiA amiC mutant cells was identical for that of SL1344 . These results suggest that AmiA and AmiC do not contribute to the robustness of the outer membrane , and that the outer membrane barrier does not involve the impaired resistance of the amiA amiC mutant strain towards to bile acids . Next , we investigated whether attenuated resistance to bile acid confers impaired gut colonization on the S . Tm amiA amiC mutant strain by using rodent chow containing cholestimide resin , which enhances the excretion of bile acids in feces via absorption in the intestinal tract , limiting the intestinal circulation of bile acids [54–56] . Total bile acids concentrations in feces of mice fed chow containing colestimide increased significantly compared to those of mice fed control chow ( S8A and S8B Fig ) . The results indicated that feeding with chow containing colestimide resin promotes the excretion of luminal bile acids in feces through the adsorption of the colestimide resin to bile acids , leading to a reduction in the concentrations of free luminal bile acids , which can interact with luminal substances including bacteria . Thus , we confirmed an inhibitory effect on luminal bile acids by feeding with colestimide resin . C57BL/6 mice were then fed chow containing cholestimide resin or control chow , and subjected to the streptomycin mouse model experiment with mixed infection of S . Tm ssaV and ssaV amiA amiC mutants . In mice fed normal chow , competitive colonization advantages of ssaV mutant increased gradually compared to input CI ( Fig 7A ) . In contrast , the advantages were slightly increased compared to input CI in mice fed chow containing cholestimide resin , and were significantly reduced compared to mice fed normal chow at day 4 and 6 p . i . ( Fig 7A ) . S . Tm-infected mice fed control chow displayed no increase in the concentration of total bile acids in feces , whereas in the mice fed chow containing cholestimide , total bile acids in feces increased significantly at day 4 p . i . ( Fig 7B ) . Furthermore , feeding with chow containing cholestimide resin tended to increase the concentrations of total bile acids of mice infected with S . Tm at day 6 p . i . ( Fig 7B ) . The results indicated that luminal colestimide resin binds to bile acids and thereby forms an insoluble complex , resulting in a decrease in the free luminal bile acids that can interact with S . Tm and an increase in total bile acids in feces . Finally , measurement of fecal lipocalin-2 levels suggested that both mouse groups exhibited gut inflammation and that no difference in inflammation levels existed between the two groups ( Fig 7C ) . Collectively , these results lend support to the hypothesis that the gut colonization defect of S . Tm tatC or amiA amiC mutant strains is attributable to attenuated resistance to bile acids . The above data suggested the possibility that certain environmental cues might impose AmiA- and AmiC-dependent cell division in the inflamed gut . To address this issue , we examined the expression pattern of amiA and amiC genes . As S . Tm amiA and amiC expression is positively regulated by the CpxRA two-component system [57] , the bacterial strains were grown under CpxRA-inducible conditions; i . e . , high osmolarity ( here , 0 . 5 M NaCl ) [58] or the presence of antimicrobial peptide ( here , 1 μg/ml polymyxin B ) [47 , 59] , and subjected to microscopy analysis . When grown in LB broth containing 0 . 5 M NaCl , in all cell configurations , the proportion of chains of GFP-expressing ssaV amiC and GFP-expressing ssaV amiA amiC strains were dramatically increased ( Fig 8A and 8B ) . Similar results were obtained upon growth in LB broth containing polymyxin B ( S9 Fig ) . This was confirmed by a complementary experiment , showing that the introduction of a plasmid expressing AmiC in trans partially restored cell division ( Fig 8C ) . These results indicated that high osmolarity and the presence of antimicrobial peptides , which comprise a similar condition to that of the gut lumen , impose AmiA- and AmiC-dependent cell division . The previous results suggest that the chain form of the S . Tm amiA amiC mutant strain may confer impaired resistance to bile acids . To clarify this possibility , we examined a causal link between the chain form and bile acid resistance by comparing S . Tm cells grown in LB medium with those grown in LB plus 0 . 5 M NaCl . GFP-expressing S . Tm ssaV or ssaV amiA amiC strains grown in LB medium or LB plus 0 . 5 M NaCl were incubated with 1% deoxycholate , with the killing effect subsequently evaluated by using the membrane integrity indicator dye , propidium iodide ( PI ) . If the bacterial membrane was damaged by deoxycholate , S . Tm would not be able to express GFP , resulting in PI-stained cells . Consistent with the previous results of MICs ( Table 1 ) , GFP-expressing ssaV cells grown in both LB medium and LB plus 0 . 5 M NaCl were resistant to 1% deoxycholate ( Fig 9A and 9B ) . In contrast , approximately 20% of the ssaV amiA amiC mutant cells grown in LB medium were killed by 1% deoxycholate . Furthermore , the proportion of deoxycholate-killed ( PI-stained ) cells in the ssaV amiA amiC mutant cells grown in LB plus 0 . 5 M NaCl was significantly increased ( Fig 9A and 9B ) . These results suggest that the chain form of the S . Tm amiA amiC mutant strain is more susceptible to deoxycholate-mediated bactericidal effect than the single or paired cells , and that the chain form therefore likely confers attenuated resistance to bile acids on luminal S . Tm .
The Tat system is widely conserved in many bacterial pathogens and plays crucial roles in virulence [25] . Therefore , it is expected that the Tat system may represent a therapeutic target for bacterial pathogen infection . Earlier studies revealed that the S . Tm tatC mutant exhibits attenuated virulence [39] , consequent to an export defect of certain Tat substrates including AmiA and AmiC [40] . However , the lack of intestinal inflammation in the previously utilized Salmonella typhoid fever mouse model precluded determination of whether the Tat system is involved in Salmonella-induced enterocolitis . Thus , to clarify the role of the Tat system in Salmonella enterocolitis , in this study we utilized the streptomycin mouse model [28 , 41 , 60] in which infected S . Tm induces severe intestinal inflammation and colonizes the gut [28 , 60] . Our results showed that the S . Tm tatC mutant can elicit intestinal inflammation , albeit at a lower level as compared to that of the wild-type strain SL1344 . This appears to be correlated with the colonization levels in the cecum , which exhibited attenuated colonization of the S . Tm tatC mutant . As deletion of the tatC gene causes reduced expression of the hilA gene , which encodes a central regulator for ttss-1 [40] , the attenuated inflammation might be due to reduced expression of ttss-1 . Moreover , we also demonstrate here that gut colonization of the S . Tm tatC mutant is attenuated in a mixed infection experiment . Based on our data , we concluded that the Tat system in S . Tm is not essential for colitis but is involved in the induction of gut inflammation and colonization in the early infectious course . To date , numerous substrates ( approximately 30 proteins ) exported by the S . Tm Tat machinery have been identified and predicted according to the presence of Tat signal peptides [4] . Conversely , the S . Tm tatC mutant cannot export any substrates , thereby rendering this mutation likely to confer attenuated virulence . Earlier research showed that the attenuated virulence of the S . Tm tatC mutant strain in the Salmonella typhoid fever model could mainly be attributable to an export defect of AmiA , AmiC and SufI [40] . Here , we demonstrate that in the Salmonella enterocolitis model , the attenuated gut colonization of the S . Tm tatC mutant is attributable to an export defect of AmiA and AmiC . Moreover , single mutation of the amiA gene contributes little to the virulence , whereas combined mutations of amiA and amiC genes lead to significant attenuation of the gut colonization compared to that from single mutation of the amiC gene , suggesting that the roles of AmiA and AmiC in the virulence are redundant . In contrast , our results with complemented strains suggest that only AmiC is involved in certain phenotypes; i . e . , the cell division defect , attenuated motility , and high sensitivity to deoxycholate observed in the S . Tm amiA amiC mutant strain . Conceivably , a difference in localization of these amidases might cause such distinct roles . Although both AmiA and AmiC act in the periplasmic space , AmiC is specifically localized to the septal ring during cytokinesis [61 , 62] . This suggests that the attenuated virulence of the S . Tm tatC mutant may be mainly attributable to the loss of properly localized AmiC . Here , we demonstrate that the bactericidal action of bile acids is responsible for impaired luminal growth of the S . Tm amiA amiC mutant strain in the mouse colitis model . Bile acids act as a detergent that solubilizes fats in intestinal tract . In addition , another important role of bile acids in vivo is to affect the luminal bacterial community via potent antimicrobial properties that are mainly explained by membrane damage [63] . This apparently contributes to host mucosal defense against enteropathogens by bacterial killing . In contrast , certain enteropathogenic bacteria are known to be highly resistant to bile acids; for example , Salmonella spp . , E . coli , Campylobacter jejuni , Listeria monocytogenes , and Clostridium perfringens [63] . Therefore , it is tempting to suspect that the bile acid tolerance of enteropathogenic bacteria including S . Tm represents a critical factor that determines the competitive fitness advantage in the gut . Consistent with this , bile acid tolerance conferred by very long O-antigen contributes to the luminal fitness of S . Tm [51] . These observations therefore suggest that agents that cause the attenuation of bile resistance in enteropathogenic bacteria might constitute new promising antimicrobials . Our data show that the chain form of S . Tm cells confers attenuated resistance to bile acids , leading to impaired colonization in the inflamed gut . In the case of Gram-negative bacteria , robustness of the outer membrane plays a critical role in resistance to bile acids; thus , changes in membrane architecture and composition often bring about a bile resistance defect . For example , bacterial freezing , which causes structural damage of the outer membrane , leads to an increase in the susceptibility of E . coli to bile salts [64] . Furthermore , in Lactobacillus acidophilus , changes in fatty acids composition are related to the susceptibility to bile acids through the enhancement of lipid membrane stability [65] . In line with previous reports , our data from this study show that S . Tm phoP mutants containing an attenuated outer membrane barrier are highly susceptible to bile acids . In contrast , our data also show that even though the outer membrane barrier of the S . Tm amiA amiC mutant strain is as robust as that of the wild-type strain , this mutant strain displays impaired resistance to bile acids . These findings indicate that the robustness of the outer membrane might be not related to the bile acids resistance of S . Tm amiA amiC strain . Therefore , at present , it remains unclear why the chain form of the S . Tm amiA amiC mutant strain lacks the ability to resist killing by bile acids . Deciphering the mechanism underlying attenuated resistance to bile acids in the chain form of S . Tm thus deserves further investigation . Recently , vaccine-induced IgA has been shown to contribute to S . Tm elimination from the gut lumen by enchaining the pathogen [66] . This high-affinity IgA forms large monoclonal clumps by coating and cross-linking the S . Tm cells in the gut , resulting in attenuated tissue invasion and accelerated elimination of S . Tm . This accelerated elimination is attributable to an increase in the rate of clonal extinction . Based on this novel role of IgA [66] , we suspect that the S . Tm tatC or amiA amiC mutant may also be eliminated from the gut lumen by the same mechanism . Notably , the chain form of the S . Tm amiA amiC mutant strain is induced in the inflamed gut but not in normal gut , indicating that AmiA- and AmiC-dependent cell division is required in the inflamed gut . The question therefore arises regarding which environmental signal ( s ) in the inflamed gut impose the cell division on S . Tm . Our results from the in vitro experiments presented here indicate the possibility that high osmolarity and the presence of antimicrobial peptide induce AmiA- and AmiC-dependent cell division in the gut , as the gut luminal osmolarity is known to be quite high ( 0 . 3 M NaCl or higher ) [67] and antimicrobial peptides are constitutively present in the gut lumen [49] . Considering that during very high osmotic stress ( 1 . 2 M NaCl or higher ) , S . Tm undergoes filamentous growth in vitro accompanied with changes of outer membrane integrity [68] , osmotic stress may generally induce a cell division defect in this bacterium . Furthermore , the CpxRA envelope stress response , which is activated by high osmolarity and antimicrobial peptides [47 , 58 , 59 , 69] , has been shown in turn to activate transcription of the amiA and amiC genes in response to periplasmic stress , which occurs in the inflamed gut [47 , 57] , and contribute to S . Tm gut colonization during Salmonella-induced colitis [47] . These findings suggest that environmental stresses such as high osmolarity and antimicrobial peptide in the inflamed gut may elicit envelope perturbations of S . Tm , leading to an increase in amiA and amiC expression in a CpxRA-dependent manner . Although the induced peptidoglycan amidases likely allow S . Tm to cope with the environmental stress , the envelope perturbation is likely to induce a cell division defect . Deciphering the mechanism that specifically imposes AmiA- and AmiC-dependent cell division will be an important topic for future work . In conclusion , we demonstrate that the Tat system and the Tat-exported peptidoglycan amidases , AmiA and AmiC , constitute virulence factors in Salmonella-induced enterocolitis . In turn , they also represent promising therapeutic targets against Salmonella gut infection . Moreover , controlling bacterial cell shape by inhibiting certain types of cell division might constitute a new therapeutic intervention strategy against infection with bacterial pathogens .
All animal experiments were approved by the Kitasato University Institutional Animal Care and Use Committee ( Permit Number: A13-6 , J96-1 , J13-1 , 17–52 , 17–54 and 17–55 ) . Bacterial strains and plasmids used in this study are listed in Table 2 . S . Tm strain SL1344 is wild-type and a mouse virulent . S . Tm strains harboring chromosomal in-frame deletions were created using lambda/red homologous recombination system [70] . Primers used for construction of the mutant strains are listed in S1 Table . Complementary plasmids were constructed using DNA fragments containing amiA gene or amiC gene which amplified by PCR with primer sets: amiA-FW-SacI and amiA-RV-SphI or amiC-FW-SacI and amiC-RV-SphI , and S . Tm strains SL1344 chromosomal DNA as template , which were digested with SacI and SphI , and then ligated between the same sites of pMW118 , yielding to pamiA and pamiC respectively . Primers used for construction of complementary plasmid are listed in S1 Table . Animal infection experiments were performed in 6 to 12 week old mice as described previously [41 , 73] . C57BL/6 mice were maintained at the institute of experiments of animals at School of Pharmacy , Kitasato University or purchased from Japan SLC . To trigger artificial and S . Tm virulence independent colitis when required , a dextran sulfate sodium ( DSS ) was treated prior infection . In brief , sterile-filtered drinking water supplemented with 3 . 5% DSS ( molecular mass 36 , 000–50 , 000 g/mol , MP Biomedicals ) was provided to the mice ad libitum for 7 days . C57BL/6 mice were pretreated with 25 mg streptomycin 24 hours prior to infection . For infection , bacteria were grown for 12 h in LB medium containing 0 . 3 M NaCl supplemented with appropriate antibiotic ( s ) under mild aeration ( 160 rpm ) , diluted 1:20 and sub-cultured for 4 h in the same medium without supplementation of antibiotics . Bacteria were washed twice with PBS and mice infected with 5×107 CFU S . Tm strains by gavage . To determine bacterial population sizes , fecal pellets , cecal content , mLN and spleen were freshly collected in sterile PBS containing 0 . 5% tergitol , and subjected to bead-beating and plated on MacConkey agar plates ( Nissui Pharmaceutical ) supplemented with the appropriate antibiotics ( 50 μg/ml streptomycin; 50 μg/ml kanamycin; 10 μg/ml chloramphenicol ) . MacConkey medium including bile acids are suitable for determining the total number of S . Tm in the gut , even if the S . Tm strains display the impaired resistance to deoxycholate . This was demonstrated by comparison of growth on LB or MacConkey medium shows that the impaired resistance of S . Tm strains to deoxycholate had no effect on growth MacConkey medium ( S10 Fig ) . Furthermore , bead-beating and plating in the mouse infection experiments in this study are suitable for determining total S . Tm CFUs in the gut . This was verified by comparing the bacterial shapes of the long chained S . Tm cells with or without bead-beating using microscopy , showing that the long chained cell are sheared into single cells after bead-beating treatment ( S11A and S11B Fig ) . A CI was calculated by dividing the population size of background strains of S . Tm by its derivative mutants . Parts of cecal tissue were fixed in 4% formaldehyde ( Mildform , Wako Pure Chemical Industries , Ltd . ) and embedded in paraffin . Cryosections were prepared and air-dried , and then stained hematoxylin/eosin ( H&E ) . To determine the degree of inflammation , pathological score was monitored as previously described [41] , evaluating submucosal edema , polymorphonuclear leukocyte infiltration , goblet cell numbers , and epithelial damage , a total score of 0–13 . More than 3 scores are considered as a sign of inflammation . To reduce luminal bile acids which interact with S . Tm in the gut , C57BL/6 mice were fed a normal rodent chow supplemented with the bile acid sequestrant cholestimide resin ( 1 . 5% , Mitsubishi Tanabe Pharma ) . Fecal pellet collected at the indicated time points were homogenized , and diluted in PBS . The resulting dilutions were then analyzed using the mouse lipocalin-2 ELISA duoset ( R&D ) according to the manufacturer’s instructions . S . Tm strains were grown overnight in LB at 37°C , diluted 1:100 in fresh LB broth , or LB containing 0 . 5 M NaCl or 1 μg/ml polymyxin B , and grown for 2 . 5 h . The resulting bacteria were placed on a 1 . 5% agarose pad , sealed under a glass coverslip , and imaged at 400× using the Zeiss Axiovert A1 microscope . S . Tm strains grown in LB containing 0 . 5 M NaCl were harvested , resuspended in 2 . 5% glutaraldehyde in PBS . The samples were post-fixed with 2% osmium tetroxide for 3 h at 4°C , then dehydrated through a series of ethanol concentrations . Specimens were critical-point dried using carbon dioxide . The samples were coated with osmium plasma and examined at 5 kV accelerating voltage in a JSM-6320F SEM . Fecal pellet was suspended gently in PBS . The resulting suspension was placed on a 1 . 5% agarose pad , sealed under a glass coverslip , and imaged at 400× using the Zeiss Axiovert A1 microscope . S . Tm strains grown overnight in LB at 37°C , subcultured in fresh LB broth and further grown for 2 h . A 5-μl aliquot at an OD600 of 1 . 0 was placed on 0 . 3% agar LB plate ( for swimming ) or 0 . 5% agar LB plate supplemented with 0 . 5% glucose ( for swarming ) , and left for 5 min . The plates were incubated at 37°C for 5 h ( swimming ) or 10 h ( swarming ) . S . Tm strains grown to the logarithmic growth phase were diluted to 1×106 CFU per ml with different concentrations of magainin 2 ( LKT Laboratories , Inc . ) or deoxycholate ( Nacalai tesque ) in sterile LB broth , and incubated for 15 h at 37°C . A positive control contained no antimicrobials whereas in the negative control , S . Tm cells were not present . After incubation , the A595 values were determined using a microplate reader ( Bio-Rad ) . MICs were determined as the lowest concentrations of antimicrobials that were shown to prevent bacterial growth by more than 50% in comparison with the growth of the positive control . Outer membrane permeability was evaluated by EtBr influx assay [53] . S . Tm grown to the stationary growth phase was washed with PBS , and diluted to OD600 = 0 . 4/ml in PBS . The resulting bacterial suspensions were mixed with EtBr ( 24 μM ) , followed by measuring the fluorescence signal intensity of the EtBr-nucleic acid complex using SpectraMax M5 spectrofluorometer ( Molecular Devices ) with excitation and emission wavelengths of 544 and 590 nm , respectively . Fecal pellets were weight , homogenized in ethanol . Total bile acids in feces was extracted by hot ethanol method , and analyzed using Total bile acids-Test wako ( FUJIFILM Wako Pure Chemical ) according to the manufacturer’s instructions . S . Tm strains were grown overnight in LB at 37°C , diluted 1:100 in fresh LB broth or LB containing 0 . 5 M NaCl , and grown for further 2 . 5 h . The bacteria were mixed with 1% deoxycholate in PBS , incubated at 37°C for 20 min . After the incubation , propidium iodide ( PI ) solution ( Dojindo ) was added , and incubated further 5 min at room temperature . PI-stained S . Tm cells were counted by fluorescence microscopy . Statistical significance was determined by Mann Whitney U-test or Student t-test using the software Graphpad Prism . P values of less than 0 . 05 were considered significant ( *P < 0 . 05; ** P < 0 . 01; *** P < 0 . 001 ) .
|
For proteins residing outside the bacterial cytoplasm , transport is an essential step for adequate function . The twin-arginine translocation ( Tat ) system enables the transport of folded proteins across the cytoplasmic membrane in prokaryotes . It has recently become clear that this system plays a pivotal role in the detrimental effects of many bacterial pathogens , suggesting Tat as a novel therapeutic target against their infection . In particular , the bacterial enteropathogen Salmonella Typhimurium causes foodborne diarrhea by colonizing the gut interior space . Here , we describe that the S . Typhimurium Tat system contributes to intestinal infection by facilitating colonization of the gut by this pathogen . We also identify that two Tat-exported enzymes , peptidoglycan amidase AmiA and AmiC , are responsible for the Tat-dependent colonization . S . Typhimurium strains having nonfunctional Tat systems or lacking these enzymes undergo filamentous growth in the gut interior owing to defective cell division . Notably , this chain form of S . Typhimurium cells is highly sensitive to bile acids , rendering it less competitive with native bacteria in the gut . The data presented here suggest that the Tat system and associated amidases may comprise promising therapeutic targets for Salmonella diarrhea , and that controlling bacterial shape might be new strategy for regulating intestinal enteropathogen infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"cell",
"cycle",
"and",
"cell",
"division",
"drugs",
"immunology",
"bile",
"cell",
"processes",
"light",
"microscopy",
"microbiology",
"salmonellosis",
"animal",
"models",
"bacterial",
"diseases",
"mutation",
"streptomycin",
"model",
"organisms",
"signs",
"and",
"symptoms",
"microscopy",
"antibiotics",
"experimental",
"organism",
"systems",
"pharmacology",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"mutant",
"strains",
"inflammation",
"animal",
"studies",
"fluorescence",
"microscopy",
"mouse",
"models",
"immune",
"response",
"diagnostic",
"medicine",
"anatomy",
"cell",
"biology",
"physiology",
"genetics",
"microbial",
"control",
"biology",
"and",
"life",
"sciences"
] |
2018
|
Tat-exported peptidoglycan amidase-dependent cell division contributes to Salmonella Typhimurium fitness in the inflamed gut
|
Ryanodine receptors ( RyRs ) are ion channels that regulate muscle contraction by releasing calcium ions from intracellular stores into the cytoplasm . Mutations in skeletal muscle RyR ( RyR1 ) give rise to congenital diseases such as central core disease . The absence of high-resolution structures of RyR1 has limited our understanding of channel function and disease mechanisms at the molecular level . Here , we report a structural model of the pore-forming region of RyR1 . Molecular dynamics simulations show high ion binding to putative pore residues D4899 , E4900 , D4938 , and D4945 , which are experimentally known to be critical for channel conductance and selectivity . We also observe preferential localization of Ca2+ over K+ in the selectivity filter of RyR1 . Simulations of RyR1-D4899Q mutant show a loss of preference to Ca2+ in the selectivity filter as seen experimentally . Electrophysiological experiments on a central core disease mutant , RyR1-G4898R , show constitutively open channels that conduct K+ but not Ca2+ . Our simulations with G4898R likewise show a decrease in the preference of Ca2+ over K+ in the selectivity filter . Together , the computational and experimental results shed light on ion conductance and selectivity of RyR1 at an atomistic level .
Muscle contraction upon excitation by nerve impulse is initiated by a rapid rise in cytoplasmic Ca2+ . In skeletal muscle , the rise in cytoplasmic Ca2+ is brought about by the opening of the ryanodine receptor ( RyR1 ) , which releases Ca2+ from intracellular stores [1] , [2] . RyRs are large tetrameric ion channels ( molecular weight of 2 . 26 MDa ) present in the membranes of endoplasmic/sarcoplasmic reticulum . They have high conductance for monovalent ( ∼800 pS with 250 mM K+ as conducting ion ) and divalent cations ( ∼150 pS with 50 mM Ca2+ ) , while being selective for divalent cations ( PCa/PK∼7 ) [3] . RyRs are important mediators of excitation-contraction coupling and congenital mutations of RyRs result in neuromuscular diseases such as malignant hypothermia and central core disease ( CCD ) [4] . Although RyRs are physiologically important , the molecular basis of their function is poorly understood . RyRs have unique properties such as their modes of selectivity and permeation not seen in other ion channels with known structures . Next to the putative selectivity filter ( 4894GGGIG ) , there are two negatively charged residues ( D4899 and E4900 ) in RyR1 that are essential for normal selectivity and conductance [5] . K+ channels have an analogous selectivity filter , but in contrast to RyR1 , have only one adjacent negative residue that is not even conserved while other Ca2+ channels have only one conserved negative residue in the equivalent position [6] . In the selectivity filter , mutations result in non-functional channels [4] leading to CCD . A structural model of the pore region that would reveal the location and function of these residues will be useful in understanding the role of these residues in channel function . An early model of RyR ion permeation postulated potential barriers within the pore corresponding to three putative binding sites [7] . Without any knowledge of the structure of the pore , the model was able to quantitatively reproduce conductance data of various ions . A PNP-DFT ( Poisson Nernst Planck-Density Functional Theory ) model [8] accurately modeled the role of residues D4899 and E4900 in RyR1 in generating the high ion conductances of RyRs established by mutagenesis [5] , [9] . Selectivity was attributed to charge-space competition , as Ca2+ could accommodate the most charge in least space compared to K+ . However , since the channel model used in these simulations relied on a fixed structure , it could not predict changes due to mutations that potentially alter the structure of the channel . Additionally , there are two homology models of the RyR pore region [10] , [11] based on KcsA , a bacterial K+ channel whose solution structure is known [12] . Shah et al . [11] used bioinformatics approaches to construct models for RyR and the related inositol triphosphate channel . The luminal loop in their RyR model begins at 4890G resulting in the selectivity filter being 4890GVRAGG . However , mutagenesis has shown that residues I4897 , G4898 , D4899 and E4900 are important for channel conductance and selectivity , which suggests that they are part of the conduction pathway of RyR1 resulting in the predicted selectivity filter being 4894GGGIGDE . Welch et al . also constructed a homology model for the cardiac ryanodine receptor ( RyR2 ) using the structure of the KcsA channel [10] and performed simulations to identify residues important for channel function . Their simulations failed to identify D4899 as an important residue for ion permeation contrary to what has been shown experimentally [5] . Furthermore , cryo-electron microscopy ( cryo-EM ) of RyR1 ( which has revealed the pore structure at highest resolution yet ) revealed significant differences between the pore region of KcsA and RyR1 [13] . Experimental structure determinations of the RyRs have been mainly performed by cryo-EM [14]–[17] . These studies revealed conformational changes that accompany channel opening [18] and the binding sites of various effectors of RyRs [19]–[21] . Cryo-EM has a resolution of ∼10–25 Å and thus is able to provide only limited structural information regarding the pore structure . Samso et al . [22] manually docked the KcsA pore structure into the transmembrane region of their cryoEM map of the intact closed RyR1 . Furthermore , they predicted the presence of at least 6 transmembrane helices from simple volumetric constraints . Ludtke et al . [13] identified several secondary structure elements in their ∼10 Å resolution cryo-EM map of the closed RyR1 . The pore-forming region as visualized by Ludtke et al . consists of a long inner helix made up of 31 residues and a pore helix made up of 15 residues that are presumably connected by a long luminal loop made up of 24 residues . Since the structure is derived from cryo-EM , the positions of pore residues' side chains and the structure of loops connecting the helices are unknown . We build a molecular model of the pore region of RyR1 based on their cryo-EM study by adding the luminal loop and the missing side chains of residues forming the helices of the pore . Furthermore , in our molecular dynamics simulations we examine the interactions of the pore region with mono- and divalent cations known to permeate the channel ( Table 1 ) .
We present in Figure 1 an atomistic model of the pore-forming region of the tetrameric RyR1 constructed from cryo-EM data [13] . Figure 1A shows two of the four inner helices and pore helices connected by a long luminal loop . Site-directed mutagenesis [22] , [23] predicts that RyR1 has a selectivity filter ( 4894GGGIGD ) , which is analogous to K+ channels ( Figure 1B ) . Since the pore helix immediately precedes the selectivity filter , we assign M4879-A4893 to the 15-residue pore helix . In K+ channels , the sequence GXXXXA in the inner membrane-spanning helix has been proposed to form the gating hinge [24] . The analogous glycine in RyR1 occurs in the 4934 position , which determines the sequence of the 40-residue inner helix as I4918-E4948 . Thus , the pore corresponds to residues M4879-E4948 [13] , which includes the putative selectivity filter . We construct the luminal loop by constraining the diameter of the selectivity filter at its luminal edge to 7 Å [25] ( Figure 1B ) . In this model , the acidic residues important for channel function , D4899 and E4900 are located at the mouth of the pore and at the beginning of the selectivity filter . Both the cytoplasmic and luminal faces of the pore are highly negatively charged ( as seen in the surface representation in Figure 1C and 1D ) , which may allow cations to concentrate around the mouths of the pore . We also predict the negatively charged faces to exclude anions , which RyR1 is known not to conduct . Interestingly , with the exception of L4935 and L4943 , the hydrophobic residues lining the helices face away from the water filled pore . In K+ channels , hydrophobic residues facing the pore are known to perform important functional roles ( like stabilization of the inactivation gate [26] ) . The model exhibits structural similarities with the K+ channel MthK such as the positioning of the pore helix and the inner helix and the bending of the inner helix [27] , although in RyR1 the pore is significantly wider . To model the interactions of RyR1 pore with ions and elucidate sites of high ion occupancy along the pore , we perform molecular dynamics simulations of RyR1 ( see below ) . We describe here some of the experimental characteristics of RyR1 mutants D4899Q and G4898R , which are both present in the selectivity filter . We compare in Figure 2 the ion permeation properties of wild type RyR1-WT [5] with RyR1-D4899Q [5] and CCD associated RyR1-G4898R [28] mutant channels . Proteoliposomes containing purified 30S channel complexes were fused with planar lipid bilayers and single channel currents were recorded in 250 mM KCl on both sides of the bilayer . Figure 2 shows representative single channel traces in presence of 2 µM cis ( SR cytosolic ) Ca2+ ( left , upper traces ) and following the subsequent addition of 10 mM trans ( SR luminal ) Ca2+ ( left , lower traces ) . In presence of 2 µM Ca2+ , WT and the two mutant channels showed rapid transitions between open and closed channel states . Reduction in cis Ca2+ from 2 µM to 0 . 01 µM reduced channel open probability for WT and D4899Q close to background levels ( data not shown ) . In contrast , single channel activities of RyR1-G4898R did not respond to a change in cytosolic Ca2+ concentration from 2 µM to 0 . 01 µM ( not shown ) . Ion currents through WT and mutant channels showed a linear voltage-dependence but differed in their magnitude ( Figure 3 , right panel ) . Averaged single channel conductances were 801 pS for WT , 164 pS for RyR1-D4899Q , and 352 pS for RyR1-G4898R . The Ca2+ selectivity of WT and the two RyR1 mutants was determined by recording current-voltage curves in 250 mM symmetrical KCl with 10 mM Ca2+ in the trans bilayer chamber . At 0 mV in presence of 10 mM trans Ca2+ , WT exhibited averaged unitary Ca2+ current of −2 . 4 pA compared with −0 . 4 pA for RyR1-D4899Q and ∼0 pA for RyR1-G4898R . Addition of 10 mM Ca2+ to the trans chamber reduced single channel currents of WT and D4899Q at negative and positive potentials and the averaged reversal potentials ( Erev ) for WT and D4899Q were shifted by +9 . 5 mV and +1 . 9 mV , respectively ( Figure 2 , right panel ) . Applying constant field theory , a permeability ratio of Ca2+ over K+ ( PCa/PK ) of 7 . 0 and 1 . 0 is calculated for WT and D4899Q , respectively ( Table 2 ) . In contrast , addition of 10 mM trans Ca2+ did not generate a noticeable unitary Ca2+ current at 0 mV , and had no effect on ion currents or reversal potential of RyR1-G4898R . Taken together , the single channel data of Figure 2 indicate that the D4899Q mutation decreases K+ conductance and ion selectivity for Ca2+ over K+ compared to WT , without eliminating Ca2+ responsiveness . In contrast , the CCD associated G4898R mutation abolished Ca2+ responsiveness , Ca2+ permeation , and reduced ion conductance demonstrating that the G4898R mutation introduced major global conformational changes in RyR1 . Acidic residues lining the pore of the RyR channel have been assumed to be deprotonated at physiological pH . In support of this , site directed mutagenesis resulting in charge neutralization of acidic residues reduced ion conduction and selectivity [5] . Single channel experiments in the pH range of 6 . 5–9 . 4 ( on the luminal side ) on RyR1-WT were performed to probe the protonation status of luminal residues D4899 and E4900 . Lack of a significant effect of a change of pH on K+ ion conductance and the permeability ratio of Ca2+ over K+ ( Xu and Meissner , unpublished studies ) indicate that the protonation status of D4899 and E4900 remained unchanged . Hence we assume in our MD simulations that these residues are deprotonated . To model the interactions of RyR1 pore with ions , we perform molecular dynamics simulations of the RyR1 pore tetramer with 70 mM CaCl2 and 250 mM KCl . Ca2+ is present in the solution only on the luminal side . We plot the histogram of ion occupancies in the pore against the pore axis in Figure 3 . Ca2+ shows highest occupancy at 77 Å along the axis of the channel , which corresponds to the position of D4899 in the selectivity filter ( Figure 3C ) . We find the preferential occupancy ratio , R for the CaCl/KCl simulations to be 11 . 3±5 . 6 ( Table 3 ) . These results show a clear preference of Ca2+ over K+ in the selectivity filter . We hypothesize that this preferential localization of Ca2+ in the first binding site along the path from luminal to cytosolic side may play an important role in channel selectivity . To ensure that this preferential localization of Ca2+ is not due to selective exclusion of K+ ions in the filter , we perform simulations of the pore with only KCl present . These simulations show K+ occupancy to be highest at 81 Å ( Figure 3B ) and the total charge of the ions in the selectivity filter ( Table 3 ) to be similar to simulations with CaCl/KCl , which shows that the binding sites are amenable to K+ in absence of Ca2+ . As a control , we perform simulations with NaCl/KCl with the same starting configuration as the CaCl2/KCl simulations . We observe a decreased occupancy of Na+ when compared to Ca2+ . Na+ is still preferred over K+ in the selectivity filter , if we take relative concentration of NaCl and KCl into consideration with R = 2 . 2 ( Figure S1 ) . Simulations with only CaCl2 show similar results ( Figure 3D ) . The occupancy data qualitatively reflects the selectivity of RyR1 pore , which is in the order of Ca2+ being much greater than Na+ which is slightly greater than K+ [7] . Radial distribution functions ( RDF ) of the ions around the carboxyl oxygens of D4899 and E4900 reflect the affinities of various ions to the binding site . We plot the number of ions found at a distance r from the carboxyl oxygens of D4899 and E4900 , to confirm that the peaks that we observe in ion occupancy plots is due to localization of the ions near these residues . From the RDFs , we find that Ca2+ ions exhibit the highest affinity for D4899 and E4900 followed by Na+ and then K+ ( Figure 4A and 4B ) . It is experimentally observed that the conductance of the channel is highest for K+ , followed by Na+ and then Ca2+ [7] . These two observations lead us to postulate that that tighter binding results in ions spending more time in the channel , thus giving rise to lower ion currents . The higher occupancy of Ca2+ compared to K+ can be either primarily attributed to electrostatics ( due to the presence of 8 acidic residues in the vicinity ) or to electrostatics combined with the structure formed by the selectivity filter . By calculating the occupancy of ions inside the pore region corresponding to the selectivity filter ( formed by 4894GGGIGDE of the tetramer , as shown in Figure 1B ) , we also consider if the pore structure contributes in concentrating the ions inside the pore . Although eight negative charges in a confined space would be more selective for Ca2+ than K+ as determined by a charge/space model [29] , we find that in an all-atom model of RyR1 pore region , the selectivity filter is able to sample structures that support preferential localization of Ca2+ inside the pore ( in the region corresponding to the selectivity filter ) . Thus , the preferential localization of Ca2+ over K+ is due to both the pore structure and electrostatics . To confirm the preferential Ca2+ binding in the selectivity filter , we perform MD simulations on two experimentally characterized RyR1 mutants , D4899Q [5] and G4898R [28] . In single channel experiments , the RyR1-D4899Q exhibits a decreased selectivity for Ca2+ and lower K+ conductance ( Table 2 ) . The experimentally observed decreased selectivity of Ca2+ over K+ in RyR1-D4899Q can be rationalized from our simulations . In simulations involving D4899Q , the ion occupancy histograms show that the peaks for K+ and Ca2+ are outside the selectivity filter compared to that of the RyR1-WT ( Figure 5D ) . We observe that the preference for Ca2+ over K+ in the selectivity filter decreases to R = 3 . 1±1 . 6 from R = 11 . 3±5 . 6 in RyR1-WT . Further , the total ionic charge in the selectivity filter is reduced ( 4 . 3 on an average compared to 7 in RyR1-WT ) . RDF of ions around Q4899 ( Figure 4C ) strongly suggests that there is no binding of K+ and Ca2+ at the site of mutation , as also seen in Figure 5D . In simulations of RyR1-D4899Q , the RDFs and ion histogram taken together prove that Ca2+ and K+ bind only at the end of the selectivity filter especially near E4900 . Transition of an ion from a fully hydrated state to inside the channel where hydration is low is facilitated by presence of high affinity binding sites near the entrance of the channel . According to our structural model of the RyR1 pore , D4899 and E4900 residues are present at the entrance of the channel and form the first binding site of ions as they traverse from the luminal side to the cytosolic side . Hence , mutation of glutamic acid to asparagine resulting in a net loss of 4 negative charges weakens the initial binding event of cations as they enter the pore , which would result in the decrease of overall conductance of cations . Another important observation is that when just one of the two acidic residues in the selectivity filter is neutralized , Ca2+ binding is affected much more than K+ , whose binding remains the same as RyR1-WT ( Table 3 ) . These results suggest the requirement of higher magnitude of negative charges in the selectivity filter to bind Ca2+ efficiently . Simulations with RyR1-G4898R show a decrease in preference of Ca2+ in the selectivity filter ( Figure 5B ) . The preferential occupancy ratio of Ca2+ over K+ ions ( R = 4 . 5±0 . 9 ) is lesser than RyR1-WT ( R = 11 . 3±5 . 6 ) . The highest occupancy of Ca2+ ions occurs near the edge of the selectivity filter similar to that of RyR1-D4899Q , which implies an exclusion of ions from the selectivity filter . The exclusion of Ca2+ ions is likewise seen from the RDF of ions around D4899 in RyR1-G4898R that shows decreased affinity of Ca2+ to D4899 compared to wild type ( Figure 4D ) . Thus , the introduction of a basic residue next to the acidic residues of the selectivity filter decreases ion binding in the selectivity filter , with the effect on Ca2+ being greater than on K+ .
Elucidation of the structure-function relationship in RyR1 necessitates an atomistic model of its pore region . We constructed a model of the pore region that identifies the positions of residues critical to channel function . Furthermore , molecular dynamics simulations help us confirm the potential binding sites of ions along the pore . Considering the permeation time for different ions in RyR , our molecular dynamics simulations cannot sample statistically significant number of permeation events . However , the correlation between preferential ion occupancies seen in our simulations and experimentally measured selectivity both in RyR1-WT and its mutants suggests that preferential Ca2+ binding to the selectivity filter is a necessary but not a sufficient condition for selectivity . The charge space competition ( CSC ) model [30]–[32] provides one explanation for the selectivity for Ca2+ over K+ and Na+ . The model attributes Ca2+ selectivity to the ability of Ca2+ ( with a higher charge ) to neutralize the carboxylate rich selectivity filter of Ca2+ channels by occupying the same space as Na+ ( or lesser space than K+ ) . The ion occupancies seen in the selectivity filter in our simulations agree well with the CSC model for RyR1 [8] , [29] . An important consequence of both models is the identification of sites on the pore ( 4899D and 4900E ) that have preferential affinity for Ca2+ compared to K+ and are experimentally shown to be sensitive to mutations . Even though the cryoEM studies of Ludtke et al . [13] provide strong evidence for the structure and the orientation of helices forming the pore region , assigning the right sequence to this structure is essential for our simulations . Several biochemical and bioinformatics studies have predicted the pore-forming region of RyR1 with good agreement between each other . Balshaw et al . [33] first proposed that the pore forming region in RyR1 is located around 4894GGGIGD due to the striking similarity between this sequence and the selectivity filter of K+ channels including KcsA . Zhao et al . [22] performed several mutations on this highly conserved region in RyR2 and observed dramatic effects on ion conduction properties . Gao et al . [23] performed functional studies on similar mutants in RyR1 that also highlighted the importance of these residues in ion conduction . Using the putative selectivity filter as an anchor and the predicted positions of the membrane spanning helix from hydrophathy plots , Welch et al . constructed a homology model of the RyR2 pore region from KcsA [10] . In K+ channels , the sequence GXXXXA in the inner membrane spanning helix has been proposed to form the gating hinge [24] . The analogous glycine in RyR1 occurs in the 4934 position , in center of the inner helix predicted by hydropathy plots [34] . Studies on triadin ( a transmembrane protein known to interact with RyRs [35] ) indicate that three of the acidic residues of RyR1 namely , D4878 , D4907 and E4908 are essential for binding of triadin to RyR1 [36] . D4878 is located in the luminal loop that connects the pore helix to the rest of the transmembrane region of RyR1 . D4907 and E4908 occur in the luminal loop connecting the pore helix to inner helix and are positioned after the selectivity filter . Binding of triadin in the luminal region of RyRs inhibits channel function experimentally , which can be inferred from our model too with respect to the positions of D4907 and E4908 . Thus , the sequence assignments in our model are in good agreement with results of biochemical studies . Although the cryo-EM data for the pore-region of RyR1 is obtained from the closed state of the channel , the inner helices of RyR1 resembles that of the potassium channels in the open state ( MthK ) . The similarity in the inner helices of RyR1 and MthK implies that the inner helices of RyR1 need not undergo major conformational change during transitions from closed to open state . The structure of the selectivity filters of the closed and open states of potassium channels ( MthK and KcsA ) are essentially the same , which suggests that RyR1 selectivity filter may be modeled even from its overall closed state . Moreover , the results of our simulations are in good agreement with experimental data despite modeling the pore-structure from an overall closed state of the channel . The residues forming the selectivity filter ( GGGIGDE motif ) are not the only determinants for selectivity and high conductance . There are acidic residues present in the inner helix , towards the cytosolic side , D4938 and D4945 whose mutation as predicted by PNP/DFT is experimentally shown to reduce K+ conductance and selectivity [8] , [9] . The peaks in the K+ and Ca2+ occupancies in the cytosolic vestibule are found near the positions of these residues ( Figure 4 ) . Thus there are two regions in the pore that have high affinity for ions and are known to determine selectivity . One region is present in the luminal side along the selectivity filter , while the other is present in the cytoplasmic side of the pore . The presence of negatively charged sites on either side of the channel seen in our structural model is comparable to nicotinic acetylcholine receptor ion channel [37] . RyR1-G4898R is not responsive to Ca2+ and caffeine and does not bind ryanodine unlike RyR1-WT which points to major altered protein conformation of the mutant channel [28] . In single channel measurements , RyR1-G4898R mutant is constitutively open to K+ and it loses Ca2+ conductance and regulation by pharmacological agents . Since we model only the pore region , the present study cannot predict the global structural changes that occur due to G4898R mutation . However , the local structural changes in the selectivity filter and the consequent loss of ion binding as seen in our simulations could account for our experimental observations on RyR1 mutants . The selectivity filters of ion channels are highly dynamic as evinced experimentally by the flickering of ion currents in single-channel measurements and structurally by the preponderance of glycines . The long , polar side chain of arginine in the selectivity filter could interact with other regions of the channel resulting in a selectivity filter that is structurally different from RyR1-WT This distortion may render the selectivity filter in a conformation that does not allow Ca2+ conduction . In contrast to RyR1-G4898R , RyR1-D4899Q shows Ca2+ and caffeine dependent Ca2+ release and ryanodine binding comparable to RyR1-WT [5] . Maintenance of activity suggests that changes upon mutation are localized to the selectivity filter and our simulations can identify these changes . Our simulations identify ab initio the preferential localization of Ca2+ ions near the side chain carboxyl groups of D4899 , E4900 , D4938 and D4945 . This result is achieved without any prior knowledge of the ion positions with respect to the selectivity filter since the initial positions of ions in the simulations are all random . Without prior bias , we are able to reproduce Ca2+ occupancies up to 11 . 3 times higher than that of K+ , which supports the accuracy of the model of a Ca2+ selective channel . RyRs are important players in excitation-contraction coupling , which is fundamental to muscle contraction for movement and heart function . A mechanistic model of RyR1 will help us understand not only a functionally unique ion channel , but also shed light on an important physiological process .
Our model is confined to the pore-forming region of the homotetrameric RyR1 channel . Both site directed mutagenesis and cryo electron microscopy ( cryoEM ) [13] have suggested that the RyRs have a pore architecture similar to K+ channels whose structure has been determined . Single particle cryo-EM studies on RyR1 detected several helix-like densities using the program SSEhunter [38] , [39] . SSEhunter quantitatively identifies densities in a cryoEM map that may represent secondary structure elements and outputs the length and orientation of the secondary structures that can be unambiguously identified from the cryoEM map . This tool has been validated in many studies [40]–[42] . Two of these helices in each subunit , a long membrane spanning helix kinked in the middle ( inner helix ) and a short helix in the luminal side of the membrane ( pore helix ) face each other and form the backbone of the channel pore . At the resolution of the cryo-EM densities the side chains of the helices could not be resolved . The coordinates of the backbone atoms of the pore helix and the inner helix were obtained from the cryo-EM studies [13] . The kink in the middle of the inner helix ( G4934 ) was proposed to be analogous to the gating hinge of MthK channel [27] . Sequence comparison indicates that 4894GGGIG motif is analogous to selectivity filter motif T[VI]GYG of K+ channels . Site directed mutagenesis suggests that 4899DE motif is also part of the selectivity filter of the RyRs . The sequence used in constructing the pore region corresponds to M4879–E4948 ( Swissprot ID: P11716 ) [13] . The amino acid sequence corresponding to M4879-A4893 is assigned to the pore helix while the sequence I4918-E4948 is assigned to the inner helix . In this sequence assignment , RyR1 has a long , 24-residue luminal loop ( G4894-D4917 ) , which is not visible in the cryo-EM reconstruction . Initial structure of the luminal loop is obtained by searching a database of loop structures found in SYBYL ( Tripos , CA ) . To remove steric clashes between the loop and the helices , we further refine the loop conformation using the MODLOOP server [43] , which predicts the loop conformations by satisfying spatial restraints . Using constrained all atom discrete molecular dynamics ( DMD ) [44] , we perform simulations on the whole pore-forming region to constrain the distance between carboxyl oxygens of D4899 in the opposite monomers around 7 Å [25] . Finally , we optimize the rotamer states of all side chains using MEDUSA [45] , a molecular modeling and design toolkit . Thus , we have directly used the structure of the helices identified by SSEhunter and the loop structure created using molecular modeling and low-resolution experimental constraints to create the final model of the pore structure of RyR1 . We perform molecular dynamics simulations using GROMACS [46] , [47] with the OPLSAA force field [48] modified with additional parameters for lipids [49] and ion parameters of Aqvist [50] . The RyR1 pore is placed in a pre-equilibrated DPPC bilayer with explicit solvation using genbox program in GROMACS , which removes all the water and DPPC molecules that have clashes with the pore as it is placed in the bilayer . We use the SPC model for water [51] . The simulation system shown in Figure S2A consists of the pore-forming tetramer , 405 DPPC molecules , ∼14000 water molecules and ions to make a neutral system . The concentrations of K+ and Ca2+ used in different simulations are shown in Table 1 . Berendsen weak temperature coupling [52] is used with a relaxation constant of 0 . 5 ps . A cut-off of 10 Å is used for Van der Waals interactions and long range electrostatics is treated with Particle Mesh Ewald [53] with a grid-spacing of 12 Å and a cutoff of 10 Å . We perform the simulations at constant volume , with a vacuum of 2 nm thickness at the top and bottom of the simulation box , to maintain asymmetry of ion concentrations at the cytoplasmic and luminal side even in the presence of periodic boundary conditions [54] . In order to ensure that the introduction of a 2 nm slab of vacuum above and below our simulation system will not affect our results , we perform one set of simulations of RyR1-WT with KCl and CaCl2 by replacing the 2 nm slabs of vacuum with water molecules . We calculate the ion occupancy ratios and the histogram of ion occupation along the channel axis ( Table S1 and Figure S3 ) . These results do not change the primary conclusions of this study . We observe a preferential occupation of Ca2+ over K+ in the selectivity filter with a ratio of 6 . 9 ( within 1 SD of simulations with the vacuum slabs ) . However , due to the periodic boundary conditions , replacing the vacuum slab with water eliminates the partition between the cytosolic and luminal compartments , which then removes the partition between what was originally two compartments , and hence the concentration gradient . Since the pore forming region is a small part of the entire membrane-spanning domain , the dynamics of the helices would depend on interactions with other regions of the membrane-spanning domain , which is not included in the present model . Hence , we focus on studying the interactions of the ions with the luminal loop and also the dynamics of the selectivity filter and the luminal loop . Therefore , harmonic restraints with force constants of 1000 kJmol−1 nm−2 are imposed on the C , Cα , N , and O atoms of the pore lining helices and the pore forming helices during the simulations . In the complete channel , the pore forming helix may not be in direct contact with the membrane , but to ensure minimal contact of water with the membrane spanning , pore forming helix , we surround the pore forming helix with lipids . Applying harmonic constraints on pore forming region can have dramatic effects on observed ion binding events [55] . However , in our simulations , the harmonic restraints is placed only on the pore helix and the inner helix , while the luminal loop ( containing the selectivity filter ) is unrestrained . The narrowest region of the pore is formed by the luminal loop , which is flexible in our simulations . The cytoplasmic side of the pore , like MthK has a much bigger volume and we predict the movement of the helix to be minimal in the time scale of our simulations and hence potential artifacts on ion binding due to the restraints on the protein should be minimal . The simulation system is first subjected to 1000 steps of steepest descent energy minimization . We then perform an equilibration run for 5 ns where the protein is restrained , while the lipids , water molecules and the ions are allowed to equilibrate around the protein . A 15 ( or 25 ) ns production run is initiated after the equilibration run . The trajectory used for analysis corresponds to 5–15 ( or 5–25 ) ns of the production run . We carry out simulations in many ionic conditions , namely CaCl2 , KCl , NaCl , CaCl2+KCl , NaCl+KCl . The concentrations of ions used are shown in Table 1 . Simulations are also performed on two pore mutants: RyR1-G4898R , RyR1-D4899Q . We calculate ion occupancy as a function of z coordinate ( along the axis of the pore ) by counting the number of ions in the shaded region shown in Figure S1B . We calculate the histogram of ion occupancy with a bin width of 1 Å , which provides a picture of the ion binding sites and ion occupancies along the axis of the channel . To quantify the preferential localization of the ions , we calculate the ratio of number of each type of ion in the selectivity filter , while accounting for the differential concentration of each ionic species . We denote this ratio as R:where nsi is the number of ions of type i present in the selectivity filter , nti is the total number of ions of type i present in the luminal side . Since there are no ion translocation events during our simulations , nt remains constant throughout the simulation time . ns is calculated from the area under the curves shown in Figures 3 and 5 , for each corresponding simulation . To determine the specific binding locations of the ions , we calculate radial distribution functions ( RDFs ) of the ions around the side chain carboxyl oxygens of D4899 and E4900 of the selectivity filter . To confirm that the system is equilibrated , we perform our analysis on shorter stretches ( 5 ns ) of our trajectories and them with analysis performed over the whole length of the trajectories . The ion-occupancy histograms and the RDFs of the shorter stretches of the trajectory are similar to the longer complete trajectory , which confirms that the simulations are well equilibrated ( data not shown ) . Single channel measurements are performed using planar lipid bilayer method [5] . Proteoliposomes containing the purified recombinant RyR1s are added to the cis ( SR cytosolic side ) chamber of a bilayer apparatus and fused in the presence of an osmotic gradient ( 250 mM cis KCl/20 mM trans KCl in 20 mM KHepes , pH 7 . 4 , 2 µM Ca2+ ) . After the appearance of channel activity , trans ( SR lumenal ) KCl concentration is increased to 250 mM . A strong dependence of single channel activities on cis Ca2+ concentration indicates that the large cytosolic “foot” region faces the cis chamber of the bilayers . The trans side of the bilayer is defined as ground . Electrical signals are filtered at 2 kHz ( 0 . 5 kHz for Ca2+ currents at 0 mV ) , digitized at 10 kHz , and analyzed as described [5] .
|
Ryanodine receptors ( RyRs ) are ion channels present in the membranes of an intracellular calcium storage organelle , the sarcoplasmic reticulum . Nerve impulse triggers the opening of RyR channels , thus increasing the cytoplasmic calcium levels , which subsequently leads to muscle contraction . Congenital mutations in a specific type of RyR that is present in skeletal muscles , RyR1 , lead to central core disease ( CCD ) , which leads to weakened muscle . RyR1 mutations also render patients to be highly susceptible to malignant hyperthermia , an adverse reaction to general anesthesia . Although it is generally known that CCD mutations abort RyR1 function , the molecular basis of RyR1 dysfunction remains largely unknown because of the lack of atomic-level structure . Here , we present a structural model of the RyR1 pore region , where many of the CCD mutations are located . Molecular dynamics simulations of the pore region confirm the positions of residues experimentally known to be relevant for function . Furthermore , electrophysiological experiments and simulations shed light on the loss of function of CCD mutant channels . The combined theoretical and experimental studies on RyR1 elucidate the ion conduction pathway of RyR1 and a potential molecular origin of muscle diseases .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"computational",
"biology/molecular",
"dynamics",
"biochemistry/membrane",
"proteins",
"and",
"energy",
"transduction",
"cardiovascular",
"disorders",
"biophysics/membrane",
"proteins",
"and",
"energy",
"transduction"
] |
2009
|
A Structural Model of the Pore-Forming Region of the Skeletal Muscle Ryanodine Receptor (RyR1)
|
Interactions of proteins regulate signaling , catalysis , gene expression and many other cellular functions . Therefore , characterizing the entire human interactome is a key effort in current proteomics research . This challenge is complicated by the dynamic nature of protein-protein interactions ( PPIs ) , which are conditional on the cellular context: both interacting proteins must be expressed in the same cell and localized in the same organelle to meet . Additionally , interactions underlie a delicate control of signaling pathways , e . g . by post-translational modifications of the protein partners - hence , many diseases are caused by the perturbation of these mechanisms . Despite the high degree of cell-state specificity of PPIs , many interactions are measured under artificial conditions ( e . g . yeast cells are transfected with human genes in yeast two-hybrid assays ) or even if detected in a physiological context , this information is missing from the common PPI databases . To overcome these problems , we developed a method that assigns context information to PPIs inferred from various attributes of the interacting proteins: gene expression , functional and disease annotations , and inferred pathways . We demonstrate that context consistency correlates with the experimental reliability of PPIs , which allows us to generate high-confidence tissue- and function-specific subnetworks . We illustrate how these context-filtered networks are enriched in bona fide pathways and disease proteins to prove the ability of context-filters to highlight meaningful interactions with respect to various biological questions . We use this approach to study the lung-specific pathways used by the influenza virus , pointing to IRAK1 , BHLHE40 and TOLLIP as potential regulators of influenza virus pathogenicity , and to study the signalling pathways that play a role in Alzheimer's disease , identifying a pathway involving the altered phosphorylation of the Tau protein . Finally , we provide the annotated human PPI network via a web frontend that allows the construction of context-specific networks in several ways .
The advent of high-throughput techniques to measure and perturb molecular species in a systematic way has enabled researchers to assess the different layers of cellular metabolism under different experimental conditions . Protein-protein interaction ( PPI ) networks created by a variety of methods including yeast-two-hybrid ( Y2H ) , mass-spectrometry ( MS ) and computational predictions [1] , [2] are valuable research resources , and have been used heavily in the last decade . However , a major drawback of these data is that the artificial expression systems used to reconstruct PPI networks do not take into account two of the many factors that are essential to understand the biology of the cell: first , the time-point at which the proteins are expressed ( e . g . , cell-cycle or developmental stage ) and second , the tissue or intracellular compartment where the proteins are expressed or located ( different organs and tissues have very specific protein compositions ) . Therefore , two proteins may be reported as interaction partners , although they are expressed in different tissues or at different time-points . While high-throughput studies acknowledge these caveats , PPI databases collect these data without mechanisms explicitly directed to discern the biological plausibility of a reported interaction . Therefore , the selection of proteins expressed in a specific cell type or compartment would allow the generation of subnetworks that more realistically represent biological processes in the respective cell types or cellular compartment . Several attempts have been made to investigate the tissue-specific binding behavior of single proteins and the spatio-temporal dynamics of PPI networks [3] , [4] , [5] , [6] , [7] , [8] . In a recent study evaluating the characteristics of publicly available PPI databases , we demonstrated that the use of subnetworks ( which include only interactions of proteins expressed in the same tissue ) identifies potential mechanisms or pathways that would remain obscured if the complete PPI database was used [9] . In addition , many proteins have multiple functions , carried out in cooperation with distinct sets of interacting partners . Networks of interacting proteins with coherent function have been termed context networks [10] . Here , we adopt this notion of context and extend it to PPIs or networks of proteins being expressed in the same tissue or cooperatively transmitting signal flow . There is a lack of studies testing systematically the potential of adding context information to PPI networks in recovering meaningful PPI subsets and , although there are a few approaches that allow to add expression or functional information to PPI data [11] , [12] , [13] , convenient methods for the creation of such context-specific subnetworks are generally missing . Here , we introduce an approach to add context to PPI networks using annotations and relations between the interacting partners and demonstrate that context-specific PPI networks are enriched in high-confidence interactions . We use this approach to investigate how the proteins of the human influenza virus interfere with the immune response of the host cell in a tissue-specific manner , finding novel potential regulators of influenza virus pathogenicity , and to study the brain-specific signaling pathways that play a role in Alzheimer's disease , identifying a pathway involving the altered phosphorylation of the Tau protein . Thereby , we illustrate how the addition of context to PPI networks can guide researchers in the discovery of meaningful interactions and pathways , which would otherwise be obscured by the vast amount of irrelevant ( for a specific question ) and partly erroneous amount of PPI data .
Our approach to add context-specific information to human PPI data was implemented in the HIPPIE database [14] . HIPPIE is an integrated PPI database that currently contains more than 101 , 000 interactions of ∼13 , 500 human proteins . HIPPIE is regularly updated by incorporating interaction data from major expert-curated experimental PPI databases ( such as BioGRID [15] , HPRD [16] , IntAct [17] and MINT [18] ) in an automated manner using the web service PSICQUIC [19] . All interactions have an associated confidence score based on the sum of cumulative supporting experimental evidence . Individual proteins were associated with tissues , subcellular locations and biological processes in the following manner . First , proteins were associated with tissues ( based on their gene expression profiles retrieved from BioGPS [20] and using the method defined in [9] ) or defined as housekeeping ( using a list from [21] ) . Next , associations with biological processes and subcellular locations were determined according to the EBI Gene Ontology ( GO ) annotation ( release from October 28 , 2011; reduced to GO slim terms ) [22] , and to MeSH terms belonging to “Diseases” ( class C ) or “Tissues” ( class A10 ) that annotate the biomedical references associated to them in MEDLINE ( release 2012; gene2pubmed at NCBI ftp site ) . We associated an interaction with a tissue when both interactors are expressed in the same tissue ( e . g . “lung” ) . Given a term of a functional ontology , we associated an interaction with this function when both interactors are annotated with either the given functional term or with children of it in the hierarchy of the ontology . For example , the GO term “transport” would be associated with an interaction between a protein annotated as involved in “vacuolar transport” and another protein annotated as involved in “nucleocytoplasmic transport” . Functional terms considered were either GO terms or MeSH terms . We excluded the rather unspecific top-level terms ‘biological process’ , ‘cellular component’ and ‘cell’ . Additionally , we ignored categories that are associated to less than 20 interactions . Our approach includes a method to infer directed PPIs . This inference of interaction ( edge ) directionality needs sets of proteins predefined as sinks and sources . As default sources and sinks , we connected all proteins annotated with the GO terms ‘receptor’ and ‘sequence-specific DNA binding transcription factor activity’ , respectively , in the UniprotKB [23] . This is done assuming that signal pathways follow the transmission of information through interacting proteins starting in cell surface receptors that collect external cues and ending in transcription factors as final effectors on gene regulation , following [24] . To infer edge directionality , all pairwise shortest paths between proteins of the source and the sink sets present in the generated output network are calculated . We do not consider edge weights and , hence we are able to determine each shortest path in linear time via a breadth-first search . An edge of the network is considered to be directed if at least one shortest path goes through that edge . The direction of the path ( from source to sink ) determines the direction of the edge . Edges with conflicting orientations of passing paths are not assigned directionality . For the evaluation of the influenza virus host factor network generation we performed pathway enrichment analysis with ConsensusPathDB ( run on August 30 , 2012; [25] ) . We used a cut-off of 0 . 05 on the q-value , which is the false discovery rate ( FDR ) adjusted equivalent to the p-value . The background control for the tests was the complete list of proteins annotated as expressed in the given tissues ( and with PPI information in HIPPIE ) . We retrieved the preprocessed microarray data described in [26] measuring gene expression changes over multiple time points in a lung adenocarcinoma cell line ( Calu-3 ) infected with influenza A/Netherlands/602/2009 ( H1N1 ) . To select steadily up-regulated genes we filtered for probes differentially expressed at the last three time-points in the time series ( 30 , 36 and 48 h ) with a q-value less than 0 . 01 and a log2 fold change greater than 1 . To generate a list of PPIs related to Alzheimer's and protein phosphorylation , first , we used the webserver MedlineRanker [27] to retrieve a list of ranked PubMed abstracts ( corresponding to manuscripts published within the last 5 years ) according to their relevance to the search term “Alzheimer phosphorylation” , which relates loosely to the question of interest . Next , we input the top 50 abstracts from MedlineRanker into the webserver PESCADOR [28] , which extracts a network of potential PPIs based on a set of PubMed abstracts . In our example , PESCADOR outputs 10 interaction pairs ( type 2; co-occurrence of genes or proteins within a sentence containing a biointeraction term ) , of which only 4 pairs existed in HIPPIE as scored interactions ( PSEN1:PSEN2 , GSK3B:MAPT , APP:BACE1 , PPP2R4:SET ) . These confirmed PPIs were then used as input for further analysis .
We inferred context information for all interactions in the human PPI database HIPPIE [14] . This database collects human PPIs for which there is experimental evidence . The amount and quality of the experimental evidence supporting each PPI is evaluated with a confidence score that ranges from 0 to 1 . In a first step , we associated all 13 , 477 proteins in HIPPIE with the following attributes: tissue-expression , GO biological process and cellular compartment , and inferred annotations for the MeSH categories disease and tissue . We then inferred context associations to the PPIs according to the annotations of the interacting proteins and taking into account the hierarchical structure of GO and MeSH terms ( see Materials and Methods for details ) . By assuming that a large fraction of signaling events transmits information from proteins sensing environmental changes to effector proteins altering the cellular state , we computed shortest paths from membrane-bound receptors to transcription factors ( TF ) through the network . From the predicted information flow we assigned edge directionality to interactions on these paths ( see Materials and Methods for details ) . Overall , we were able to associate context to more than 97 , 000 of the 101 , 131 interactions of the current version of HIPPIE . Interactions for which we inferred or collected annotations had significantly better experimental evidence ( Figure 1A ) . This suggests that annotated interactions might have higher biological significance than non-annotated ones . As expected , we observed that more specific context categories were associated to interactions with higher experimental reliability: while the confidence scores of interactions with rather unspecific and ubiquitous terms resemble the overall confidence score distribution , interactions with highly specific terms usually have a higher than average confidence score ( Figure 1B-C ) . For example , the 43 , 372 interactions associated with the GO category ‘cytoplasm’ ( of depth 1 in the GO hierarchy ) have an average confidence score of 0 . 675 as compared the average of 0 . 670 over all interactions . On the other hand , the 159 interactions associated with the ( depth 3 ) GO category ‘ribonucleoprotein complex assembly’ have an average confidence score of 0 . 754 . We observed a similar tendency for more specific MeSH terms to have a higher experimental reliability . To demonstrate that our automated context association approach allows identification of relevant interactions , we tested if networks of interactions of our inferred MESH-based disease-annotation are enriched in well-known disease proteins . Therefore , we repeatedly generated disease-context networks around a set of canonical disease proteins . As a canonical disease protein specification , we retrieved the manually curated UniProt Knowledgebase disease protein annotation . For each of the canonical disease proteins , we generated two types of networks: ( a ) disease networks consisting only of interaction partners of the disease proteins that we had associated with the equivalent MeSH disease term and ( b ) unfiltered PPI network consisting of all interaction partners of the disease protein from HIPPIE . We did this for all disease proteins where the disease was associated with at least two disease proteins in UniProt and at least two interactions that we had associated with this disease . To quantify the enrichment of disease proteins in these networks we repeatedly calculated the F1 score , the harmonic mean of precision and recall ( F1 = 2*precision*recall/ ( precision+recall ) ) . A one-sided Mann-Whitney-test comparing the distribution of F1 scores between the disease networks and the non-filtered networks indicated that the F1 scores for the disease networks were significantly larger ( p<0 . 05 ) proving an enrichment of disease proteins in the disease filtered networks ( without losing sensitivity by removing disease proteins in the filtering step ) . The mean precision on the filtered networks was 0 . 47 and on the unfiltered networks 0 . 21 . The mean recall for the filtered networks was 0 . 14 and for the unfiltered networks 0 . 15 . This illustrates that in exchange for a small decrease in recall the precision can be more than doubled by applying the MeSH disease filter . We then investigated the potential of edge directionality inference based on the shortest paths between membrane-bound receptors and TFs through the PPI network to recover known pathways . We retrieved pathway annotations ( extracted from WikiPathways download March 29 , 2012 ) and computed the shortest paths through HIPPIE between all pairs of receptors and TFs within the same pathway ( excluding only pairs that directly interact or could not be connected by any path ) . We counted the number of proteins of each pathway found on the shortest paths . We found for 3163 of the 5063 pairs that this approach correctly identified proteins of the selected pathway . The mean precision ( the fraction of proteins on the paths that indeed belonged to the correct pathway ) over all combinations of receptors with transcription factors was 0 . 20 . The mean recall ( the fraction of the pathway that was recovered by considering the paths between one receptor and one transcription factor ) was 0 . 02 . To assess if the agreement between shortest paths and canonical pathways was larger than expected by chance , we generated a background distribution by computing repeatedly the shortest paths between a receptor and a TF from different pathways and computed the overlap between the proteins on the shortest paths to either the TF- or the receptor-containing pathway . We found that the overlap distribution was significantly higher when the receptor and the TF were members of the same pathway ( p<0 . 001; Mann-Whitney-test ) proving the potential of shortest paths to recover the signal flow between TFs and receptors when functionally related pairs of receptors and transcription factors are chosen . We wondered if we could further increase the overlap between the shortest paths and the canonical pathways by filtering the networks for tissue expression . To associate pathways with tissues , we determined for each pathway which tissues were enriched among the genes of the pathway ( Supplementary Table S1 lists pathway that are associated to more than 2-fold enriched tissues ) . Inspection of the tissues enriched among proteins forming a pathway revealed that in many cases they indeed reflect plausible locations for pathway activity . For example , immune response pathways were enriched among blood cells and pathways associated with neurodegenerative diseases and addiction in brain-related tissues . We repeated the computation of shortest paths linking receptors to transcription factors in tissue-specific networks for combinations of pathways and tissues listed in Supplementary Table S1 and for all pairs of receptors and transcription factors that were expressed in the respective tissue . Indeed , we observed an increase of the mean precision to 0 . 24 , which indicates that we could increase the amount of meaningful interactions by additionally filtering for tissue expression . The recall remained low ( at 0 . 03 ) , which is not surprising since many pathway-related proteins were not present in the considered tissue-specific networks and , hence , could not be detected . Again , the amount of pathway proteins on the tissue-specific shortest paths between receptors and TF from the same pathway was significantly larger as compared to shortest paths between receptors and TF from different pathways ( p<0 . 05 ) . To further investigate if the described context-associations can help to extract pathway information from networks , we compared the frequency of protein pairs being member of the same pathway ( as defined by WikiPathways ) among tissue-specific PPIs ( both proteins where required to be co-expressed in at least one tissue ) and compared this frequency to PPIs between proteins that are not expressed in the same tissue . We observed that interacting protein pairs that are expressed in the same tissue are indeed more likely to be in the same pathway as compared to interacting protein pairs that are expressed in disjoint sets of tissues ( p<0 . 001 ) . This , again , demonstrates that the annotations have captured properties related to pathways and suggests that the filtering helps revealing pathway information . In the next sections we use the context-associated PPI network to obtain novel insights into the mechanisms of human disease: we perform a targeted study of the PPI network surrounding the human proteins that interact with influenza virus proteins to find potential regulators of viral pathogenicity , and we explore the question of whether and how altered protein phosphorylation might be a cause of Alzheimer's disease . We analyzed PPI data of human proteins that interact with influenza virus proteins . Influenza viruses infect bronchial epithelial tissue and many cell types in the lung , sometimes resulting in viral pneumonia [29] . We started by obtaining a list of 87 human proteins that have been shown to interact with at least one influenza virus protein in a previous study [30] . From this list , we observed that 23 proteins were expressed in bronchial epithelial tissue ( BET ) , in whole lung , or in both tissues - we refer to these proteins as first layer host factors . We created the second layer by filtering tissue-specific proteins ( expressed in BET or whole lung ) that interact with members of the first-layer ( Figure 2A ) . Together , the first and second layers compose the tissue-specific PPI subnetworks . Next , we identified known pathways enriched in the BET- and lung-specific PPI subnetworks , and found both similarities and differences in the cellular functions of each ( see Materials and Methods for details on the enrichment analysis and a full list of enriched pathways in Supplementary Table S2 ) . Both subnetworks showed enrichment for processes related to programmed cell death and eukaryotic translation . These results are consistent with functions known to be activated or disrupted by influenza virus infection [31] , [32] , [33] . In addition , proteins in the BET subnetwork exhibited a stronger signature in processes involved with transcriptional regulation , sumoylation , and the regulation of mRNA stability ( in particular , the stability of AU-rich element-containing mRNAs ) . Although these processes tend to be associated with general housekeeping functions , we point out that many cytokine and interferon mRNAs contain AU-rich elements [34] . This observation suggests , hypothetically , that influenza virus proteins may function to dysregulate cytokine mRNA stability in BET , a function that could impact influenza virus pathogenesis through modulation of immune cell infiltration and function . In relation to sumoylation , it has been noted recently that influenza virus can gain protein functionality during infection by interacting with the sumoylation system of the host cell [35] . On the other hand , the lung subnetwork was uniquely enriched for processes related to cell-substrate adhesion ( pathway “signaling events mediated by focal adhesion kinase” ) . Because cell adhesion is important for maintaining cellular viability and epithelial barrier function , it is possible that influenza virus protein-mediated interference with this process could impact both the amount of virus-inflicted damage upon the lung and dissemination of influenza virus into extra-pulmonary sites . Cells respond to influenza infection by producing cytokines and chemokines [36] , [37] , while viral proteins counteract this innate immune response . One example of a viral protein that directly interferes on the protein level with cellular immune pathways is NS1 ( its involvement in immune response suppression is reviewed in [38] ) . Here , we noted that the lung PPI subnetwork – which was centered on viral protein-host protein interactions – was enriched for several curated pathways involving Toll-like receptor ( TLR ) and IL-1 receptor ( IL-1R ) signaling ( e . g . , “TLR JNK” , “TRAF6 mediated IRF7 activation in TLR7/8 or 9 signalling” , “IL-1 JNK” , “TLR ECSIT MEKK1 JNK” and “IL1-mediated signaling events” ) . Although these pathways are expected to be activated in response to viral infection , no previous study has identified any role for any influenza virus protein in perturbing TLR or IL-1R signal transduction . Several host proteins were consistently observed in most/all of the enriched TLR/IL-1R pathways from the influenza PPI lung subnetwork , including IRAK1 , TOLLIP and MyD88 . Under normal conditions , the IRAK1 kinase associates with TOLLIP ( an inhibitory molecule ) , and upon receptor stimulation , IRAK1 is recruited to the TLR/IL1R-receptor complex through its interaction with MyD88 ( reviewed in [39] ) . Recruitment results in activation of IRAK1 kinase activity and subsequent activation of MAP kinase pathways , NF-κB-dependent gene expression and interferon α induction . Altogether , these observations suggest the novel possibility that influenza virus proteins interfere with TLR/IL-1R signaling in lung – possibly by accessing a critical regulator of TLR/IL-1R signal transduction ( i . e . , IRAK1 ) – an observation that may have implications for the regulation of pathogenesis associated with influenza virus infections . A recent study demonstrated that signaling through the IL-1 receptor has a protective effect in mice infected with the pandemic 1918 influenza virus [40] . Another study reported that IL-1 receptor-deficient mice succumbed more easily than wild-type mice to infection with an H5N1 virus of low pathogenicity ( A/Hong Kong/486/1997 ) [41] . Moreover , IL-1 receptor-deficient mice showed reduced inflammatory pathology upon infection with A/Puerto Rico/8/34 ( H1N1 ) influenza virus [42] . Several studies also established that influenza virus infection is sensed by TLR7 in plasmacytoid dendritic cells [43] , [44] , [45] , [46] , [47] , [48] . However , none of these studies addressed the significance of IRAK1 in influenza virus pathogenicity . Our study thus exemplifies how our network analysis can identify potential regulators of influenza pathogenicity for experimental testing , for example , by assessing influenza virus infections in IRAK1-deficient cells or mice . Next , we aimed to predict more specific novel interference mechanisms by constructing directed and tissue-specific protein networks linking the viral proteins with proteins whose corresponding transcript was up-regulated after influenza virus infection . We selected steadily up-regulated transcripts from a microarray experiment measuring gene expression changes over time in a lung epithelial cell line infected with a 2009 pandemic H1N1 virus [26] ( 228 transcripts were selected in total; see Materials and Methods for more details ) . As expected , all ten most strongly enriched known pathways among the selected transcripts were involved in infection and the immune response . For example , the most highly overrepresented pathway was interferon alpha-beta signaling ( p<10e-20 ) . We constructed BET- and lung-specific networks connecting the viral proteins with the 228 up-regulated factors by shortest paths . From the shortest paths we assigned directions to the edges on these paths . The directed networks consisted of 577 ( BET ) and 1056 ( lung ) PPIs . To examine if these networks might reveal relevant information on how viral proteins interfere with the cellular immune response , we tested for enrichment of known pathways in the directed networks . We found that the directed networks were strongly enriched in immune response-related pathways ( especially cytokine-related ) even after excluding the 228 up-regulated transcripts , indicating that enrichment was independent of the high fraction of immune response factors in the transcriptomics data ( Supplementary Table S3 ) . For example , we observed a significant enrichment in both the directed BET- and lung-specific networks for proteins related to IL-2 and IL-6 signaling and focal adhesions ( q-values<0 . 05 ) . This suggested that we , indeed , might have captured relevant crosstalk between the viral proteins and immune pathways . The full networks are included in the File S1 . To mine the directed networks for interactions that are involved in interference mechanisms of the viral proteins with the cellular immune response , we concentrated , again , on layer one and two host factor proteins on the shortest paths . From the list of curated pathways enriched in both the BET and the lung directed networks ( Supplementary Table S3 ) , we selected several cytokine-related pathways ( marked in Supplementary Table S3 ) and filtered for interactions where the second layer protein was in one of these pathways but the layer one protein was not ( to specifically detect novel , indirect interference mechanisms ) . This resulted in a comprehensive BET network consisting of 49 interactions and a lung network formed by 67 interactions including viral proteins and host factors up to layer two ( see Supplementary Table S4 for the comprehensive networks and Figure 2 for a manually curated subset of these networks ) . Close inspection of these comprehensive cytokine-related networks in both BET and lung revealed several points of potential viral protein-mediated interference with inflammatory pathways ( Figure 2 ) . For example , the BET network showed interactions between viral polymerase complex proteins ( i . e . , PB1 and PB2 ) and BHLHE40 , a transcriptional regulator that cooperates with HDAC1 to repress STAT1 activity [49] ( Figure 2B ) . STAT1 is essential for the activation of interferon stimulated genes , which repress viral replication , and while influenza virus has an established ability to impair STAT1 [50] , no such function has been assigned to any of the viral polymerase complex subunits . BHLHE40 also interacts with TOLLIP , a suppressor of TLR signaling [51] ( see also the discussion of lung-specific inflammatory pathways above ) . This implies that the BHLHE40 protein could act as an important access point for influenza virus-mediated interference with host antiviral and inflammatory regulation in BET , and further that viral polymerase subunits may have an important – yet unappreciated – role in this activity . As in BET , lung-specific cytokine-related networks revealed that influenza virus proteins interface with TOLLIP ( Figure 2C ) . However , it is notable that , in lung , this interaction occurs through BHLHE40 and two additional routes ( i . e . , MAGED1 and RBPMS ) , potentially involving up to four viral proteins: ( i ) the aforementioned polymerase complex subunits , PB1 and PB2; ( ii ) the viral ion channel protein , M2; ( iii ) and the viral RNA-binding nucleoprotein , NP . Thus , access to TOLLIP might be particularly important in lung . The PB1/PB2-BHLHE40 interaction is maintained in this tissue type , although the nature of the interaction may differ compared to BET . Specifically , BHLHE40 may favor interaction with STAT3 ( Figure 2C ) , and previous evidence indicates that BHLHE40 stimulates STAT3 activity rather than inducing inhibition [52] . Thus , analysis of context-specific PPIs – in combination with influenza virus-induced changes in the cellular transcriptome – reveal important , putative tissue-specific differences in the ability of viral proteins to interact with cellular immune response signaling networks . Additional experiments will be necessary to further establish the functions of these interactions . Assuming no prior expert knowledge on a given topic , we applied a systematic protocol which can , in principle , be used to interrogate the PPI network about the involvement of protein interactions in a complex biological question according to current knowledge . In general , altered states of protein phosphorylation affect the PPI network and can lead to pathogenesis . Our goal in this example was to investigate the possible role of protein phosphorylation in Alzheimer's disease ( AD ) , the most common form of dementia . AD is a degenerative disease manifesting in the brain , and its cause has been hypothesized to be the formation of protein aggregates leading to neuron death , in particular related to the abnormal phosphorylation of the microtubule-associated protein tau [53] . First , we need to input a list of proteins related to the topic . Using a literature mining protocol ( see Materials and Methods for details ) we generated a list of PPIs related to Alzheimer's and protein phosphorylation: PSEN1:PSEN2 , GSK3B:MAPT , APP:BACE1 , and PPP2R4:SET . We then studied the network surrounding these interactors ( Figure 3 ) . The initial PPI network contained 727 interactions ( Figure 4A ) . Interactions could be further filtered on the basis of reasonable criteria , namely by tissue filtering for housekeeping and genes expressed in the brain ( we selected “whole brain” and “prefrontal cortex” ) , and filtering for genes related to the GO term “cell death” , reflecting that AD is characterized by death of neural cells ( Figure 4B ) . Finally , to reveal potential signal transduction pathways we used the inference of edge directionality from receptors to TFs described above ( Figure 4B ) . Within the resulting network , we highlighted the following path ( Figure 4 ) : LRP6-GSK3B-MAPT-AATF . The low density lipoprotein receptor-related protein 6 ( LRP6 ) interacts with glycogen synthase kinase 3B and attenuates the kinase's ability to phosphorylate microtubule associated protein tau ( MAPT ) [54] . Tau protein can contribute to AD in different ways: 1 ) the hyperphosphorylation of tau protein can affect microtubule stability , leading to a disassociation of tau protein from the microtubule , possibly followed by the aggregation of phosphorylated tau into neurofibrillary tangles , which are observed in the brains of AD patients [55]; 2 ) mediated by protein phosphatase 1 and GSK3 activity , Tau filaments interfere with axonal transport in the neuron , which is consistent with deficiencies in axonal transport in AD [56] . Tau protein has been found to co-localize in the cytoplasm with Che-1 ( AATF ) , which is an evolutionarily conserved RNA polymerase II binding protein that accumulates in the cell upon DNA damage [57] . It appears that Che-1/Tau proteins dissociate during neuronal cell death [58]; however , the function of Che-1 in the cytoplasm is unclear , as Che-1 is a nuclear protein that is involved in gene regulation of E2F1 targets and p53 and has pro-proliferative and anti-apoptotic functions [59] . Together , these interactions suggest a complex interplay whereby the Tau phosphorylation state and structure , and context-dependent protein distribution within the cell may contribute to neuronal cell death and AD pathology . An unbiased search for protein phosphorylation in relation to cell death in AD pointed us to this interesting pathway .
The incorporation of tissue-specific expression information to create PPI subnetworks is a useful method to elucidate biological processes that cannot be observed when using the complete PPI network . Here we have shown an approach for the inference of associated context for PPIs based on the annotations of the interacting partners , which enhances the relevance of the annotated interactions . Interactions between proteins expressed in the same location ( e . g . lung ) or at the same time or developmental stage ( e . g . embryo development ) can then be selected . Directed pathways can be inferred and highlighted in the filtered network according to sets of sources and sinks corresponding to receptors and transcription factors . Using this approach we were able to identify novel , tissue-specific interactions between influenza virus proteins and cellular inflammatory signaling pathways that may regulate pathogenesis associated with infection , and to describe a brain-specific protein phosphorylation pathway relevant for Alzheimer's disease . Several methods exist to create subnetworks of the human interactome based on context criteria . For example , POINeT [11] integrates the major PPI databases and allows the creation of tissue-specific networks . To our knowledge we are the first to combine edge directionality , gene expression and functional information for the detection of meaningful interactions . Some approaches exist that infer information flow in a network from the shortest paths ( or ‘lowest costs’ if costs are associated with edges ) that connects a set of source nodes with sink nodes . Cytoscape plug-ins such as BisoGenet [60] and GenePro [61] find the shortest paths between nodes of the gene and protein network and represent properties of the nodes . SPIKE [62] includes curated pathway data and also calculates pathway inference . The task of identifying signaling events from PPI data and functional protein annotation alone has been addressed in several studies [24] , [63] , [64] and implemented in tools ( e . g . ANAT [65] ) . Here , we proposed a protocol for edge directionality prediction based on calculating the shortest paths between sources and sinks . This protocol is runtime-efficient , which allowed us to provide it as a web tool that is the first to combine both PPI analysis for inference of edge directionality and PPI filtering by tissue and function ( available from http://cbdm . mdc-berlin . de/tools/hippie/ ) . In summary , we have presented and made available an approach to associate context to PPI networks , which provides novel biological insight into mechanisms of disease . The continuing generation of PPI data and further incorporation into databases , and an increasing quality of annotations attached to genes and proteins will result in further improvements of our methodology .
|
Protein-protein-interactions ( PPIs ) participate in virtually all biological processes . However , the PPI map is not static but the pairs of proteins that interact depends on the type of cell , the subcellular localization and modifications of the participating proteins , among many other factors . Therefore , it is important to understand the specific conditions under which a PPI happens . Unfortunately , experimental methods often do not provide this information or , even worse , measure PPIs under artificial conditions not found in biological systems . We developed a method to infer this missing information from properties of the interacting proteins , such as in which cell types the proteins are found , which functions they fulfill and whether they are known to play a role in disease . We show that PPIs for which we can infer conditions under which they happen have a higher experimental reliability . Also , our inference agrees well with known pathways and disease proteins . Since diseases usually affect specific cell types , we study PPI networks of influenza proteins in lung tissues and of Alzheimer's disease proteins in neural tissues . In both cases , we can highlight interesting interactions potentially playing a role in disease progression .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"protein",
"interactions",
"influenza",
"signaling",
"pathways",
"alzheimer",
"disease",
"infectious",
"diseases",
"proteins",
"regulatory",
"networks",
"biology",
"dementia",
"systems",
"biology",
"biochemistry",
"signal",
"transduction",
"neurological",
"disorders",
"neurology",
"viral",
"diseases",
"computational",
"biology",
"molecular",
"cell",
"biology"
] |
2013
|
Adding Protein Context to the Human Protein-Protein Interaction Network to Reveal Meaningful Interactions
|
A large amount of short interfering RNA ( vsiRNA ) is generated from plant viruses during infection , but the function , structure and biogenesis of these is not understood . We profiled vsiRNAs using two different high-throughput sequencing platforms and also developed a hybridisation based array approach . The profiles obtained through the Solexa platform and by hybridisation were very similar to each other but different from the 454 profile . Both deep sequencing techniques revealed a strong bias in vsiRNAs for the positive strand of the virus and identified regions on the viral genome that produced vsiRNA in much higher abundance than other regions . The hybridisation approach also showed that the position of highly abundant vsiRNAs was the same in different plant species and in the absence of RDR6 . We used the Terminator 5′-Phosphate-Dependent Exonuclease to study the 5′ end of vsiRNAs and showed that a perfect control duplex was not digested by the enzyme without denaturation and that the efficiency of the Terminator was strongly affected by the concentration of the substrate . We found that most vsiRNAs have 5′ monophosphates , which was also confirmed by profiling short RNA libraries following either direct ligation of adapters to the 5′ end of short RNAs or after replacing any potential 5′ ends with monophosphates . The Terminator experiments also showed that vsiRNAs were not perfect duplexes . Using a sensor construct we also found that regions from the viral genome that were complementary to non-abundant vsiRNAs were targeted in planta just as efficiently as regions recognised by abundant vsiRNAs . Different high-throughput sequencing techniques have different reproducible sequence bias and generate different profiles of short RNAs . The Terminator exonuclease does not process double stranded RNA , and because short RNAs can quickly re-anneal at high concentration , this assay can be misleading if the substrate is not denatured and not analysed in a dilution series . The sequence profiles and Terminator digests suggest that CymRSV siRNAs are produced from the structured positive strand rather than from perfect double stranded RNA or by RNA dependent RNA polymerase .
The RNA silencing based antiviral plant response is one of the best studied antiviral strategies in plants . The key element of RNA silencing based antiviral strategies is the virus derived small interfering RNA ( vsiRNA ) , which guides the RNA induced silencing complex ( RISC ) to target viral genomes in plants and invertebrates [1] . siRNAs are processed from double-stranded RNAs ( dsRNA ) or structured single-stranded RNAs ( ssRNAs ) by RNase III-like enzymes such as DICER [2] , [3] ( in plants there are several Dicer-like ( DCL ) genes ) . siRNAs guide the sequence-specific inactivation of target mRNAs by RISC [4] . Plant RNA viruses are strong inducers as well as targets of RNA silencing and high levels of vsiRNAs accumulate during the viral infection . However , despite of the extensive studies of siRNA biogenesis the origin of plant viral siRNA is still not understood . vsiRNAs are thought to be processed from ds viral RNA replication intermediates , local self-complementary ds regions of the viral genome or through the action of RNA-dependent RNA polymerases ( RDRs ) on viral RNA templates [1] . In plants two distinct classes of vsiRNAs have been identified: the primary siRNAs , which result from DCL mediated cleavage of an initial trigger RNA , and secondary siRNAs , whose biogenesis requires an RDR enzyme [5] , [6] . DCL4 and DCL2 are the most important plant DICERs involved in virus induced RNA silencing and they can process ds or hairpin viral RNAs into vsiRNAs of 21 and 22 nt , respectively . Although DCL4 is the major player in vsiRNA production , in the absence of DCL4 , DCL2 is also sufficient to produce 22 nt vsiRNA , which are biologically active in antiviral silencing response [7] , [8] . siRNAs are associated with distinct Argonaute ( AGO ) -containing effector complexes to guide them to their RNA target molecules [1] , [9] , [10] . In plants , loading of siRNAs into a particular AGO complex is preferentially , but not exclusively , dictated by their 5′ terminal nucleotides [11] . AGO1 is the major slicer in plants but other AGO paralogs are likely to be involved , potentially also mediating translational repression [1] , [11] . The accumulation of vsiRNAs may also depend on the presence of virus expressed silencing suppressor proteins . Several silencing suppressor proteins sequester the primary vsiRNAs thus inhibiting the accumulation of secondary vsiRNAs and the antiviral response [1] , [12] . In Arabidopsis , RDR6 and RDR2 are required for virus ( but not for all viruses ) or sense transgene triggered RNA silencing , the spread of long-range cell-to-cell silencing , and the reception of the long-distance RNA silencing signal [1] , [13] , [14] , [15] . RDR-dependent biogenesis of vsiRNAs was first demonstrated by Diaz-Pendon et al . [16] . Recently it was also shown that the production of Tobacco rattle virus ( TRV ) derived vsiRNAs and antiviral silencing are strongly dependent on the combined activity of the host-encoded RNA-dependent RNA polymerases such as RDR1 , RDR2 , and RDR6 suggesting that viral single-stranded RNAs , might be converted by RDR enzymes to dsRNAs , which could serve as a substrate for vsiRNA production [17] . However , this model is not supported by previous observations that the majority of vsiRNAs are derived from the plus ( mRNA sense ) viral strand [18] , [19] , [20] . In addition , it has been also shown that RDR6 is not required for virus-induced gene silencing when the endogenous phytoene desaturase ( PDS ) gene was silenced using the crucifer strain of tobacco mosaic virus ( crTMV ) and TRV based vectors [13] . These conflicting observations indicate that our knowledge of vsiRNA biogenesis is still far from complete . Cymbidium ringspot virus ( CymRSV ) is a member of the Tombusvirus genus containing a positive single-stranded RNA genome with five open reading frames ( ORFs ) [21] . It is widely assumed that positive-strand RNA viruses replicate their genomes via dsRNA intermediates that may activate the siRNA generating machinery [22] . However , the accessibility of long viral dsRNA intermediates for DICER cleavage have not been proven experimentally . In addition , our previous studies suggested that highly structured viral RNAs might also be processed into vsiRNAs in virus-infected plants [18] , [19] , [23] . We have also shown that sensor RNAs with negative viral polarity are better targets of RISC mediated cleavage in CymRSV infected plants than those with positive polarity , which further suggests an excess of biologically active vsiRNAs with positive polarity [24] . In this work we analysed the composition and the molecular nature of vsiRNAs in virus infected plants in order to gain a better insight into the biogenesis of vsiRNAs using the high throughput 454 and Solexa sequencing strategies .
To establish the profile of viral siRNAs , cDNA libraries of short RNAs were generated using total RNA extracted from the first systemically infected leaves of Nicotiana benthamiana infected with in vitro transcripts of either wild type CymRSV or with a mutant form of the virus that did not express the p19 silencing suppressor protein ( Cym19stop ) . The cDNA libraries were sequenced on the 454 platform yielding around 100 000 sequences for each library ( Table 1 ) . The short RNA sequences were mapped to the viral genome and we found 65% of the reads on the positive strand of wild type CymRSV without mismatches . The mutant virus also produced vsiRNAs at a similar ratio from the positive strand ( 63 . 5% ) indicating that the silencing suppressor did not affect strand preference and that the higher representation of plus strand is consistent . Next we assessed whether deep sequencing also identifies vsiRNA hot spots as small scale vsiRNA sequencing did [19] . The abundance profiles of vsiRNAs on the positive and negative strands of viral genomic RNA were very similar for wild type and mutant viruses ( Figure S1A and B ) : vsiRNAs were produced from the entire genome with several hot spots on the positive strand and a few low abundance hot spots on the negative strand . The position of the hot spots were identical between wild type and mutant viruses indicating that p19 does not influence the generation of high abundance vsiRNAs and that the 454 platform yielded reproducible vsiRNA profiles . Next we wanted to experimentally validate the positions of the hot spots obtained by 454 sequencing . A 520 nt region ( nucleotides 2650–3169 on the positive strand ) containing one highly abundant and a few less abundant hot spots on the positive strand was selected for validation . Five hundred 21mer oligonucleotides ( complementary to the plus strand of CymRSV ) in a one nucleotide sliding window were designed to cover the entire selected region and spotted on a membrane . The short RNA fractions ( 19–24 nt ) were isolated from first systemic leaves of plants infected with in vitro transcripts of wild type and mutant virus , labelled at the 5′ end and used as probes to hybridise to the membrane . The two dot blots gave very similar if not identical patterns ( Figure 1A and B ) providing experimental evidence that hot spots do exist and p19 does not influence the position of hot spots . The membranes were also hybridised to labelled short RNAs isolated from wild type CymRSV in vitro transcript infected Nicotiana clevelandii and a N . benthamiana where the RDR6 gene [13] , [14] was silenced by an inverted repeat construct [15] ( Figure 1C and D ) . These probes gave identical patterns to the previous two probes suggesting several conclusions . First , it indicated that viral siRNAs were generated from hot spots in the absence of RDR6 therefore this RNA dependent RNA polymerase was not absolutely essential for the increased production of CymRSV vsiRNAs from certain positions of the viral genome . Second , it showed that the positions of hot spots only depend on the viral genome and are not determined by the host since the hot spots mapped to the same regions in two different hosts . The fact that four experiments gave identical patterns of hot spots proved that the experimental approach is reproducible . Intensity of the dots was quantified and plotted along the selected region to obtain vsiRNA profiles . Comparison of vsiRNA profiles obtained by 454 and hybridisation , however , showed no correlation ( Figure 1E and F ) , although both approaches were reproducible . We carried out another dot blot experiment covering a different region on the viral genome ( 4300–4505 ) and obtained similar result ( Figure 1G , H and I ) . One of the differences between the two methods is that the short RNAs were de-phosphorylated before labelling the 5′ end while the 5′ 454 adapter was ligated directly without de-phosphorylation . If many short RNAs contain a 5′ end that cannot ligate directly to an adapter but can be labelled after de-phosphorylation that could explain the difference between the two sets of results . Alternatively , the two approaches have different preferences in detecting vsiRNA sequences . To distinguish between these possibilities first we analysed the 5′ end structure of vsiRNAs using the Terminator™ 5′-Phosphate-Dependent Exonuclease and then we profiled vsiRNAs using a different high-throughput sequencing platform after ligating the 5′ adapter either directly or following de- and re-phosphorylation to short RNAs extracted from plants inoculated with in vitro transcripts of either wild type or Cym19stop viruses . To decide whether a substantial percentage of vsiRNAs cannot be ligated directly to an adapter through the 5′ end we used the terminator exonuclease to study the 5′ end of vsiRNAs . This enzyme degrades RNA molecules in a 5′ to 3′ direction but it can only process molecules with a monophosphate at the 5′ end [25] , however it is not known whether short dsRNA molecules can be processed by this enzyme . It was important to address this question since vsiRNAs are double stranded [19] therefore we digested a perfect duplex ( with two nucleotide overhangs at the 3′ ends ) , which consisted of two 19mer in vitro synthesised and phosphorylated RNA oligonucleotides in decreasing concentration without or after denaturation and compared these with undigested samples ( Figure 2A ) . The enzyme was not able to digest the non-denatured perfect duplex at all ( Figure 2A left panel ) indicating that these molecules are not substrates of this enzyme even if they have 5′ monophosphate end . Most of the RNA was digested by the exonuclease when only one of the strands from the duplex was digested , which demonstrates that the enzyme was functional and that the RNA oligonucleotides were efficiently phosphorylated . The enzyme was able to partially digest the duplex RNA if it was denatured before incubating with the enzyme ( Figure 2A right panel ) . This experiment was repeated using two other perfect duplexes and the same result was obtained ( Figure S3B ) . RNA extracted from CymRSV in vitro transcript infected plants was treated with terminator exonuclease without or after denaturation and the level of different short RNAs was compared with an untreated sample by northern blotting . We used a miRNA probe as a positive control for the treatment because it is known that the 5′ end of miRNAs contain a monophosphate and so they are expected to be processed by terminator exonuclease . We also added the 19mer perfect duplex to the samples as an additional control . The miRNAs were indeed efficiently degraded by the terminator ( Figure 2B miR159 probe ) providing an internal control for the enzyme reaction . The perfect duplex gave the same result as previously: it was not digested at all without denaturation and was partially digested after denaturation ( Figure 2B 19mer probe ) . Two different probes were used to analyse the viral siRNAs ( Figure 2B Cym probes ) and they gave identical results: vsiRNAs were partially digested with or without denaturation and the efficiency of the digestion was increased dramatically by diluting the sample ( two more viral probes are shown on Figure S2 ) . Comparing the digestion efficiency of the vsiRNAs and the perfect duplex after denaturation we can conclude that the vsiRNAs are even more efficiently digested than the 19mer duplex . This indicates that most vsiRNAs must have monophosphate at the 5′ end . The weak signal in the digested lanes of the diluted samples could be either due to renaturation or a small amount of 5′ triphosphate siRNAs . Based on this experiment we cannot rule out that a small amount of vsiRNAs have 5′ triphosphate but we can conclude that the majority of CymRSV vsiRNA possess 5′ monophosphate . We also compared the digestion efficiency of vsiRNAs and the control perfect duplex without denaturation and found that vsiRNAs were much more efficiently digested without denaturation than the perfect duplex . This raises the possibility that the vsiRNAs are not perfect duplexes but imperfect dsRNA molecules with mismatches between the two strands as it was proposed previously [19] . To explore this we used synthetic imperfect duplexes with 2 , 3 or 4 mismatches between the strands ( Figure 3A and Figure S3C ) . Two kinds of duplexes were used with 4 mismatches that differed in the position of the mismatches ( Figure 3A ) . First we digested all four imperfect duplexes at 2 µM concentration either without or after denaturation ( Figure 3B ) . None of the imperfect duplexes were digested without denaturation but all of them were efficiently digested after denaturation . This showed that imperfect duplexes are digested more efficiently after denaturation than perfect duplexes because the perfect duplex was not digested at all at this concentration even after denaturation ( compare right panels of Figure 2A and Figure 3B ) . Next we compared the digestion pattern of imperfect duplexes and vsiRNAs without denaturation by diluting each of the four imperfect duplexes and digesting them without denaturation ( Figure 3C ) . The perfect duplex was used as a control and it was not digested even at the lowest concentration we tried ( 0 . 002 µM ) . The imperfect duplex with two mismatches showed a very small amount ( if any ) digestion at 0 . 002 µM but not at higher concentrations . The imperfect duplex with three mismatches was digested much more efficiently , showing big differences between digested and non-digested samples at concentrations 0 . 002 µM and even 0 . 004 µM . The imperfect duplex with four mismatches at the end of the duplex showed the most efficient digestion , detected even at 0 . 1 µM , while the same number of mismatches but at the middle of the duplex resulted efficient digestion at a slightly lower concentration ( 0 . 02 µM ) . Two more imperfect duplexes showed similar pattern ( Figure S3C ) . This set of experiments demonstrated that perfect and imperfect duplexes show different digestion pattern by terminator exonuclease without denaturation and that vsiRNAs showed a similar pattern to the imperfect duplexes: they are both digested at the “right concentration” . The right concentration depends on the number and position of mismatches . We also studied the 5′ end of vsiRNAs by ligating another adapter molecule ( different from the 454 5′ adapter ) to the 5′ end of vsiRNAs . cDNA libraries of short RNAs were generated using total RNA extracted from early systemic leaves of N . benthamiana infected with in vitro transcripts of either wild type CymRSV or with a mutant virus , Cym19stop . The library generation includes an adaptor ligation to the 5′ end of the short RNAs and this can be done either after removing the phosphate ( s ) from the 5′ end or without modifying the 5′ end . When short RNAs are dephosphorylated the 5′ monophosphate , diphosphate or triphosphate is removed and then the short RNAs are re-phosphorylated ensuring that all short RNAs have a 5′ monophosphate . Solexa sequencing of the four libraries resulted in over one and a half million reads for each library ( Table 2 ) . The ratio of vsiRNAs mapping to the positive and negative strands was identical for the wild type virus infected plants when the short RNAs were directly ligated to the adapter or after de- and re-phosphorylation . The high percentage of vsiRNAs derived from the positive strand ( 93% ) confirmed once more that most vsiRNAs are not processed from the double stranded ( ds ) replicative form . This result also shows that the bias for positive strand sequences is not due to a modification at the 5′ end of the vsiRNAs because the two ligation approaches gave identical ratios . The ratio of positive/negative strand vsiRNAs and their distribution was also very similar in the libraries obtained by the two different ligation methods using RNA extracted from the mutant virus ( Cym19stop ) infected plants . However , the percentage of vsiRNAs generated from the negative strand of the mutant virus ( Cym19stop ) was higher than it was for the wild type virus ( 22–29% compared to 5–6%; Table 2 ) . We hypothesised that this increase was due to higher ratio of negative strand genomic RNA to positive strand genomic RNA in the mutant virus infected plants compared to wild type virus infected plants and this was confirmed by Northern blot analysis ( Figure S4 ) . Although most redundant vsiRNAs matched perfectly the viral genome , we observed a higher ratio of not perfectly matching non-redundant reads/perfectly matching non-redundant reads for the positive strand vsiRNAs compared with the negative strand vsiRNAs ( p<0 . 05 ) . One possible explanation for this is the fact that a lot more sequences are derived from the positive strand therefore there is a higher probability to observe sequencing errors in positive strand vsiRNAs . In fact the difference is more pronounced for the wild type virus compared to the mutant virus most likely because the positive/negative strand vsiRNA ratio is higher for the wild type ( 15× ) than for the mutant ( 2 . 7× ) sample . We also simulated the effect of sequencing errors on positive and negative strand vsiRNAs . The perfectly matching positive and negative strand vsiRNAs were separated from the wild type data set and different rates of sequencing errors ( 1 error in 1000 , 100 or 10 sequences ) were simulated . The ratio of perfectly matching sequences to those matching with 0 or 1 mismatch was calculated . We found a similar effect on the positive and negative strand vsiRNAs when redundant reads were used but a much more pronounced effect on the positive strand vsiRNAs when the simulation was applied to non-redundant reads ( Figure S5 ) . This result confirms that the observed difference in ratios of not perfectly matching non-redundant reads/perfectly matching non-redundant reads for the positive strand vsiRNAs compared with the negative strand vsiRNAs can be due to sequencing errors . The other contributing factor could be if the not perfectly matching reads are not sequencing errors but vsiRNAs derived from mutated viral genomes . Positive strands can contain point mutations that are not present on the negative strand if no negative strand is produced from the mutant positive strand . However , it is less likely that the negative strand contains a mutation that is not present on the positive strand since the positive strands are produced from the negative strands . The abundances of reads in all four libraries were plotted on the viral genome ( Figure 4A , B , C and D ) and hot spots were located . Comparison of the vsiRNA profiles obtained by Solexa and the dot blot approaches revealed a high similarity between the two profiles ( Figure 4E and F ) . Since the Solexa platform gave more similar result to the dot blot than the 454 we analysed the Solexa profile further ( direct comparison of the vsiRNA profiles obtained by 454 and solexa is shown on Figure S6 ) . Size distribution and 5′ nucleotide bias was compared between the four libraries and specifically between vsiRNAs derived from hot spots ( Figure S7 ) . A slight bias was found for A and U at the 5′ position , especially in wild type virus infected plants but regardless of the ligation method . We also found a slight G bias in the first position of 22mer reads in the libraries from plants infected with the mutant virus . This bias was slightly more pronounced in the vsiRNAs derived from hot spots but there was no other significant difference in the size distribution and 5′ nucleotide bias between the total population of siRNAs and siRNAs generated from hot spot regions . The main difference revealed by this analysis was that the most abundant class of siRNAs in plants inoculated with wild type virus was the 21 nt sequences followed by the 22 nt class while in plants inoculated with the mutant virus the 22 nt siRNAs were more abundant than the class of 21 nt ( the same was found by 454 sequencing; data not shown ) . To further investigate this difference we plotted the 21 and 22 nt sequences separately for all four libraries ( Figure S8 ) . The plots show a very similar pattern for 21 and 22 nt sequences in all four libraries and a very similar high positive/negative strand ratio . All of our results ( including the presence of hot spots , the + strand bias , the independence of hot spots from RDR6 , the similar pattern of libraries generated by direct ligation or after dephosphorylation and the similar profile of wild type and mutant viruses; please note that p19 blocks the generation of primer dependent secondary siRNAs ) suggest that vsiRNAs are produced from structured + or − strands , rather than from the double-stranded replicative form or from host RDR generated double-stranded RNA . To explore this further one could predict the secondary structure of the viral genomic RNA or the hot spot regions . However , prediction of RNA secondary structure of long molecules is not reliable . It is also very likely that long RNA molecules have several possible conformations , which is difficult to model by using computational methods . Prediction of shorter regions is more reliable but this approach completely ignores that sequences far away from each other can pair with each other . Therefore secondary structures of hot spot regions are not informative either . We took a different approach and investigated whether the sequence reads derived from the + strand can potentially form duplexes with each other allowing up to four mismatches . We found a large number ( 8492 ) of such potential duplexes ( Table S1 ) suggesting that vsiRNA duplexes indeed can be derived from the positive strand only . A similar approach identified 3617 imperfect duplexes that can derive from the − strand ( Table S2 ) . In order to understand the biological role of hot spots , we carried out a functional analysis of viral siRNAs . Different regions ( 200 nt length ) of the viral genome in both plus and minus orientation were selected from hot spots and from non hotspots regions ( Figure 5 ) . These were cloned downstream of a green fluorescent protein ( GFP ) reporter gene and expressed from a binary vector . The reporter constructs were agroinfiltrated into systemically infected leaves of N . benthamiana plants inoculated with in vitro transcript of Cym19stop virus . In the absence of the silencing suppressor protein the viral siRNAs are incorporated into RISC and target the GFP mRNA if they are complementary to the viral fragment in the sensor construct [24] , [26] . Some constructs contained fragments including hot spots and other constructs carried fragments that did not include hot spots . All sensor constructs were targeted by vsiRNAs , although with variable efficiency and without clear correlation with vsiRNA hotspots ( Figure 5 ) . In fact , some fragments not homologous to any hot spot ( Figure 5 , lanes F- ) showed similar down-regulation to constructs with hot spot regions ( Figure 5 , lanes G- ) . This result indicates that vsiRNAs generated from hot spots are not more efficient than other vsiRNAs in spite of their much higher abundance .
We used three approaches to profile vsiRNAs: 454 and Solexa high-throughput sequencing and hybridisation of 5′ labelled vsiRNAs to two arrays of DNA oligonucleotides covering a 520 and a 206 nt region of the viral genome . All three approaches gave reproducible results because total RNA extracted from wild type or suppressor deficient CymRSV infected plants gave very similar results for each technique . The hybridisation and Solexa sequencing showed good correlation with each other whereas the profile obtained using 454 sequencing was not in agreement . This suggests that ligation of the 454 adaptors have some sequence preference and not necessarily the most abundant molecules are present at high frequency in the cDNA library . This was also reflected by a bias to sequence reads starting with C nucleotide in the 454 library ( data not shown ) compared with a bias for U and A in the Solexa library ( Figure S7 ) . This however , does not mean that 454 is not appropriate to profile short or mRNAs in different samples . The 454 profile was reproducible demonstrating that the preference shown for certain sequences is the same in different samples therefore it is still a valid technique to compare the abundance of transcripts/sRNAs between samples . During revision of this manuscript it was reported that different high-throughput sequencing platforms and different library construction protocols led to different profiles within the same sample [27] . The reason for these differences is not known but is probably influenced by the sequence of the short RNAs to be profiled . Because of this it is possible that different platforms/protocols would give a more accurate profile in different systems . Clearly more studies are necessary that use different platforms in the same system , and in different systems in order to obtain a comprehensive picture about the reliability of the different platforms . Both sequencing approaches identified more negative strand vsiRNAs in the p19 mutant virus infected plants , which mainly mapped to the last 1700 nt of the viral genome ( downstream of nucleotide 3000 ) . This shows a good correlation with the strongly increased level of subgenomic RNA 2 negative strand in the mutant virus infected plants compared with wild type virus infected plants , suggesting that the subgenomic RNAs also contribute to the vsiRNA pool . The Solexa and array approaches both identified regions from which vsiRNAs were generated at a higher than average frequency . Although this was reported previously based on small scale sequencing of vsiRNAs it could not be ruled out that those observations were influenced by the small data size . This comprehensive study confirmed that there are hot spots on the viral genome that produce specific vsiRNAs at a very high abundance . We also showed that the positions of these hot spots are the same in at least two different hosts indicating that it is determined by the virus itself . The profile of vsiRNAs was also identical in wild type plants and in plants where RDR6 accumulation was suppressed . The symptoms of these two plants following CymRSV infection were also very similar ( data not shown ) demonstrating that RDR6 is not required for vsiRNA production from the CymRSV genome . Some viruses , such as cucumber mosaic virus [14] and potato virus X [15] cause more severe symptoms in the absence of RDR6 than on wild type plants but many other viruses including tobacco rattle virus , tobacco mosaic virus , turnip crinkle virus cause similar symptoms in the presence and absence of RDR6 [13] . This indicates that production of siRNAs from some viruses requires RDR6 but this enzyme is not involved in vsiRNA production from other viruses , including CymRSV . Since there are several other RDR family members in plants it is possible that one or more other RDRs are involved in CymRSV vsiRNA production through two potential mechanisms . For example it was reported recently that RDR1 is involved in vsiRNA production in TMV-Cg infected plants [28] . The first mechanism is similar to the generation of DCL-dependent ( DCL-D ) secondary siRNA production from transgenes [1] where primary siRNAs prime the synthesis of complementary RNA of the silenced RNA by RDR . The resulting dsRNA is processed by one of the DCL genes producing DCL-D secondary siRNAs . These DCL-D secondary siRNAs contain 5′ monophosphate like primary siRNAs that are also produced by one of the DCLs but independently of any RDR enzyme . The second possible mechanism was described in C . elegans where an RDR directly produces siRNAs without Dicer activity [25] ( DCL-I; DCL independent ) . However , our data indicate that in the case of CymRSV none of these mechanisms play an important role as discussed next . The CymRSV encoded silencing suppressor ( p19 ) was shown to bind primary siRNAs and suppress the accumulation of DCL-D secondary siRNAs by blocking the activity of the primary siRNAs [26] . This indicates that if DCL-D secondary siRNAs were produced at a high level in CymRSV infected plants than we should observe a different vsiRNA profile in the absence of the viral suppressor . In this scenario Cym19stop infected plants would contain more DCL-D secondary siRNAs than the wild type virus infected plant and the vsiRNA profile would be different in the two plants . However , we did not obtain different vsiRNA profiles in wild type and mutant virus infected plants indicating that most CymRSV siRNAs are primary siRNAs and not DCL-D secondary siRNAs . Moreover the vsiRNA profile of CymRSV infected RDR6 silenced plants was the same as the virus infected wild type plants further supporting our conclusion since RDR6 is involved in DCL-D secondary siRNAs biogenesis . Although DCL-I secondary siRNAs have not been described in plants , plant virus derived siRNAs are good candidates for being direct products of an RDR . Most plant viruses are RNA viruses therefore vsiRNAs could be produced by the viral RDR or alternatively by one of the host RDRs . In addition , vsiRNAs were shown to be generated from hot spots on the viral genome that could be preferred sites of unprimed RNA synthesis . We took two different experimental approaches to analyse the 5′ structure of vsiRNAs at a genomic scale because DCL-I secondary siRNAs possess 5′ triphosphates ( since the 5′ ends are not generated by Dicer cleavage but by RNA synthesis ) and primary siRNAs have 5′ monophosphate . First we digested RNA extracted from virus infected plants with the terminator exonuclease and compared the level of vsiRNAs in treated and untreated samples by Northern blot . A similar level of vsiRNA in treated and untreated samples would suggest that most vsiRNAs carry 5′ triphosphates since the enzyme can only process molecules with 5′ monophosphate but not molecules with 5′ triphosphates . Our experiments revealed that most of the CymRSV vsiRNAs were processed by the enzyme when the samples were diluted thus we concluded that most vsiRNAs carry a 5′ monophosphate , although it cannot be ruled out that there is a small amount of vsiRNA with 5′ triphosphates . We also used another approach to study the 5′ structure of vsiRNAs where we profiled the short RNAs after ligating an adaptor to the 5′ end either directly or after de- and re-phosphorylation . The direct ligation only reveals vsiRNAs with 5′ monophosphate while the other approach also identifies vsiRNAs with 5′ triphosphate . If the profiles are different that suggests that there are vsiRNAs with 5′ triphosphate or diphosphate ends because they would be included in the dephosphorylated library but not in the directly ligated library . However , if the sequence profiles of the two libraries are the same that would indicate that there are not vsiRNAs with 5′ triphosphate or diphosphate because removing the 5′ phosphates does not have any effect . The patterns of vsiRNA hot spots in plants infected with the wild type virus were identical when the short RNAs were ligated to the 5′ adapter directly or after de- and re-phosphorylation . This argues against the idea that vsiRNAs in hot spots are mainly siRNAs with triphosphate 5′ end because in that case we should have seen a shift between the patterns obtained in the two experiments . The mutant virus also generated a similar vsiRNA profile regardless of the ligation method . To investigate the 5′ end of vsiRNAs we also used another approach that is based on the ability of the Terminator™ 5′-Phosphate-Dependent Exonuclease to distinguish between 5′ monophosphate and 5′ triphosphate ( i . e . only digests RNA with 5′ monophosphate ) . Since a negative result was possible in the terminator exonuclease experiments ( i . e . : no digestion of vsiRNAs ) , it was important to monitor the enzyme activity . A possible internal control is to monitor the level of miRNAs before and after exonuclease treatment since this class of short RNAs has 5′ monophosphate and present in the RNA extracted from virus infected plants . However , we observed during our preliminary experiments that different samples showed different ratios of digested vsiRNAs while the miRNA was always completely digested ( data not shown ) . We hypothesised that a big difference between vsiRNAs and miRNAs is their accumulation level . Small scale short RNA cloning and sequencing experiments found that the majority of the short RNA population in virus infected plants is virus derived [19] . Therefore we decided to use a synthetic siRNA duplex as an additional control . Another difference between miRNAs and vsiRNAs is that most miRNAs are incorporated into RISC quickly while most vsiRNAs are not co-purified with RISC [24] . As a consequence , most miRNAs are present as single stranded RNA and most vsiRNAs accumulate as dsRNA [19] . This raised the question whether the terminator exonuclease have lower affinity to dsRNA than ssRNA . We characterised the terminator exonuclease using a perfect duplex of two 19mer in vitro synthesised and phosphorylated RNA oligonucleotides . We found that the enzyme did not digest the non-denatured duplex at all . The duplex was partially digested following denaturation but only at lower concentration . The fact that the duplex was not digested at all at high concentration suggests that short RNAs can re-anneal very quickly but renaturation happens less frequently at lower concentrations therefore the enzyme activity increased dramatically . These results demonstrate that any RNA sample analysed by the terminator exonuclease has to be denatured and digested at different concentrations . This conclusion has an important implication beyond the plant virology field since the terminator exonuclease can be used to study the 5′ structure of RNA molecules in diverse systems , including endogenous siRNA biogenesis in animals . Interestingly , vsiRNAs were digested with a similar efficiency without or after denaturation , which is different from the efficiency of the digestion of a perfect duplex and very similar to the digestion pattern of imperfect duplexes . This suggests that vsiRNAs are not perfect duplexes but there are mismatches between the two strands . The most likely explanation is that most vsiRNAs are not generated from long perfectly matched double stranded RNAs but from precursor-miRNA like imperfect intramolecular structures . This is also supported by previous observation [19] and by the very strong strand bias of sequenced vsiRNAs: 93% of sequences obtained by the Solexa profiling derived from the positive strand . Theoretically , the imperfect duplex controls could be used to determine the number of mismatches in the vsiRNAs . However , one should know the concentration of vsiRNAs in order to compare with the concentrations of the imperfect duplexes . Unfortunately , this is very difficult since the vsiRNAs are part of the total RNA , which also contains many different plant short RNAs . Identifying the precursor-miRNA like structures is not easy on long RNA molecules . Qi et al ( 2009 ) searched for stem-loop structures around hot spots of vsiRNAs on the TMV-Cg genome but did not find these structures [28] . However , this approach ignores the fact that distant regions of a long RNA can anneal to each other . Therefore instead of looking for local secondary structures , we identified imperfect duplexes in the sequenced library that can derive from the + or − strand . Although these imperfect duplexes do not prove that vsiRNAs derive from structured single stranded viral RNA , all our results support this model over other possibilities . It was shown in TCV ( turnip crinkle virus ) and TRV infected Arabidopsis plants that DCL4 is the primary nuclease to produce vsiRNAs but if DCL4 is absent or suppressed DCL2 can also produce vsiRNAs [7] . Based on this we expected to find mainly 21 nt vsiRNAs produced by DCL4 . In wild type virus infected plants indeed the 21 nt class was the dominant vsiRNA class followed by the 22 nt vsiRNAs . However , in the p19 mutant virus infected plants we found more 22 nt than 21 nt vsiRNAs . The first conclusion is that the viral structures are probably recognised by multiple DCLs . DCL4 predominantly produces 21 nt vsiRNAs and DCL2 mainly generates 22 nt vsiRNAs [7] . There are at least two possible explanations for the difference in the dominant size class of vsiRNAs between mutant and wild type virus infected plants . One possibility is that the 22 nt sequences are the consequence of primary vsiRNA activity , which is inhibited by p19 and therefore less 22 nt sequences are produced in wild type virus infected plants . If the 22 nt sequences are the products of primer-dependent secondary siRNA production , they would be expected to show a similar ratio of positive to negative strand since they would be produced from a perfect dsRNA consisting of the positive and negative strand of viral genome . Alternatively , 21 and 22 nt sequences are produced both in the presence and absence of p19 ( i . e . independently of primary vsiRNA activity ) from the same substrate RNA ( i . e . mainly the positive viral RNA strand ) but because p19 binds the 21 nt duplexes more strongly , these are preferentially stabilised by p19 in wild type virus infected plants . In this scenario , both 21 and 22 nt sequences would show a similar , high positive/negative strand ratio . By plotting the 21 and 22 nt sequences separately for all four solexa libraries we found that the profiles of 21 and 22 nt sequences were very similar to each other and both class displayed a high positive to negative strand ratio . This indicates that the 22 nt sequences are not the consequence of primary vsiRNA activity . In fact the plots suggest that more 22 nt sequences are generated from the viral RNA but because p19 binds to 21 nt duplexes more efficiently , the shorter duplexes are preferentially stabilised in wild type infected plants . The fact that in the absence of p19 the 22 nt sequences accumulate at a higher level than the 21 nt class raises the possibility that DCL2 is dominant over DCL4 in N . benthamiana at least in the case of processing CymRSV RNA . However , this hypothesis has to be tested using different dcl mutants but these do not exist in N . benthamiana at the moment . It was shown before that vsiRNAs do mediate cleavage of viral RNAs [24] . We asked the question whether viral sequences complementary to abundant vsiRNAs ( i . e . : generated from hot spots ) are targeted more efficiently than sequences that are recognised by vsiRNAs generated from non-hot spot regions . Interestingly , the abundance of vsiRNAs complementary to the sensor constructs did not show any correlation with targeting efficiency of the different sensors . There are several possible explanation for this . First , if hot spot regions form partially double stranded structures with other regions of the viral RNA it would mean that these regions are less efficiently targeted by vsiRNAs since it was reported that RNAs with strong local structures are less accessible for RISC mediated cleavages [18] , [29] . The other potential explanation is that primary vsiRNAs are not good effectors of gene silencing . Indeed only small portion of vsiRNAs were found in high molecular weight complexes [24] . Moreover a recent report demonstrated that vsiRNAs produced in the absence of RDR6 are poor effectors of gene silencing [30] . We showed that most CymRSV vsiRNAs are primary siRNAs including the ones derived from hot spots , although , it is possible that the vsiRNAs , which efficiently target the viral RNA are the small amount of secondary siRNAs ( DCL-D or DCL-I ) that may be generated in CymRSV infected plants . However , we do not have evidence for this latter possibility .
Adaptors were removed from both 454 and Solexa samples by searching for the last eight bases of the 5′ adaptor ( in the case of 454 samples ) and the first eight bases of the 3′ adaptor ( for both 454 and Solexa samples ) using the UEA sRNA tools adaptor removal application [31] The sequence after the 5′ adaptor match and before the 3′ adaptor match was retained for further analysis . vsiRNA sequences were mapped to the CymRSV genome using PatMaN [32] allowing a single mismatch and zero gaps . vsiRNA abundances in each sample were normalised by dividing each count by the total number of trimmed reads for a given sample and then multiplying this value by one million . This gave a value of counts per million reads and ensured that the profiles were comparable even though the sample sizes varied . Profiles were plotted for each sample by taking the sum of the normalised abundance for each vsiRNA sequence covering a given nucleotide position on the viral genome . Boundaries of hot spot regions were initially determined using a simple algorithm designed to detect peaks in the profile then checked manually on the graphical plots . Sequence logos ( Figure S7 ) were drawn for vsiRNA size classes of 20 , 21 , and 22 nucleotides , respectively , using the program seqlogo from the WebLogo package version 2 . 8 . 2 [33] . A Perl script was written to identify vsiRNAs that could potentially base-pair with each other . Firstly sRNA sequences from the CymRSV PHOS Solexa sample were mapped to the plus strand of the CymRSV genome using PatMaN , allowing for one mismatch to the reference genome . All sequences matching to the plus strand of the CymRSV genome were then given as input to the script which used FASTA3 [34] to search each vsiRNA against the reverse complement of all plus strand genome matching sequences . Any resulting matches were then aligned to the query sequence using ClustalW [35] before finding the complement of the hit sequence to obtain the correct orientation . Any potential duplexes with four or fewer mismatches including a maximum of one bulge were accepted as potential pairs . This process was then repeated separately for the sequences matching to the minus strand of the CymRSV genome . A . tumefaciens infiltration was carried out according to Silhavy et al . ( 2002 ) . For coinfiltration of N . benthamiana leaves , mixtures of strains carrying sensor constructs ( OD600 = 0 . 15 ) and strains carrying suppressor constructs ( OD600 = 0 . 3 ) were used . Total RNA from Agrobacterium-infiltrated N . benthamiana and virus infected plant leaves was isolated using Trizol reagent [36] . Denaturing RNA gel blot hybridisation and analyses were done as described previously [37] . All sensor constructs were prepared from the previously reported 35S-green fluorescent protein ( GFP ) binary plasmid [38] . First , a SmaI restriction site was inserted downstream of the GFP ORF by using the QuikChange site-directed mutagenesis kit ( Stratagene ) by following the instruction manual . The PCR fragment of 200 bp , corresponding to the indicated positions ( Figure 5 ) of the CymRSV genome ( accession no . NC 003532 ) , was amplified by using appropriate 5′-phosphorylated oligonucleotides and placed into the unique SmaI site of the modified 35S-GFP plasmid . The plus and minus orientations of the inserts were selected , thus generating the different sensor constructs shown in Figure 5 . In vitro transcription of CymRSV and Cym19stop RNAs from linearized template plasmids and inoculation of RNA transcripts onto Nicotiana benthamiana , N . benthamiana GFP16c/RDR6i line [15] and N . clevelandii plants were performed as described previously [39] . Total RNA was extracted from 100 mg of systemicly infected leaf tissue [40] . Briefly , the homogenized plant materials were resuspended in 600 µL of extraction buffer ( 0 . 1 M glycine-NaOH , pH 9 . 0 , 100 mM NaCl , 10 mM EDTA , 2% SDS , and 1% sodium lauroylsarcosine ) and mixed with an equal volume of phenol . The aqueous phase was treated with equal volumes of phenol and chloroform , precipitated with ethanol , and resuspended in sterile water and used in subsequent reactions . Small RNA between 19–24nt were cloned from systemic leaves of N . benthamiana as described by Moxon et al . [41] . Briefly , the sRNA fraction was purified and ligated to adaptors without de-phosphorylating and re- phosphorylating the sRNA . The RNA was reverse transcribed , and amplified by PCR before being sent to 454 Life Sciences for pyrosequencing . The small RNA libraries are submitted to GEO and can be accessed through a super series number: GSE17278 . Small RNA fraction ( 19–24 nt ) of total RNA extracted from systemic leaves of N . benthamiana was isolated from 15% denaturing polyacrylamide gel . The eluted sRNA fraction was divided into two and one half was not treated ( called CymRSV-PHOS and Cym19stop-PHOS ) while the other half was de-phosphorylated with Shrimp Alkaline Phosphatase and re- phosphorylated with T4 Polynucleotide Kinase ( called CymRSV-CIP and Cym19stop-CIP ) . The resulting sRNAs were ligated to adaptors in the following reaction . The purified , adaptor ligated short RNAs were converted to DNA by RT-PCR and the DNA was sequenced on a Solexa platform ( Illumina ) . The small RNA libraries are submitted to GEO and can be accessed through a super series number: GSE17278 . For preparation of 19–24 nt RNA fraction , 15 to 20 µg of total RNA from in vitro transcribed virus infected plants was subjected to electrophoresis through an 8% denaturing polyacrylamide gel followed by staining in 1× Tris-borate-EDTA and 0 . 5 µg/mL ethidium bromide solutions for 5 min . The 19–24 nt fraction was visualized by UV light and excised from the gel . The gel slice was crushed , covered with 2 volumes of elution buffer ( 80% formamide , 40 mM Pipes , pH 6 . 4 , 1 mM EDTA , and 400 mM NaCl ) and incubated overnight . The gel residues were pelleted by centrifugation and the supernatant was precipitated with ethanol . The RNA ( ∼100 ng ) was dephosphorylated and then labeled in a 10-µl reaction in the presence of γ-32P-ATP and RNasin with 8 units of T4 polynucleotide kinase . The labeled 19–24 nt RNAs were used for hybridization either to membranes ( Zeta-Probe GT , BioRad ) containing five hundred 21mer DNA oligonucleotides in a one nucleotide sliding window covering a 520 nt region ( nucleotides 2650–3169 on the positive strand , see oligonucleotide sequences in Table S3 ) or to a membrane containing hundred eighty-six 21mer DNA oligonucleotides in a one nucleotide sliding window covering a 206 nt region ( nucleotides 4300–4505 on the positive strand , see oligonucleotide sequences in Table S3 ) . As a control for cross hybridization two set of oligonucleotides , which were one to five nucleotides shorter at their 5′ ends were also spotted on the membranes ( 12-1 , 12-2 , 12-3 , 12-4 , 12-5 , 87-1 , 87-2 , 87-3 , 87-4 and 87-5 ) . As a negative control to check background hybridization we spotted pUC/M13 Forward and Reverse primer onto the membranes or only water . Hybridization was performed in Ambion ULTRAhyb-Oligo Buffer at 37°C following the manifacturer's instruction . Signals were quantified with a Genius Image Analyzer ( Syngene ) . 19mer and 21 mer synthetic RNA molecules were phosphorylated and 30 µl of sense RNA ( 10 µM ) and 30 µl of antisense RNA ( 10 µM ) were incubated in 90 µl annealing buffer ( 30 mM HEPES-KOH pH 7 . 4 , 100mM KCl , 2mM MgCl2 , 50mM ammonium acetate ) to form siRNA duplex ( 19 nt siRNA , siRNA171a and siRNA171b ) . We also prepared mismatch containing siRNA duplexes on the same way by incubating sense RNA with different mismatch containing antisense RNA ( 19mer antisense RNA 1mism_end , 19mer antisense RNA 2mism_end , 19mer antisense RNA 3mism_end , 19mer antisense RNA 3mism_middl , 2mm antisense siRNA171a and 3mm antisense siRNA171c ) . We digested the resulting siRNA duplexes for one hour at 30°C in decreasing concentration ( 2 µM , 0 , 4 µM , 0 , 2 µM and 0 , 04 µM or 0 , 1 µM , 0 , 02 µM , 0 , 004 µM and 0 , 002 µM and as a control 2 µM or 0 , 002 µM sense phosphorylated RNA oligo ) in a 10 µl reaction in the presence of 0 . 25 U Terminator™ 5′-Phosphate-Dependent Exonuclease ( Epicentre Biotechnologies ) without or after 1 minute 90°C denaturation . Terminator treated and non-treated samples were loaded onto 0 . 5× TBE 15% UREA-Polyacrylamide gel . After electrophoresis the samples were transferred to Zeta-probe GT BioRad membranes by Semi-dry blotting . The membranes were hybridized with γ-32P-ATP labelled DNA oligonucleotide ( 19mer sense detector , siRNA171a sense detector and siRNA171b detector ) complementary to the 19mer sense RNA or to the sense RNE of siRNA171a or to the sense RNA of siRNA171b respectively . Hybridization was performed in Ambion ULTRAhyb-Oligo Buffer at 37°C following the manifacturer′s instruction . RNA extracted from CymRSV infected plants was mixed with 19mer siRNA duplex and was diluted containing decreasing concentration of total RNA and siRNA duplex ( 2 , 3 µg RNA+2 µM siRNA duplex , 0 , 46 µg RNA+0 , 4 µM siRNA duplex , 0 , 23 µg RNA+0 , 2 µM siRNA duplex , 0 . 046 µg RNA+0 , 04 µM siRNA duplex , and as a control 2 , 3 µg RNA+2 µM sense phosphorylated RNA oligo ) . We digested the resulting RNA and siRNA duplex mixture in a 10 µl reaction in the presence of 0 . 25 U Terminator™ 5′-Phosphate-Dependent Exonuclease without or after 1 minute 90°C denaturation for one hour at 30°C . Terminator treated and non-treated samples were loaded onto 0 . 5× TBE 15% UREA-Polyacrylamide gel . After electrophoresis the samples were transferred to Zeta-probe GT BioRad membranes by Semi-dry blotting . The membranes were hybridized with γ-32P-ATP labelled DNA oligonucleotides . Cym mix probe was a pool of five oligonucleotides ( cym3125 , cym3196 , cym3420 , cym3950 and cym4240 ) and complenetary to the negative strand of the virus genome . Cym3025 was also complementary to the negative strand of the virus genome . Cym minus probe was a pool of two oligonucleotides ( cym1021 minus and cym3710 minus ) and complementary to the positive strand of the virus genome . Cym 4632 LNA was also complementary to the positive strand of the virus genome . U6 detects the U6 snRNA while miR159 LNA detects the mature strand of miR159 and contains locked nucleic acid ( LNA ) nucleotides . The oligonucleotide sequences used for hybridisation are available in Table S3 . Hybridization was performed in Ambion ULTRAhyb-Oligo Buffer at 37°C following the manifacturer's instruction .
|
Viral RNA is processed into short RNAs in plants , which guide a complex to the viral RNA and cause cleavage of the viral RNA . We profiled Cymbidium ringspot virus ( CymRSV ) derived short RNAs using three different methods . Profiling of viral short interfering RNAs revealed a different sequence bias for the 454 and Solexa high-throughput sequencing platforms . We also found that viral short RNAs are primarily produced from the positive strand of the virus and produced with very different frequency along the viral genome . The hybridisation approach showed that the profile of viral short RNAs is determined by the virus itself because the profiles were the same in different species and it also showed that the process was RDR6 independent . We used the Terminator exonuclease to study the 5′ end of viral short RNAs and discovered that this enzyme cannot digest double stranded RNA . A control perfect duplex was only partially processed even after denaturation . Since double stranded short RNAs can quickly re-anneal , this assay must be carried out using different concentrations of the substrate . We found that most of the CymRSV short RNAs had 5′ monophosphate and were not perfect duplexes . Taken together , these results suggest that CymRSV short RNAs are produced from the structured positive strand rather than from perfect double stranded RNA or by RNA dependent RNA polymerase . We also found that regions from the viral genome that are not complementary to highly abundant viral short RNAs were targeted in the plant just as efficiently as regions recognised by abundant short RNAs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"plant",
"biology/plant-biotic",
"interactions",
"virology/host",
"antiviral",
"responses"
] |
2010
|
Structural and Functional Analysis of Viral siRNAs
|
Anaplastic Lymphoma Kinase ( Alk ) is a Receptor Tyrosine Kinase ( RTK ) activated in several cancers , but with largely unknown physiological functions . We report two unexpected roles for the Drosophila ortholog dAlk , in body size determination and associative learning . Remarkably , reducing neuronal dAlk activity increased body size and enhanced associative learning , suggesting that its activation is inhibitory in both processes . Consistently , dAlk activation reduced body size and caused learning deficits resembling phenotypes of null mutations in dNf1 , the Ras GTPase Activating Protein-encoding conserved ortholog of the Neurofibromatosis type 1 ( NF1 ) disease gene . We show that dAlk and dNf1 co-localize extensively and interact functionally in the nervous system . Importantly , genetic or pharmacological inhibition of dAlk rescued the reduced body size , adult learning deficits , and Extracellular-Regulated-Kinase ( ERK ) overactivation dNf1 mutant phenotypes . These results identify dAlk as an upstream activator of dNf1-regulated Ras signaling responsible for several dNf1 defects , and they implicate human Alk as a potential therapeutic target in NF1 .
Receptor Tyrosine Kinases ( RTKs ) are transmembrane proteins with intrinsic kinase activity directed in part towards tyrosine residues within their own carboxy-terminal tails . They play pivotal roles in most tissues , including the central nervous system ( CNS ) , by transducing extracellular ligand binding events into intracellular signals . A major signaling pathway activated by RTKs is the Ras/ERK ( Extracellular signal Regulated Kinase ) cascade [1]–[4] . Initially thought to be mostly involved in cell proliferation and differentiation , recent work has increasingly implicated various components and regulators of this signaling cascade in neuronal plasticity and memory formation [4] . However , although most RTKs should , in principle , be able to activate Ras/ERK signaling , only few among the 58 human receptors have been functionally linked to cognitive processes [5] . Even in Drosophila , a system with powerful genetics and resident homologs of most mammalian RTKs [6] , evidence implicating these receptors in learning and memory remains scant [5] . The linotte/derailed RTK ( an ortholog of RYK ) is the only Drosophila family member implicated in learning and memory to date [7] . However , deficits in adult neuroplasticity associated with mutations in this gene appear at least partially attributable to abnormal brain development [8] . In addition , linotte/derailed is an atypical RTK , devoid of intrinsic kinase activity . Evidence suggesting involvement of at least one typical RTK in olfactory associative learning and memory in the fly comes from work on Drk , an adaptor protein that binds active tyrosine phosphorylated receptors [1] , [2] . Reducing Drk levels results in defective olfactory learning and memory [9] , suggesting that at least one RTK may be involved in this process . To identify RTKs potentially involved in Drosophila learning and memory , we determined the family members which are expressed in the adult CNS . The fly ortholog of Anaplastic Lymphoma Kinase ( Alk ) was among genes showing prominent expression in this screen . Vertebrate Alk , and its dAlk Drosophila ortholog , are members of the insulin receptor subfamily of RTKs , [10] , [11] . Two related secreted proteins , pleiotrophin and midkine , can activate vertebrate Alk , although whether they do so directly by interacting with Alk , or indirectly by modulating the activity of a transmembrane tyrosine phosphatase , remains controversial [11] . As for most RTKs , Alk activation results in the recruitment of adaptor proteins , such as IRS-1 , Shc and FRS2 and initiation of intracellular signaling pathways , including the canonical Ras/ERK cascade [11] , [12] . Aberrant activation of the Alk kinase by chromosomal translocations or point mutations has been causally implicated in anaplastic large cell lymphoma , non-small cell lung cancer , and neuroblastoma [11] , [13]–[17] . Alk signaling may also be a rate limiting factor controlling the growth of glioblastoma cells [14] and non-synonymous polymorphisms in the gene may be associated with schizophrenia [18] . While recent reports have generated much excitement about Alk as a therapeutic target in lung cancer [19] , the normal roles of vertebrate Alk remain poorly understood [11] . Drosophila dAlk functions in visceral muscle development in the embryo [20]–[22] , in axonal targeting in the retina [23] and in synaptic signaling at the larval neuromuscular junction [24] . Although the Drosophila miple1 and miple2 genes predict pleiotrophin- and midkine-related proteins , the bona fide dAlk-activating ligand is the secreted protein Jelly belly ( Jeb ) [11] . As reported here , we found dAlk to be widely expressed in the adult brain , but to be especially abundant in the calyces of the mushroom bodies ( MBs ) , neuronal structures essential for olfactory learning and memory [25] , where Drk is also preferentially expressed [9] . Prompted by these observations , we investigated whether dAlk functions in associative learning . Our results identify dAlk as the first active RTK involved in olfactory learning , but also in body size determination . Intriguingly , dAlk shares both of these disparate functions with dNf1 , the ortholog of the human neurofibromatosis type 1 ( NF1 ) tumor suppressor gene . NF1 affects 1 in 3 , 000 individuals worldwide and is the most common genetic disease associated with an increased cancer predisposition . Hallmark NF1 tumors are benign neurofibromas and malignant peripheral nerve sheath tumors . Learning defects in most children with NF1 , skin pigmentation abnormalities , and short stature are among other common symptoms of this multi-system disorder . NF1 is caused by loss of function mutations in a gene , termed NF1 , encoding an evolutionary conserved Ras GTPase Activating Protein ( GAP ) , termed neurofibromin . While excess Ras signaling due to loss of neurofibromin is widely believed to be responsible for most , if not all NF1 defects , no effective therapy for any NF1 defect has yet been devised . Moreover , Ras is a central component of multiple signaling pathways , and which of these specifically contribute to disease development remains only partially understood [26] , [27] . Drosophila neurofibromin is approximately 60% identical to the human protein , and dNf1 null mutants are viable , fertile , and normally patterned , but reduced in size and defective in associative olfactory learning and memory [28]-[31] . Previous work showed that the size and learning defects reflect roles for NF1 in larval and adult neurons respectively , and that loss of dNf1 is associated with neuronal ERK overactivation [31] . However , no upstream activator of dNf1 regulated Ras-ERK signaling has yet been identified [32] . Here , we present evidence that dAlk functions as an activator of dNf1-controlled neuronal Ras/ERK pathways , responsible for defects in body size determination during larval development and in olfactory learning in adult flies . Our results argue that Alk may provide a therapeutic target for human NF1 .
An expression profiling screen revealed robust expression of dAlk mRNA in the adult CNS ( data not shown; see Materials and Methods for details ) . To confirm and extend this finding , adult brain sections were immunostained with an anti-dAlk antibody [33] . These experiments identified prominent dAlk staining throughout the neuropil , whereas cortical areas , including the regions occupied by mushroom body cell bodies , stained to a much lesser extent ( Figure 1A1-1A4 ) . Particularly strong staining was observed in mushroom body calyces ( Figure 1A1 ) , compared to near background level staining in other mushroom body parts , including the α , β and γ lobes ( Figure 1A3 and 1A4 ) , or the pedunculus ( Figure 1A2 ) . Other dAlk staining positive CNS structures included the optic lobes ( Figure 1A1 and 1A2 ) , the protocerebral bridge ( Figure 1A2 ) , the antennal lobes , the suboesophageal ganglion ( Figure 1A3 and 1A4 ) , the medial bundle and lateral horns ( not shown ) . In contrast , staining similarly treated sections incubated in the absence of the primary antibody did not reveal any signal , demonstrating the specificity of the anti-dAlk antibody and of the staining pattern ( Figure S4A ) . The intense staining of mushroom body calyces may reflect a specific accumulation of dAlk in dendrites , and the same dendritic accumulation may be reflected by the neuropil staining . Because homozygous dAlk mutants die as early larvae [33] , we used two independent RNA-interference ( RNAi ) transgenes and the TARGET system [34] , to abrogate dAlk . Both AlkRNAi transgenes were specifically expressed in the adult CNS with the Elav-Gal4 driver ( See Materials and Methods ) and resulted in significant reduction of dAlk ( Figure 1B ) . To confirm the specificity of dAlk immunostaining , we next abrogated dAlk expression specifically in mushroom bodies , using the c772-Gal4 driver . This resulted in a specific reduction of the dAlk signal in mushroom body calyces , whereas staining in the remaining neuropil remained largely unaffected ( Figure 1C ) . In conjunction with the results in Figure S4A , we conclude that the antibody is highly selective for dAlk and that the AlkRNAi transgene allows effective ablation of dAlk expression . Previous work indicated that dAlk can activate ERK in larval imaginal disks [33] . To determine whether the same is true in the adult CNS , we analyzed phospho-ERK ( p-ERK ) levels in flies expressing various dAlk or Jeb transgenes . Transgenes designed to enhance dAlk signaling included those encoding wild type ( AlkWT ) [33] , a constitutively active truncated form of the protein ( AlkCA ) , or the dAlk-activating ligand Jeb [23] . Conversely , the AlkRNAi transgene , or a transgene encoding a dominant negative AlkDN mutant [23] , were used to block dAlk signaling . In addition to Elav , we used the Ras2-Gal4 driver [31] , which although not as widespread , is also expressed in the majority of CNS neurons ( Figure S1C and S1D ) . Phospho-ERK levels were significantly elevated in head lysates of flies expressing AlkWT , AlkCA , or Jeb under Ras2-Gal4 and reduced in flies expressing the interfering AlkDN or AlkRNAi transgenes . Thus , dAlk modulates ERK activation in the adult CNS , likely by engaging the Ras/ERK cascade , as previously demonstrated for the vertebrate protein [11] . To investigate whether dAlk plays a role in associative learning as suggested by its presence in the mushroom bodies , we again used the TARGET system to modulate dAlk expression specifically in the adult CNS , thus avoiding potential defects stemming from its known developmental roles [22]–[24] . Surprisingly , adult-specific pan-neuronal expression of wild type dAlk , its constitutively active AlkCA mutant , or its activating ligand Jeb , caused significant deficits in associative olfactory learning ( Figure 2A-Induced ) . Because the only known function of Jeb is to activate dAlk , it appears unlikely that the observed learning impairment is a non-specific consequence of expressing the transgene ectopically . Moreover , it is unlikely that the attenuated learning upon dAlk over-expression or over-activation is due to ectopic expression , because the same effect was observed in response to transgenic Jeb elevation . Interestingly , interfering with endogenous dAlk activity with the dominant negative protein , or attenuating its levels in the CNS by RNAi , did not precipitate deficits , but rather reproducibly elevated performance ( Figure 2A , open bars , # p = 0 . 0016 ) . Learning deficits were not observed in flies raised at the restrictive temperature ( Figure 2A , ‘Uninduced’ ) , and task-relevant olfactory responses and shock reactivity were normal ( Table S1 ) . Similar learning deficits were observed upon increasing dAlk expression or activity using the Ras2 neuronal driver ( Figure 2B ) . Again , abrogating endogenous dAlk by expressing UAS-AlkRNAi , or attenuating its activity with AlkDN in Ras2-positive neurons yielded a modest , but significant ( # p = 0 . 01 ) learning enhancement ( Figure 2B ) . To ascertain that these effects were not a consequence of ectopic transgene expression , we generated a novel Gal4 driver Alk ( 38 ) -Gal4 . The pattern of Alk ( 38 ) -Gal4 expression did not overlap precisely with endogenous dAlk distribution in few tissues and cell types as commonly observed with reporter expression and may in part reflect differences in the localization of dAlk and the mCD8-GFP reporter in the membranes . However , Alk ( 38 ) -Gal4 was abundantly expressed and largely recapitulated the endogenous dAlk distribution in larval and adult CNS ( Figure S1A and S1B ) . Alk ( 38 ) -Gal4-mediated adult-specific induction of the above mentioned transgenes yielded similar results to those obtained with Elav and Ras2 ( Figure 2C ) . Therefore , increased dAlk levels or signaling impairs olfactory learning , whereas its attenuation improves it . Both Alk ( 38 ) and Ras2-Gal4 are expressed in the MBs , neurons that contain an abundance of dAlk in their dendrites ( calyces ) . Thus , we were surprised to find that expressing dAlk or Jeb transgenes with the MB247 and c772-Gal4 MB drivers did not affect associative learning ( Figure 2D and 2E ) . It is unlikely that the lack of effect is due to limited expression in the approximately 2000 MB neurons located in each adult brain hemisphere . In fact , the MB247 and c772 drivers are expressed in 1600 and 1800 MB neurons respectively and c772 in particular is also expressed in α3′/β3′ cells [35] . Therefore , despite its obvious presence in the calyces , dAlk does not appear to play a role in olfactory associative learning in most , if not all , MB neurons . Because enhanced learning is an uncommon phenotype [36] , we sought additional substantiation that dAlk inhibition caused this effect . To increase assay resolution , we limited the number of conditioned/unconditioned stimulus ( CS/US ) pairings , as described previously [9] . Limited training with just 3 pairings also resulted in enhanced learning when Alk was abrogated with the Ras2 ( Figure 2F ) , or Alk ( 38 ) -Gal4 drivers ( Figure 2G ) , validating this finding . Enhanced learning was also apparent after training with 6 , but not after the more intense 12 pairing training , suggestive of a performance “plateau” [9] . We conclude that inhibiting dAlk enhances performance after suboptimal training , apparently by increasing learning per CS/US pairing . This increased learning rate allows dAlk inhibited flies to reach a performance plateau faster than controls . Thus , dAlk appears to be a limiting negative regulator of olfactory associative learning and surprisingly this function seems to involve neurons outside the mushroom bodies . In contrast to the normal-sized animals obtained with use of the TARGET system , we noted that pan-neuronal expression of wild-type dAlk , or of the constitutively active dAlkCA throughout development yielded ∼10%–15% smaller pupae compared to isogenic controls ( Figure 3A ) . Similar results were obtained by over-expression of Jeb , which again argues against an artifact caused by ectopic dAlk expression , since Jeb should only activate its endogenous dAlk receptor . Further arguing against a non-specific effect , abrogation of endogenous dAlk activity by pan-neuronal expression of the dominant-negative dAlkDN or AlkRNAi transgenes yielded pupae that were ∼10%–15% larger than wild-type ( Figure 3A ) . Except for these conspicuous size differences , normally patterned and viable adults emerged from these pupae ( Figure S2 ) . Driving the dAlk and Jeb transgenes with Ras2 and Alk ( 38 ) -Gal4 caused similar size changes ( Figure 3B and 3C ) . Notably , increasing the levels of wild type dAlk in neurons normally expressing the protein with the Alk ( 38 ) driver resulted in larval lethality , suggesting that animals below a certain size threshold are not viable , possibly because of neuromusculature defects [22] , [24] . To further define the role of dAlk in body size determination , we sought to identify the types of cells in the larval CNS where it is required for this novel function . Therefore , we used 29 additional Gal4-drivers to drive the dAlk transgenes in various CNS cell types throughout development ( Table S2 ) . Body size was altered by manipulating dAlk expression or activity in cells marked by the peptidergic drivers 386Y-Gal4 ( Figure 3D ) , c929-Gal4 and the cholinergic Cha-Gal4 ( Table S2 ) . In contrast , increasing dAlk activity in glia , MBs , dopaminergic and GABAergic neurons did not affect pupal size ( Table S2 ) . Interestingly , transgenic Jeb expression in glial cells , GABAergic and dopaminergic neurons and the potentially peptidergic c316-Gal4-marked DPM neurons [37] , resulted in small pupae . These results are consistent with a paracrine mode of action for Jeb as previously suggested [22] , [38] . Significantly , driving dAlk in Insulin-producing cells ( IPCs ) , known to be involved in Drosophila body size determination [39] , [40] , also had no effect ( Figure 3E ) . This suggests that dAlk operates in a different , potentially independent size determination system than that requiring Insulin . Measuring wing surface areas confirmed that flies with altered body size possess proportionally sized wings ( Figure 4A and 4B ) . This likely reflects an altered size of the larval wing disk , and agrees with the notion that dAlk is required in the larval CNS to determine body size in a non-cell autonomous manner . To determine whether size differences reflect changes in cell size and/or cell number , we took advantage of the fact that each wing cell produces a single hair , which provides a convenient way of determining wing cell densities [41] . Comparing the cell densities of wings from Ras2-Gal4 driven transgenic flies to those of wild type controls , we found that size alterations are largely due to changes in cell size ( Figure 4C ) and not cell number ( Figure 4D ) . Collectively , these data demonstrate that dAlk and Jeb are novel non-cell-autonomous regulators of organismal growth , affecting cell size but not proliferation . Growth of insects occurs exclusively during larval development . Therefore , it was not surprising that pan-neuronal expression of dAlk specifically in adult flies with TARGET did not modify their size . Rather , it appears that the size-related function involves larval neuroendocrine and cholinergic neurons . This conclusion is supported by the spatial specificity of Gal4 drivers that modify size ( Table S2 ) , as well as by the prominent dAlk immunostaining ( Figure 5A ) in discrete parts of the larval ventral ganglion and the central brain where neuroendocrine and cholinergic neurons reside [42] . Notably , the Alk ( 38 ) driver is also expressed in these areas of the larval ventral ganglion ( Figure S1A ) . The size and learning defects observed upon dAlk over-activation are highly reminiscent of phenotypes exhibited by Drosophila dNf1 mutants [26] , [29]–[31] , [43] . An important goal of Neurofibromatosis 1 research has been to identify upstream activators of NF1 regulated Ras signaling , which might be exploited as potential therapeutic targets . To test whether dAlk plays such a role , we compared dAlk and dNf1 expression patterns , and we investigated whether the two genes interact genetically . In order for dAlk to function as an upstream activator of dNf1 regulated Ras signals , both proteins should at least partially colocalize . Using an antibody specific to dNf1 ( Figure 5C ) , we found that both proteins are broadly distributed in the larval CNS , and colocalize extensively in the larval ventral nerve cord , the brain lobes and particularly in the larval mushroom body calyces and the developing adult visual system ( Figure 5A ) . Both proteins are also broadly present throughout the adult brain and show similar extensive colocalization in mushroom body calyces , the protocerebral bridge , mushroom body satellite neuropil and ventral bodies ( Figure 5B ) . Furthermore , both proteins appeared in close proximity with the cell membrane in different neuronal types ( Figure 5 insets ) . To functionally test the colocalization of the two proteins , we took advantage of the fact that loss of dNf1 also results in a non-cell autonomous decrease in cell size yielding overall smaller pupae [30] , [31] and adults ( Figure S2 ) . Targeted re-expression of dNf1 in Alk ( 38 ) -Gal4-expressing cells reversed the size deficit of Nf1-null mutants ( Figure 6A ) indicating that dNf1 is indeed required in dAlk-expressing cells . Thus , dAlk and dNf1 show widespread and extensively overlapping expression in both the larval and adult CNS . Moreover , both proteins appear to function in overlapping neuronal populations to non-autonomously regulate organismal size . To substantiate this conclusion , we evaluated functional links between dAlk and dNf1 using genetic epistasis . Pupal size in Alk1-null and Alk9-kinase-dead heterozygotes is normal ( Figure 6B ) . However , heterozygosity for either Alk mutant allele increased significantly the small size of Nf1E2 homozygotes ( Figure 6B ) . This phenotypic suppression was a consequence of increased cell size and not cell number ( Figure S3A ) . Importantly , additional support for the hypothesis that dAlk and dNf1 function in the same growth regulating pathway was provided by pharmacological abrogation of dAlk with the selective inhibitor TAE684 [44] . Feeding 1–10 µM TAE684 to homozygous null Nf1E2 larvae throughout development rescued their size deficit . No phenotypic rescue was observed at 100 µΜ , indicating a sharp concentration optimum and perhaps reflecting an off-target effect at the higher concentration ( Figure 6C ) . Therefore , dAlk and dNf1 interact genetically to determine pupal size . To define more precisely the cells where dAlk and dNf1 function for size determination , AlkWT , AlkCA and Jeb transgenes were expressed with Elav , Ras2 and the more restricted 386Y-Gal4 in Nf1E2 mutants . All these manipulations caused larval lethality , suggesting that further reduction of the already small size of Nf1E2 is not tolerated ( crosses in Figure 6D ) . In agreement with this interpretation , pan-neuronal expression of AlkWT and AlkCA in Nf1E2/+ heterozygotes was also lethal , whereas Jeb expression enhanced their slight size reduction to 70% of controls . The Nf1E2/+ small size was also further reduced upon Jeb expression with Ras2 and 386Y , but not with the MB-driver OK107-Gal4 ( Figure S3B ) . Conversely , abrogation of dAlk signaling with AlkDN or AlkRNAi resulted in Nf1E2 homozygotes with significantly larger size , which approached control levels if driven with Elav . Driving the same transgenes with Ras2 and 386Y-Gal4 also increased pupal size , but to a lesser extent ( Figure 6D ) . Similar attenuation of dAlk activity in Nf1E2/+ heterozygotes rescued their smaller size completely ( Figure S3B ) . Importantly , rescue of the Nf1E2 small size by dAlk abrogation is the result of increased cell size and not cell number ( Figure 6E ) . These results provide additional support for the proposed functional interaction of the two proteins in size determination . In contrast , dAlk abrogation in MB neuroblasts and neurons with OK107-Gal4 , or in CCAP-producing neurons with CCAP-Gal4 , did not alter Nf1E2 size ( Figure 6D ) . Increased ERK activity as a result of dNf1 loss in larval Ras2-expressing cells has been proposed to result in the reduced body size in mutant homozygotes [31] . In agreement with this hypothesis targeted expression of an activated form of ERK ( UAS-rlsem ) [45] with Ras2-Gal4 , yielded smaller animals with wings that consisted of smaller , but not fewer cells ( Figure S3C ) . If our hypothesis that dAlk activates dNf1-regulated Ras/ERK signaling is correct , then attenuating dAlk activity should reduce the elevated neuronal phospho-ERK levels in Nf1E2 homozygotes . Indeed , abrogation of dAlk levels or activity in Ras2-expressing neurons of Nf1E2 homozygotes , restored phospho-ERK to control levels ( Figure 6F ) . Collectively , these results indicate that neuronal ERK activity levels control Drosophila size , and that attenuation of ERK over-activation by genetic or pharmacological reduction of dAlk activity rescues the small size phenotype of dNf1 mutants . Since dAlk ( Figure 2 ) and dNf1 [29] both affect associative learning independent of their developmental roles , we investigated whether the two proteins interact in neurons mediating this process . Immunostaining revealed that dNf1 is distributed in the adult brain in a pattern highly overlapping that of dAlk , exhibiting coincident staining in mushroom body calyces and the protocerebral bridge , among other regions ( Figure 5B ) . Interestingly , much like dAlk , dNf1 also appears conspicuously absent from the pedunculus and lobes of the MBs ( Figure S4 ) . Others recently reported the presence of dNf1 transcripts in mushroom bodies by in situ hybridization [28] . Our results confirm and extend this work , by providing the first documentation of the spatial distribution of dNf1 protein in the adult brain including the MBs . As previously reported for the dNf1P1 and dNf1P2 alleles [29] , Nf1E1 and Nf1E2 mutant homozygotes and Nf1E1/E2 heteroallelics also exhibit significant associative learning defects ( Figure S5A ) . In fact , upon training with 3 CS/US pairings , even the heterozygous null mutants exhibited a learning defect ( Figure S5B ) . In support of the notion that dNf1 and dAlk function in the same cells to regulate learning , the learning defect of dNf1E1/E2 animals was fully rescued by targeted re-expression of dNf1 within Alk ( 38 ) -Gal4-expressing cells ( Figure 7A ) . To investigate whether dAlk-mediated signals are potentially regulated by dNf1 presumably within Alk ( 38 ) -Gal4-expressing cells in the adult CNS , we attempted to dominantly suppress the learning deficit of Nf1E2 homozygotes by reducing the dAlk gene dosage , using the Alk1 null and the Alk9 kinase-dead alleles . Either mutant allele improved the associative learning of Nf1E2 mutants ( Figure 7B ) , consistent with the hypothesis that dAlk and dNf1 interact to mediate the process . Importantly , acute inhibition of dAlk with TAE684 [44] in adult Nf1E2 homozygotes nearly eliminated their learning deficit ( Figure 7G , 10 nM and 100 nM , p = 0 . 0013 and p<0 . 0001 respectively , compared to vehicle-fed Nf1E2 flies ) . Similar to what was observed in size rescue experiments , the pharmacological response to acute TAE684 administration had a clear optimum in the 10-100 nM range , whereas higher doses were ineffective ( Figure 7C ) . This pharmacological evidence strongly suggests that dAlk-mediated signals engage dNf1 and are required for normal associative learning in flies . To determine the neurons where these proteins are required for associative learning , dAlk was attenuated with AlkDN and AlkRNAi in defined neuronal groups specifically in adult Nf1E2 homozygotes ( Figure 7D ) . If these transgenes remained uninduced , learning was defective in these small-sized Nf1E1/E2 adults ( Figure 7D ) . However , pan-neuronal expression under Elav-Gal4 completely reversed the learning deficit of Nf1E1/E2 , and this was also attained with the Alk ( 38 ) and to a lesser extent with the Ras2 driver ( Figure 7D ) . Similar results were also obtained with a second independent UAS-AlkRNAi transgene ( Figure S5C ) . Consistent with the data indicating a function for dAlk in learning outside the MBs , rescue of the Nf1 learning deficit was not observed with c772-Gal4 ( Figure 7D ) . Importantly , over-expression of a wild-type dNf1 transgene in mushroom bodies also failed to improve performance of Nf1 mutants ( Figure 7D last graph , last column , p<0 . 0001 compared to control ) , arguing that learning requiring dAlk and dNf1 does not directly involve these neurons . This result was confirmed with additional mushroom body drivers , MB247 ( Figure 7E ) and OK107 ( Figure 7F ) . However , as with Elav [46] , both Alk ( 38 ) and Ras2-Gal4 which rescue the learning deficit are expressed in the MBs ( Figure S1 ) . Therefore , to unequivocally determine that dNf1 function is not required within the MBs for associative learning , we performed additional experiments using these drivers in combination with MB-Gal80 , which constitutively prevents transgene expression specifically in the mushroom bodies [47] . Transgenic expression of the dNf1 transgene under the pan-neuronal Elav driver resulted in complete rescue of the associative learning deficits of Nf1E1/E2 animals ( Figure 7G left side ) . Learning of Nf1E1/E2 animals was also significantly improved when UAS-dNf1 was driven with Alk ( 38 ) or Ras2-Gal4 , consistent with the notion that dAlk and dNf1 interact genetically within neurons expressing these drivers to mediate associative learning ( Figure 7G ) . These results confirm that dNf1 is not required within the MBs for normal associative learning . However , although significantly improved , learning of Nf1E1/E2 flies with UAS-dNf1 driven with Alk ( 38 ) and Ras2-Gal4 did not attain control levels as under Elav-Gal4 . This indicates that Ras2 and Alk ( 38 ) -Gal4 may not be expressed in a subset of Elav-Gal4 positive neurons where dNf1 is necessary for normal associative learning and complete rescue of the learning deficits in Nf1E1/E2 animals . Alternatively , the reduced rescue may reflect the lower expression level of Alk ( 38 ) and Ras2-Gal4 compared with that of Elav . In addition , full rescue of the Nf1E1/E2 learning deficit was observed upon adult-specific dAlk attenuation in Alk ( 38 ) -expressing neurons ( Figure 7D ) , but not upon expressing a dNf1 transgene in the same neurons . A potential explanation of this is that although dAlk functions outside the MBs , dNf1 may be required both outside and inside these neurons for normal associative learning . Experiments to resolve this issue and to define the minimal neuronal subset for full rescue of the deficit are currently ongoing .
The results presented here lead us to hypothesize that dAlk and dNf1 have opposite roles in controlling neuronal ERK activity during larval development , and therefore determine overall organismal size in a non-cell autonomous manner . In support of our hypothesis , dAlk and dNf1 co-localize extensively in larval neurons , both proteins control ERK activity , and both modulate growth by regulating cell size . In agreement with this conclusion , transgenic neuronal expression of the constitutively active ERK , RlSEM , is sufficient to reduce Drosophila size . dAlk is the second active RTK implicated in Drosophila growth control . Previous work demonstrated that the fly homolog of the insulin/IGF1 receptor dInr , regulates both body and organ size [40] , [41] , [50] . In peripheral tissues , dInr is activated by a family of Insulin-like proteins ( dILPs ) , leading to the activation of the IRS ( Chico ) , PI ( 3 ) K ( Dp110 ) , PTEN ( dPTEN ) , and Akt/PKB , ( dAkt1/dPKB ) , signaling pathway . Ablating the Insulin Producing Cells ( IPCs ) , or silencing the function of dInr pathway components in the larval CNS resulted in severe growth defects [40] , [41] , [50] . Notwithstanding the similar growth phenotypes , several lines of evidence argue that dAlk and dInr control growth in fundamentally different ways . Most importantly , organismal growth is impaired when dInr activity or signaling is reduced , whereas a similar phenotype is observed upon dAlk activation . Secondly , dAlk affects organism growth by specifically regulating cell size in a non cell-autonomous manner . In contrast , dInr signaling affects both cell size and number cell-autonomously and non-autonomously [40] , [50] . Finally , expression of Jeb or dAlk transgenes in neuroendocrine IPCs using the dILP2-Gal4 driver did not modify pupal size . Thus , although both dAlk and dInr RTKs are involved in body size determination , their mechanisms and sites of action are distinct . This interpretation is consistent with results with the C . elegans Alk homolog Scd-2 shown to function in parallel with or converge with TGF-β signaling , but act independently of the Insulin cascade in dauer determination [51] . However , given that dInr and dAlk are members of the same subfamily of RTKs , a potential explanation for the lack of rescue of dNf1 mutant homozygous larvae with systemic administration of 100 µM TAE684 ( Figure 6C ) , may be off-target inhibition of dInr at the higher drug concentration [44] . Interestingly , S6K ( dS6K ) resides on a downstream branch of the dInr/PI ( 3 ) K signaling pathway and regulates cell size without impinging on cell number [41] . Although the dS6K loss-of-function phenotype resembles the dAlk gain-of-function and dNf1 loss-of-function phenotypes , its mode of action is cell-autonomous . However , it is still tempting to speculate that dAlk and dNf1 ultimately affect neuroendocrine signals that affect dS6K activity in peripheral tissues . Increasing signaling through the cyclic AMP ( cAMP ) -dependent protein kinase A ( PKA ) pathway has been reported to suppress the size defect of dNf1 mutants [30] . This among other findings , have led some to propose that dNf1 regulates Ras activity and cAMP levels independently [52] , [53] . In contrast , an investigation of the cAMP/PKA sensitive dNf1 mutant growth defect argued that aberrantly upregulated Ras/ERK signaling in Ras2-expressing larval neurons was its proximal cause [31] . Our results further support the latter explanation implicating a Ras/ERK signaling defect downstream of dAlk as the cause of size defects in dNf1 mutants . Then , how could elevated cAMP/PKA signaling rescue decreased body size ? Because neuroendocrine signals can activate the cAMP pathway [54] , it is possible that defective dAlk/Ras/ERK signaling in dNf1 mutants may lead to a neuroendocrine deficiency , which is restored by increasing cAMP/PKA signals . In Drosophila , the dAlk/Jeb receptor-ligand pair has been shown to act in an antrerograde signaling pathway essential for assembly of the neuronal circuitry of the fly visual system [23] . However , loss of either Alk or Jeb did not appear to impair assembly of functional synapses with normal postsynaptic responses at the larval neuromuscular junction [24] , indicating that they do not participate in CNS development . In agreement , pan-neuronal , spatially restricted attenuation or unregulated activation of Alk throughout development did not appear to yield gross structural defects in the adult brain ( Figure 1 and not shown ) or alter naïve behavioral responses to the stimuli used for conditioning ( Table S1 ) . Hence , it is unlikely that the learning phenotypes we describe are the result of developmental alterations in the CNS . In fact , dAlk seems to be acutely required for normal learning as limiting modulation of its activity to the adult CNS results in phenotypes on its own and also modifies the learning deficits of dNf1 mutants . Moreover , the function of dAlk and dNf1 in associative learning is independent of body size as the learning reverted to normal by dAlk abrogation in the small-sized dNf1 mutant homozygotes . Interestingly , the C . elegans Jeb homolog Hen-1 is required non-cell autonomously in the mature nervous system for sensory integration and associative learning [55] . Collectively then , these studies in C . elegans , mice [55] , [56] and our data strongly support an evolutionary conserved role for Alk signaling in adult associative learning and memory . Elevated dAlk/Jeb signaling outside the MBs impaired olfactory learning , while its abrogation increased learning efficiency . These are results are consistent with the enhanced performance in a hippocampus-dependent task described for Alk knockout mice [56] . We propose then , that Alk signaling normally functions to limit the strength of the CS/US associations , in effect providing a putative threshold required to be overcome for specific and efficient association of the stimuli . A GABAergic neuron outside the MBs , the Anterior Paired Lateral ( APL ) , was recently reported to similarly suppress olfactory learning and its silencing yielded enhanced performance [57] . Interestingly , a decrease in presynaptic GABA release or abrogation of the GABAA receptor , RDL in the post-synaptic mushroom body neurons [57] , [58] resulted in enhanced learning . Whether Ras2-Gal4 is expressed in the APL neuron and dAlk also functions in this neuron to suppress learning are questions currently under investigation . Interestingly , a recent study [59] suggested that the learning defects of Nf1+/− mice are attributed to increased ERK-mediated phosphorylation of synapsin I in hippocampal inhibitory neurons and concomitant increase in GABA release . In accord , a GABAA receptor antagonist enhanced learning in Nf1+/− mice and controls , and reversed LTP defects in the mutants . Similarly , elimination of the dAlk-mediating inhibition in Drosophila Ras2-expressing neurons enhanced learning , potentially via attenuation of ERK phosphorylation . In support of this notion , we show that constitutive activation of ERK in adult Ras2-expressing neurons precipitates learning deficits ( Figure S3D ) . Collectively , these results together with the reported learning deficits of Drosophila synapsin mutants [60] , suggest that a mechanism similar to that proposed for vertebrates may also regulate Nf1-dependent learning in flies . In mice , a decrease in Nf1 levels in heterozygous mutants increased Ras/ERK signaling and precipitated Long-Term Potentiation ( LTP ) and spatial learning deficits [61] . These deficits were reversed upon genetic or pharmacological inhibition of Ras signaling [61] , [62] . Our own results demonstrate that dAlk inhibition reversed the impaired learning of dNf1 mutants and since this is the first ‘kinase-active’ RTK shown to be involved in this process in flies [5] , it provides independent support for Ras/ERK hyperactivation as causal of these learning defects . Then , how can the reported phenotypic reversal of Nf1 learning deficits by expressing the PKA catalytic subunit throughout the fly [29] be explained ? We hypothesize that the MBs are functionally downstream of the dAlk/dNf1 neurons and elevated PKA activity within the former could result in normal learning . Future work will focus on addressing the merits of these hypotheses regarding the mechanisms underlying the size and learning defects of dNf1 mutants . A recent report [28] suggested that dNf1 mRNA is found in the mushroom bodies and in agreement , our own immunohistochemical results demonstrate that dNf1 is present within the mushroom body calyces ( Figure 6 , Figure S4 ) . Protein synthesis-dependent memory defects in Nf1 mutants were rescued upon MB-limited expression of the same full-length transgene as we used herein [28] . Since we did not examine memory deficits in our work , this complements our data and suggests a function for dNf1 within the MBs . In contrast , our data indicate that dNf1 expression in the MBs is not sufficient for learning/3 min memory ( Figure 7G ) . Three common MB drivers ( Figure 7D , 7E , 7F ) including the most specific MB247 and the most broadly expressed OK107-Gal4 [35] , did not rescue learning in Nf1 mutants by expressing dNf1 . We suggest therefore that rescue described by Buchanan and Davis was mediated largely by c739-Gal4 transgene expression in neurons extrinsic to MBs where Elav , Ras2 and Alk ( 38 ) -Gal4 are expressed , perhaps in combination with expression within MB-intrinsic neurons . The neuronal circuits where dNf1 and dAlk are required for normal learning are the subject of ongoing investigations . Our study identifies dAlk as the first RTK to functionally interact with Nf1 in Drosophila , raising the important question whether a similar functional relationship exists in mammals . Suggestive evidence argues that this may indeed be the case . Thus , Alk and NF1 extensively colocalize in the mammalian CNS during the same developmental periods [11] , [26] , [43] . Additionally , excess Alk expression or activation has been reported in astrocytomas , gliomas , neuroblastomas and pheochromocytomas , in which loss of NF1 expression has also been found [11] , [13]–[17] , [63] . Based on our identification of Alk as a bona-fide RTK that initiates a Ras/ERK cascade regulated by Nf1 , this suggests that Alk inhibition may rescue not only the phenotypes reported here , but also other symptoms that have been previously associated with Nf1 loss and ERK over-activation . It was recently reported that knockdown of NF1 expression renders a neuroblastoma cell line resistant to retinoic acid-induced differentiation , and that NF1 deficient neuroblastoma tumors have a poor outcome [64] . Our results suggest that Alk inhibition may provide an intervention strategy in such cases . Finally , the findings reported here , combined with the lack of overt abnormalities in Alk knock-out mice [11] , [16] , [56] , provide a rationale for further explorations of Alk as a potential therapeutic target in NF1 .
Drosophila crosses were set up in standard wheat–flour–sugar food supplemented with soy flour and CaCl2 , and cultured at 25°C and 50% humidity with a 12 h light/dark cycle . The Alk1 and Alk9 , dNf1E1 and dNf1E2 mutants have been described previously [22] , [31] . Transgenic fly strains used in this work were: UAS-AlkWT , UAS-Jeb , UAS-AlkCA , UAS-AlkDN [20] , [23] , [33] , UAS-AlkRNAi ( Vienna Drosophila RNAi Center-11446 ) , UAS-AlkRNAiKK ( Vienna Drosophila RNAi Center-KK107083 ) , UAS-dNf1 [31] , UAS-rlsem [45] , UAS-mCD8:: GFP [65] , Gal80ts [34] . The MB-specific Gal80 ( MBGal80 ) , which drives expression predominantly in the MBs [47] was introduced into the UAS-dNf1 , dNf1E1 strain through standard genetic crosses . All strains were backcrossed into the Cantonised-w1118 isogenic background for six generations . The legend to Table S2 gives the origin of all Gal4-driver lines used in this study . Dissected adult w1118 brains were submitted to Trizol-based RNA extraction followed by oligo-dT primed Reverse Transcription as described previously [66] . Amplification of cDNAs encoding Drosophila RTKs was achieved with the polymerase chain reaction ( PCR ) using a set of 21 distinct forward/reverse primers designed with Oligo v6 . 71 software ( Molecular Biology Insights Inc . ) , using sequence information obtained from Flybase , Kinbase or NCBI GenBank . We probed for the following molecules that contain an RTK signature: dAlk , dTor , dRor , Nrk , sev , InR , btl , htl , Cad96Ca , Pvr , dRet , dnt , drl , Drl-2 , Eph , Ddr , otk , Tie , CG3277 , dEgfr and Wsck . All amplifications were performed in triplicate from two different pools of dissected brains for 24 and 32 cycles . RTKs whose cDNAs exhibited robust amplification at 24 cycles were selected for further investigation of their distribution in the adult brain using antibodies or reporter transposon insertions as available as a counter screen . RTKs with unequivocal presence in the MBs , or other spatial restriction in their distribution were selected for further characterization . Additional details and results from this screen will be presented elsewhere . The Alk ( 38 ) -Gal4 transgenic line was constructed by fusing a genomic region that includes the first ( non-coding ) exon and first intron of dAlk , to a cDNA encoding Gal4 , followed by the genomic region downstream of the dAlk coding region encompassing the 3′UTR . Firstly , a clone corresponding to the presumed promoter region upstream of dAlk , was generated by PCR using genomic DNA and the primers ALK-FOR and ALK-REV ( for primer sequences see below ) . These primers amplify a region extending from the first ( non-coding ) exon of the adjacent upstream gene , gprs , to 8 nt upstream of the ATG site of Alk ( i . e . includes the first exon , first intron and part of the second exon of dAlk ) . Secondly , a Gal4 cDNA was amplified from the pChS vector using the Gal4-FOR and Gal4-REV primers . Finally , the region encoding the 3′UTR was amplified using ALK 3′UTR-FOR and ALK 3′UTR-REV . Each PCR product was TA-subcloned into pCR2 . 1 and sequenced . The final Alk-Gal4 construct was made in a modified pCaSpeR1 vector containing the multiple cloning site of pBlueScript . Firstly , the upstream dAlk promoter was subcloned using Kpn I and Bam HI . Next , the Gal4 coding region was subcloned into the resulting construct with Bam HI and the orientation of the insert was verified . Finally , the dAlk 3′UTR was subcloned behind the Gal4 coding region using StuI and NotI . ALK-FOR GGTACCCACACAGAAAGCAGAAG ALK-REV GGATCCAGCTACACTTTTCACGTTT GAL4-FOR GGATCCAACATGAAGCTACTGTCTTC GAL4-REV AGGCCTTGCGGGGTTTTTCAGTATC ALK 3′-FOR AGGCCTTACGCGGAGCGATACAAG ALK 3′UTR-REV GCGGCCGCTGTGGAGCTCCTTTCTGGAG Restriction sites are underlined The selective Alk inhibitor TAE684 [44] was provided by Nathanael S . Gray , dissolved in DMSO and serial dilutions of stock solutions were prepared . For pupal measurements , the solution was subsequently mixed into 10 ml fly food to make the specific concentrations mentioned in the text . For behavioral experiments , the solution was mixed into 10 ml dead-dry-yeast paste and flies were fed for 48 h , and then transferred into normal fly-food vials 1 hour prior to conditioning . Green food dye ( two drops per 10 ml ) was added to monitor homogeneity and to ensure that larvae or adult flies were actually ingesting the food . No significant differences in the measurements of body size or learning performance were observed in flies fed without DMSO ( no vehicle ) or with DMSO alone ( 0 nM TAE684 ) , confirming that DMSO showed no adverse effects . Pupal size measurements were performed as described previously [31] , [41] . ∼50 pupae of each genotype were digitally photographed using a video-equipped stereoscope ( Zeiss stereoscope equipped with a Zeiss Axiocam CCD camera ) and measured using ImageJ software ( NIH Image 1 . 45 ) . Pupae were then allowed to eclose , and scored for sex . Measurements for ∼20 female pupae were used to calculate average size , standard deviations , and statistical significance . To allow for slight variations in experimental conditions , all controls were included in each experiment . For cell growth analyses , ∼20 wings of 20 different female flies of each genotype were dissected , placed on slides and digitally photographed under the same magnification . Images were captured as described above . ImageJ software was used to measure wing-blade surfaces viewed at 40X magnification . The size of intervein cells was obtained by initially counting the number of hairs in a rectangle of 0 . 02 mm2 ( cell density ) . In order to maintain fair morphological comparisons , the location of the rectangle was defined using wing vein landmarks ( between veins 3 and 4 of the dorsal wing-blade , up to the posterior cross vein ) . The reciprocal value of the cell density gives the cell size . The approximate number of cells in the whole wing was calculated by multiplying the wing surface by the cell density . Similar results were obtained with male flies . Experiments were replicated at least once with flies from different crosses ( biological replicates ) . Behavioral tests were performed under dim red light at 23°C–25°C and 70%–78% humidity . All animals were 2–6 days old , collected under light CO2 anesthesia one day prior to testing , and kept in food vials in groups of 50–70 at 23°C–25°C or 18°C as appropriate for strains with Gal80ts temporal restriction of transgene expression . They were transferred to fresh vials 1–1 . 5 hours before testing . Olfactory learning and memory in the negatively reinforced paradigm coupling aversive odors as conditioned stimuli ( CS+ and CS- ) with the electric shock unconditioned stimulus ( US ) [67] was used to assess learning and memory . The aversive odors used were benzaldehyde ( BNZ ) and 3-octanol ( OCT ) . For training , ∼50 flies were placed into a tube lined with an electrifiable grid and presented with air ( 500 ml/min ) for 15 s , the shock-associated odor carried in the air current for 1 min concomitant with 1 . 25 second shocks at 90 V delivered every 5 s . This was followed by delivery of air for 30 s , the control odor in the air current for 1 min , and air again for 30 s . The timing of stimulus delivery was kept proportional to that for the full 12 CS/US pairing protocol , such that 3 shocks were delivered in 15 seconds of continuous CS+ presentation , 6 pairings within 30 seconds and so on . Two groups of animals of the same genotype were trained simultaneously , one to avoid BNZ , the other OCT , while the complementary odorant was used as the respective control . The animals were transferred to a T-maze apparatus immediately and allowed to choose between the two odors converging in the middle for 100 seconds . Since the time between testing and the coupling of the conditioned with the unconditioned stimulus is 3 min , the initial performance assessment is that of 3 min memory , which we refer to as learning . Performance was measured by calculating an index ( PI ) , as the fraction of flies that avoided the shock-associated odor minus the fraction that avoided the control odor reflected learning due to one of the conditioning stimuli and represented half of the performance index . One performance index was calculated as the average of the half-learning indexes for each of the two groups of animals trained to complementary conditioning stimuli and ranges from 100 to 0 , for perfect to no learning , respectively . All behavioral experiments were carried out in a balanced design , where all genotypes involved in an experiment were tested per day . The experimenter was blind to the genotype . Behavioral experiments were replicated at least once with flies from different crosses and a different time period ( biological replicates ) . For experiments using the TARGET system ( flies bearing Gal80ts ) , all animals were raised at 18°C until adulthood and UAS-dAlk transgenes were induced maximally by placing 3–5-day old flies at 30°C for 48 h . The animals were kept at the training temperature ( 25°C ) for 30 min before training . To assess olfactory avoidance , naive animals were given 100 seconds to choose between one of the odors and air . The airflow in both arms of the maze was kept constant and equal to that used for testing conditioned animals . Avoidance to both odors was tested simultaneously for each strain and all strains used were tested in a given session . Avoidance is represented by a performance index , which is calculated as the fraction of flies that avoid the odorant minus the fraction of flies that do not . To assess the avoidance of animals to electric shock , the arms of the T-maze were lined with electrifiable grids . Naive flies were placed at the choice point and given 100 seconds to choose between an electrified and an inert grid . Throughout the choice period , 1 . 25 s shocks at 90 V were delivered to one arm and air was passed through both arms at the standard flow rate . Avoidance is measured by a performance index calculated as the fraction of flies that avoid the electrified grid minus the fraction of flies that do not . Again , all strains were tested in a given session . For paraffin sections ( Figure 1A ) , wild type animals were fixed in Carnoy's fixative ( 60% ethanol , 30% chloroform , 10% acetic acid ) for 4 hr at room temperature , treated with methylbenzoate for 12 hr , and embedded in paraffin . 6 µm sections were obtained , deparaffinized in xylene baths , rehydrated through 100%–30% ethanol series , blocked for 1 . 5 h in 10% normal goat serum in PBHT [0 . 02 M NaPO4 , 0 . 5 M NaCl , 0 . 2% Triton X-100 ( pH 7 . 4 ) ] and challenged with the rabbit anti-dAlk antibody ( 1∶1000 ) in blocking solution ( 5% normal goat serum in PBHT ) overnight at 4°C . The sections were washed in PBHT , and a 1∶400 dilution of the secondary antibody ( Vector Labs ) in blocking solution was applied at room temperature for 3 hr . Slides were washed in PBHT and exposed to HRP conjugated to streptavidin at a dilution of 1∶400 in PBHT . After a final PBHT wash , the HRP was reacted with a substrate solution of 1 mg/ml diaminobenzidine and 0 . 03% H2O2 in PBHT . The unreacted substrate was washed away with water and the slides were mounted with Glycergel ( DAKO ) . Whole-mount larval CNS and adult brains were dissected in cold PBS , fixed in 4% paraformaldehyde for 20 min , and permeabilized using 0 . 1% Triton X-100 in PBS . The primary antibodies were used as follows: rabbit anti-dAlk ( 1∶1 , 000 ) [33] , and mouse anti-Nf1 ( DNF1-21 ) ( 1∶5 ) [30] . The following secondary antibodies were used: Goat anti-mouse , anti-rabbit conjugated with Alexa-Fluor secondary antibodies ( 1∶400 , all from Molecular Probes ) . Confocal laser microscopy was performed using a Leica TCS SP5 Confocal system equipped with the Leica LAS AF image acquisition analysis software suite . Images were processed using ImageJ 1 . 45 ( NIH , Bethesda ) software . For detection of dpERK and total ERK levels , five adult heads or five larval CNS were homogenized in Laemmli buffer supplemented with protease and phosphatase inhibitors . Extract equivalent to one adult head or one larval CNS was loaded per lane and the primary antibodies were used at 1∶2 , 000 for mouse anti-pERK , ( Sigma ) , and 1∶2 , 000 for rabbit anti-ERK ( Cell Signaling ) . Mouse anti-tubulin ( Developmental Studies Hybridoma Bank ) at 1∶2 , 500 was used as an internal loading control . Four independent experiments were scanned , the band intensities were determined using ImageQuant 5 . 0 ( Molecular Dynamics ) and used to calculate ratios of dpERK/ERK . Untransformed ( raw ) data were analyzed parametrically with the JMP 7 . 1 statistical software package ( SAS Institute Inc . , Cary , NC ) as described before [9] . Following initial ANOVA , planned multiple comparisons were performed , using α = 0 . 05 . The level of significance was adjusted for the experimentwise error rate . Detailed results of all planned comparisons mentioned in the figure legends are shown in Table S3 . Data are shown as mean ± S . E . M .
|
Neurofibromatosis-1 ( NF1 ) syndrome is a common ( 1/3 , 000 births ) genetic disorder affecting multiple organ systems , including the nervous system . Its clinical features include short stature , learning disabilities , and several types of benign and malignant tumors . NF1 is caused by mutations that inactivate the NF1 gene , a crucial negative regulator of Ras signaling . Although unregulated Ras signaling is a hallmark of NF1 , the specific Ras signaling pathways responsible for disease development remain largely unknown . The Drosophila and human Nf1 genes are highly conserved; and , as in patients , mutant flies are smaller than usual and present deficient learning . Here , we identified the Drosophila Receptor Tyrosine Kinase dAlk as a negative regulator of organismal growth and olfactory learning . We show that excessive dAlk activation results in growth and learning defects similar to those of Nf1 mutants . Genetic suppression studies and pharmacological inhibition indicate dAlk as a critical upstream activator of Nf1-regulated neuronal Ras/ERK signals that contribute to size determination and learning . Importantly , our results strongly suggest that Alk represents a novel , highly specific , and promising therapeutic target in human NF1 .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion"
] |
[
"animal",
"models",
"molecular",
"neuroscience",
"neurobiology",
"of",
"disease",
"and",
"regeneration",
"drosophila",
"melanogaster",
"model",
"organisms",
"genetics",
"signaling",
"pathways",
"biology",
"neuroscience",
"learning",
"and",
"memory",
"genetics",
"and",
"genomics"
] |
2011
|
The Receptor Tyrosine Kinase Alk Controls Neurofibromin Functions in Drosophila Growth and Learning
|
In budding yeast , asymmetric cell division yields a larger mother and a smaller daughter cell , which transcribe different genes due to the daughter-specific transcription factors Ace2 and Ash1 . Cell size control at the Start checkpoint has long been considered to be a main regulator of the length of the G1 phase of the cell cycle , resulting in longer G1 in the smaller daughter cells . Our recent data confirmed this concept using quantitative time-lapse microscopy . However , it has been proposed that daughter-specific , Ace2-dependent repression of expression of the G1 cyclin CLN3 had a dominant role in delaying daughters in G1 . We wanted to reconcile these two divergent perspectives on the origin of long daughter G1 times . We quantified size control using single-cell time-lapse imaging of fluorescently labeled budding yeast , in the presence or absence of the daughter-specific transcriptional regulators Ace2 and Ash1 . Ace2 and Ash1 are not required for efficient size control , but they shift the domain of efficient size control to larger cell size , thus increasing cell size requirement for Start in daughters . Microarray and chromatin immunoprecipitation experiments show that Ace2 and Ash1 are direct transcriptional regulators of the G1 cyclin gene CLN3 . Quantification of cell size control in cells expressing titrated levels of Cln3 from ectopic promoters , and from cells with mutated Ace2 and Ash1 sites in the CLN3 promoter , showed that regulation of CLN3 expression by Ace2 and Ash1 can account for the differential regulation of Start in response to cell size in mothers and daughters . We show how daughter-specific transcriptional programs can interact with intrinsic cell size control to differentially regulate Start in mother and daughter cells . This work demonstrates mechanistically how asymmetric localization of cell fate determinants results in cell-type-specific regulation of the cell cycle .
At the Start transition in G1 , budding yeast cells integrate internal and external cues into an all-or-none commitment to a new round of cell division [1] , [2] . Cell division is asymmetric , producing a smaller daughter cell and a larger mother cell [3] . Mother cells progress through Start more quickly than daughter cells [3] , [4] . The regulation of G1 phase is composed of two independent modules separated by the nuclear exit of the transcriptional repressor Whi5 [5]: a cell size sensing module , which extends G1 in small cells to allow additional growth before Start [5] , and a subsequent size-independent module [5] , [6] . The fast and coherent transition between the two modules likely coincides with commitment to the cell cycle and is driven by transcriptional positive feedback [7] . The G1 cyclin Cln3 is the most upstream activator of the Start transition [8] , [9] , [10] , [11] , [12] and the main regulator of the size-sensing module . Cln3 initiates inactivation of Whi5 [13] , [14] and expression of SBF/MBF dependent genes , including the G1 cyclins CLN1 and CLN2 [9] , [11] , [12] , [15] , [16] . Subsequent positive feedback of Cln1 and Cln2 on SBF/MBF dependent transcription ensures fast and coherent commitment to the cell cycle [7] . Cell size control is thought to regulate the length of the G1 phase of the cell cycle [4] , [5] , [17] , [18] . In budding yeast , cell size control is readily detectable in daughter cells but much less obvious in mother cells . In part this is because mother cells are almost always born larger than daughters [3] , but it has also been shown that daughters are slower to pass Start than mothers even when both are made equally large ( greater than normal mother or daughter size ) [19] . This finding suggested some asymmetry in Start control between mothers and daughters beyond that due to different cell size; differential gene expression in mothers and daughters could provide such asymmetry . Regulation of gene expression is asymmetric in mother and daughter cells as a result of the daughter-specific localization of the transcription factors Ace2 and Ash1 . Ace2 enters mother and daughter nuclei during mitotic exit [20] , [21] . Asymmetric localization of Ace2 is due to the Mob2-Cbk1 complex [20] , [21] , [22] , which promotes nuclear retention of Ace2 specifically in the newborn daughter nucleus , leading to daughter-specific expression of a number of genes [20] , [21] , [22] , [23] , [24] . Daughter-specific localization of Ash1 is achieved through active transport of ASH1 mRNA to the bud tip and consequent preferential accumulation of Ash1 in the daughter nucleus [25] . Ash1 represses expression of the HO endonuclease gene responsible for mating type switching [26] , [27] , thus restricting HO expression to mother cells . Recently , Ace2 was shown to cause a daughter-specific G1 delay , acting indirectly through “Daughter Delay Elements ( DDE ) ” 5′ to the CLN3 coding sequence to reduce CLN3 expression in daughters [28] . In that work , it was proposed that this Ace2-dependent delay is the only reason that daughters have a longer G1 than mothers . Cell size was proposed to play no role in controlling the length of G1 [28] . This proposal is incompatible with our recent finding that small cells display very efficient size control , requiring a significantly longer period of growth to attain a sufficient size before exiting G1 [5] . Here , we resolve this conflict and further investigate the differences between mother and daughter cell cycle control by analyzing the interaction between daughter-specific transcriptional programs , cell size control , and irreversible commitment to the cell cycle at Start .
G1 ( defined operationally as the unbudded period of the cell cycle ) can be decomposed into two independent steps , of duration T1 and T2 , respectively , separated by exit from the nucleus of the transcriptional repressor Whi5 ( Figure 1A ) [5] . We previously used time-lapse fluorescence microscopy of yeast expressing WHI5-GFP and ACT1pr-DsRed [5] to simultaneously measure the duration of T1 , measured by the interval of Whi5 nuclear residence , and cell size , measured using total cell fluorescence expressed from the constitutive ACT1pr-DsRed [5] . T2 , the time between Whi5 nuclear exit and budding , is similar in mothers and daughters and is largely independent of cell size [5] , [6] . T1 is extremely short in mothers but of significant duration in daughters [5] , [6] . G1 size control is readily detected in small daughter cells , and maps specifically to the T1 interval [5] . Smaller cells have a longer T1 , allowing growth to a larger size before cell cycle entry . This links birth size to T1 duration . Given exponential growth of single cells [5] , [29] , the size at Whi5 exit , M1 , is related to the size at birth , Mbirth , through the period T1 by the simple formula: M1 = MbirtheαT1 , where α is the growth rate for exponential growth . This expression yields: αT1 = ln ( M1 ) –ln ( Mbirth ) . The correlation between αT1 and ln ( Mbirth ) characterizes the efficiency of size control . If there is efficient size control , then T1 should become larger as ln ( Mbirth ) becomes smaller , because cells born smaller require a longer period of growth to promote Start . Specifically , the slope of the linear fit of the plot of αT1 against ln ( Mbirth ) should be −1 in the case of perfect size control ( that is , an exact size at which Start is invariably executed ) and 0 in the absence of size control [5] , [30] . The different duration of the period T1 in mothers and daughters could in principle be solely a consequence of size control imposing a delay in the smaller daughter cells [3] . We analyzed the correlation between αT1 and ln ( Mbirth ) , comparing mothers and daughters binned for very similar size at birth ( binning was necessary to ensure sufficient numbers of cells of a given size for statistical comparisons ) . This comparison demonstrates an increase in αT1 in daughters compared to mothers of similar size ( Figure 1B , 1C; regions marked with bars ) ( p values<10−6; Table S4 ) . This delay in Start is most readily detectable in glycerol-ethanol medium ( Figure 1C ) ( p value<10−70; Table S4 ) , in which cell growth is much slower than in glucose medium . Slower growth means that the mother cell feeds less biomass into the daughter cell , resulting in smaller daughter size at the time of cell division [3] . The resulting population of very small daughters enhances detection of size control ( Figure 1C ) [5] . In glycerol-ethanol , across the domain of size overlap in mother-daughter size at birth , daughters exhibit clear size control ( slope ∼−0 . 8 ) while mothers exhibit essentially none ( slope ∼0 ) . This increase in αT1 in daughters with respect to mothers of equal size is consistent with previous findings of a daughter-specific delay , above and beyond the delay needed to achieve equivalent size [19] , [28] . Laabs and collaborators had previously implicated the daughter-specific transcription factor Ace2 in delayed exit from G1 in daughters [28] . Ash1 is a second daughter-specific transcription factor [26] , [27] , and Ace2 contributes to the expression of ASH1 in daughter cells [31] . Ash1 might therefore be the effector of the Ace2-induced daughter delay , or it could independently contribute to daughter delay . We analyzed the correlation between αT1 and ln ( Mbirth ) in ace2 and ash1 single and double mutants ( for a complete list of strains and plasmids used in this study , see Table S1 and Table S2 ) . ace2 ash1 mothers and daughters that were born at similar sizes exhibited similar αT1 values , failing to display the daughter-specific delay seen in wild-type ( Figure 1H , 1I , and Table 1 ) . Furthermore , only very small ace2 ash1 daughters from glycerol/ethanol cultures displayed efficient size control . It is important to note that the mutant still displayed efficient size control by our metric; the effect of the deletions was to shift the size domain where efficient size control could be detected , not to eliminate size control per se . Single mutants ( ace2 ASH1 and ACE2 ash1 ) display a phenotype similar to but less extreme than ace2 ash1 double mutants ( Figure 1D–1G , Table 1 ) . Ace2 contributes to transcriptional activation of ASH1 [31] , so some but not all of the effects of ACE2 deletion may be a consequence of reduced ASH1 expression . The characterized indirect effect of Ace2 on DDE sites 5′ of CLN3 coding sequence [28] likely accounts for at least some of the Ash1-independent effect of ACE2 deletion; we argue below that there is likely an additional direct effect of Ace2 on CLN3 transcription . In strains with ACE2 and/or ASH1 deleted , little effect on mother cell size control is expected or observed , since mother cells naturally lack Ace2 and Ash1 due to differential segregation of the factors at cell division ( see Introduction ) . ace2 ash1 daughters exhibit efficient size control only when born at a size that mothers almost always exceed , due to the budding mode of growth ( Figure 1B , 1C ) [3] . To test whether Ace2 or Ash1 can affect size control when introduced into mothers , we used mutations resulting in symmetrical inheritance of the factors to mothers and daughters . For Ace2 , we used ACE2-G128E ( indicated as ACE2* from here on ) . Ace2-G128E accumulates symmetrically and activates Ace2-dependent transcription in both mothers and daughters [20] , [32] and was shown previously to reduce mother-daughter G1 asymmetry [28] . For Ash1 , we used a mutant ( ASH1* ) in which mutation of localization elements in ASH1 mRNA results in accumulation of Ash1 in both mother and daughter nuclei [33] . As with ace2 ash1 cells , ACE2* ASH1* mothers and daughters that were born at similar sizes exhibited similar αT1 values ( Figure 2G , 2H; Tables 1 , S4 ) . Furthermore , ACE2* ASH1* mothers , when born sufficiently small , exhibit size control , essentially as observed in similarly sized wild-type daughters . Such small mother cells are observed in significant numbers only in glycerol-ethanol culture ( Figure 2H ) . Thus , making the Ace2/Ash1 daughter-specific gene expression program symmetrical between mothers and daughters ( either by deletion or by symmetrical introduction of the factors ) results in effective size control ( high negative slope in αT1 versus ln ( Mbirth ) plots ) over a similar cell size domain in mothers and daughters , eliminating the daughter-specific delay seen in wild-type . In wild-type daughters , size control is exerted at sizes where mothers do not experience size control . Mother cell size control is in principle hard to detect in any case , because these cells “passed” size control in the previous cycle , and budding yeast cell division removes little or no material from the mother cell . For this reason , even in ACE2* ASH1* cells , which presumably all have daughter-type size control , mothers small enough to allow detection of size control are relatively rare . Strains in which only Ash1 or Ace2 is symmetrically localized show intermediate phenotypes ( Figure 2C–2F; Table 1 ) , suggesting again that both transcription factors contribute to delay in T1 in partially independent ways . ACE2* and ASH1* had little effect on size control properties of daughter cells , as expected since these factors are already present in wild-type daughters . Altogether , these results show that Ace2 and Ash1 define daughter-specific programs that shift size control responses to larger cell size . Ace2 and Ash1 appear to be necessary for this shift in size control in daughters compared to mothers; in addition they are sufficient for imposing daughter-like size control properties when introduced in mothers . These results led to the idea that ACE2* ASH1* mothers should be “pseudo-daughters” with respect to size control , while ace2 ash1 daughters should be “pseudo-mothers . ” To test this , we combined data for mothers and pseudo-mothers , and daughters and pseudo-daughters , in rich and poor medium ( Figure 3A–3F ) . We define mothers and pseudo-mothers as “mother-like , ” and daughters and pseudo-daughters as “daughter-like . ” Remarkably , these combined data sets collapsed onto one plot for all mother-like cells and a different plot for all daughter-like cells ( Figure 3E , 3F ) . The individual datasets fit well with the average behavior , as shown by plots separating out the various components ( Figure S7 , S8 ) . The noise about the lines in these plots ( size-independent variation ) is of a magnitude consistent with previous results ( Table S5 ) [5] . Further analysis showed that the daughter-like plot could be transformed to the mother-like plot simply by shifting the curve 0 . 2 units of ln ( Mbirth ) ( Figure 3G , 3H ) . This implies that , with respect to Start , cells containing Ace2 and Ash1 interpret a given cell size as being ∼20% smaller than cells lacking Ace2 and Ash1 . These results can be interpreted in the classical framework of sizers and timers [18] , [34] by defining the point at which cells switch from efficient size control to a timer control ( the intersection between the two lines fitting the correlation between αT1 and ln ( Mbirth ) in Figure 3C ) as “critical size”: a precise size that cells must attain to transit Start . This analogy is imperfect ( the slopes are not −1 or 0 , as required for perfect sizers and timers [5] , [30] , and the sharpness of the transition point cannot be rigorously determined ) but provides a useful simplification using the terms of prior size control literature . Using this terminology , the effect of daughter-specific localization of Ace2 and Ash1 is to cause daughter cells to have a larger “critical size” than mother cells ( increased by 0 . 2 units of ln ( Mbirth ) , or ∼20% larger ) . We emphasize that size control remains highly effective , independent of Ace2 and Ash1; essentially , Ace2/Ash1-containing cells read a given size as smaller than the same size read in the absence of Ace2 and Ash1 . Laabs and collaborators reported symmetrical G1 durations for ace2 mothers and daughters , and for ACE2* mothers and daughters , independent of cell size [28] . In our experiments , the loss of asymmetrical “interpretation” of cell size caused by these mutations does indeed result in T1 durations in mothers and daughters that are more similar than in wild-type ( Figure S4 , S5 ) . However , our results differ in that in our experiments , size control remains present and effective despite deletion or mislocalization of Ace2 and/or Ash1 . As a consequence , the average daughter T1 is still significantly longer than the average mother T1 ( p values<10−3 in glucose; p values<10−14 in glycerol/ethanol ) even in the mutants , since the budding mode of growth ensures that most daughters are born smaller than most mothers . This discrepancy likely has a number of sources . First , our use of T1 , the time from cytokinesis to Whi5 nuclear exit as a landmark , rather than the differential time to budding for mothers versus daughters , as measured by Laabs and collaborators [28] , greatly increases the sensitivity with which size control can be detected , since the interval from Whi5 exit to budding is quite variable , cell-size-independent , and very similar in mothers and daughters [5] . Inclusion of this noisy interval blurs the mother-daughter distinction , which is restricted to T1 . Second , the use of medium supporting slow cell growth ( glycerol-ethanol ) enhances the ability to detect size control , simply because daughters ( of all genotypes ) are born much smaller; the work of Laabs et al . [28] employed only rich glucose medium , making size control harder to detect . Our time resolution is also 3 min per frame rather than 10 . Finally , our cell size estimates are based on the validated ACT1-DsRed marker [5] , while Laabs et al . [28] employed volume estimations from geometry of cell images . We have found that the latter method gives on average similar results to ACT1-DsRed but increases noise in the detection of size control effects [5] . CLN3 was proposed as the relevant indirect transcriptional target of Ace2 to account for mother-daughter asymmetry [28] . Because Ace2 could affect other genes involved in cell size control or mother-daughter asymmetry , and because we had evidence for the involvement of an independent transcription factor , Ash1 , we carried out an unbiased search for the transcriptional target ( s ) through which Ace2 and Ash1 modulate size control in daughters . We performed microarray analysis of synchronized cell populations , comparing cells lacking Ace2 and Ash1 to cells in which they localize symmetrically to both mother and daughter nuclei . Doing the comparisons in this way , rather than simply comparing wild-type to mutants , increases sensitivity of the analysis , since wild-type cultures always contain a mixture of mothers and daughters , reducing the detectable effects of manipulation of daughter-specific transcription factors . Our approach relies on three comparisons: ace2 ash1 versus ACE2* ASH1* , ace2 versus ACE2* , and ash1 versus ASH1* ( see Dataset S2 for the microarrays raw data ) . We also compared swi5 , ace2 , swi5 ace2 , and wild-type in order to obtain insight into the set of genes regulated by one or both of these factors ( see Dataset S1 for the microarrays raw data ) . Swi5 and Ace2 are closely related transcription factors that recognize the same DNA sequence and share many target genes [35] , [36] . The best characterized Ash1 target , HO , is also a Swi5 target and its regulation by Swi5 and Ash1 is required for mother-daughter asymmetry in mating type switching [26] , [27] . To synchronize cells during the critical M/G1 interval , we used strains expressing Cdc20 under the control of an inducible promoter ( the truncated GAL1 promoter , GALL [37] ) . Cells were arrested in metaphase by depletion of Cdc20 in glucose medium and released from the arrest by transfer to galactose medium to reinduce Cdc20 . This synchronization procedure provides excellent synchrony in M/G1 ( anaphase , cell division , and early G1 ) immediately following release , which is the time of nuclear localization of Ace2 , Swi5 , and Ash1 ( Figure 4A ) [36] , [38] . About 15 min after release , cells of all genotypes complete anaphase and degrade the mitotic cyclin Clb2 ( see Figure 4A ) . Subsequently , cells separate and rebud ( Figure 4A ) . Both Swi5 and Ace2 enter the nucleus at about the time of anaphase ( Figure 4A ) . On average , Swi5 nuclear entry precedes Ace2 nuclear entry by 2–3 min ( see Text S1 ) . A slightly longer ( 10 min ) Ace2 delay relative to Swi5 entry was recently reported [39] . Swi5 is rapidly degraded and disappears before cytokinesis and cell separation ( Figure 4A and Text S1 ) [40] . Ace2 is quickly excluded from the mother nucleus but remains in the daughter nucleus for a significant period during G1 ( Figure 4A and Text S1 ) [20] . Ash1 protein begins to accumulate a few minutes after Swi5 and Ace2 nuclear entry and localizes to the nucleus slightly before cytokinesis , remaining until about the time of budding ( Figure 4A and Text S1 ) [26] . The microarrays for wild-type cells show well-defined M/G1 and G1/S clusters consistent with previous results ( Figure 4B ) [38] . Furthermore , well-characterized Ace2 and Ash1 targets , such as CTS1 and HO , behave as expected upon transcription factor deletion or mislocalization ( see Figure 4C ) . Cell-cycle-regulated genes that are unaffected by the two transcription factors behave very similarly in all arrays ( Figure 4C ) . Note that the time of anaphase , which varies slightly between experiments , was used as the zero time to make the comparisons more accurate . The high reproducibility of these microarray data allows us to do a time-point by time-point subtraction of the deletion mutant data from the mislocalization mutant data . This subtraction cancels out cell-cycle-regulated changes in gene expression that are independent of Ace2 and/or Ash1 , allowing the hierarchical clustering algorithm [41] to efficiently detect changes that are specifically due to these transcription factors ( see Figure 4C ) . Clustering analysis of the subtracted data reveals a clear Ace2-dependent cluster composed of well-characterized Ace2-dependent genes , such as CTS1 , DSE1 , and DSE2 ( see Text S1 and Figure S1 for a complete list ) . Only two genes , HO and PST1 , displayed strong changes in expression upon deletion versus mislocalization of Ash1 ( see Text S1 ) . None of the genes whose expression was obviously and strongly Ace2- or Ash1-dependent appeared to be a good candidate to account for daughter-specific regulation of Start . We therefore performed a statistical analysis to obtain a list of genes specifically regulated by both Ace2 and Ash1 . We imposed an “AND” logical condition that co-regulated genes should be detected as differential signals in the subtracted ace2 versus ACE2* , ash1 versus ASH1* , and ace2 ash1 versus ACE2* ASH1* comparisons . Additionally , we imposed a temporal requirement that the observed Ace2/Ash1-dependent changes in expression be observed only at times when these factors have accumulated in wild-type nuclei ( Figure 4A ) . This criterion excludes genes whose changes in expression are long-term , indirect consequences of mutation of Ace2 or Ash1 . Using a p value cutoff sufficient for an expected false positive rate of less than one gene over the whole genome ( see Text S1 and Table S3 ) , we identified only five Ace2/Ash1 shared targets: CLN3 , HSP150 , MET6 , YRF1-1 , and YRF1-5 . Direct interactions between Ace2 or Ash1 and the promoters of three of these genes ( Ace2 with CLN3 and HSP150; Ash1 with YRF1-1 ) were previously observed in chromatin immuno-precipitation ( ChIP ) -chip experiments [42] , [43] , supporting the validity of our analysis ( see Text S1 ) . Prominent in the list of genes affected by both Ace2 and Ash1 is the G1 cyclin , CLN3 , a rate-limiting activator of the Start transition . Laabs and collaborators had likewise implicated CLN3 as a gene repressed by Ace2 , based on comparing CLN3 RNA levels with and without Ace2 , and examining mother versus daughter accumulation of GFP driven from a truncated CLN3 promoter [28] . In that paper , it was also suggested that Ace2 might regulate CLN3 indirectly through an unknown transcription factor that represses CLN3 expression in daughters by binding to DDE sites on the CLN3 promoter . Among all the identified Ace2 targets , Ash1 is the most likely candidate transcription factor for a repressive role on CLN3 expression . We observe , however , that there is no obvious homology between the Ash1 consensus and the DDE . In the next sections we provide evidence that Ash1 binds to the CLN3 promoter and that this binding is at least in part mediated by Ash1 consensus-binding sites that are different from the DDE . Together , these findings suggested the hypothesis that differential regulation of Start in mothers and daughters due to Ace2 and Ash1 may be solely a consequence of differential regulation of CLN3 . CLN3 expression in M/G1 is from 1 . 5- to 2 . 5-fold higher in ace2 ash1 cells ( pseudo-mothers ) than in ACE2* ASH1* cells ( pseudo-daughters ) ( Figure 5A ) , suggesting that CLN3 is differentially regulated in wild-type mothers and daughters . Previously published data support this idea: in populations of cells containing both mothers and daughters , CLN3 expression peaks at the M/G1 boundary [44] , while in populations of size-selected daughters CLN3 expression peaks later in G1 [45] , or shows no peak [12] , [46] , consistent with our conclusion that CLN3 expression in M/G1 is higher in mothers than in daughters . M/G1 expression of CLN3 is driven by Mcm1 through early cell-cycle box ( ECB ) elements [44]; our results and the results of Laabs and collaborators [28] suggest that in daughters , Ace2 and Ash1 antagonize this activation . Hierarchical clustering of microarrays of wild-type and swi5 cells indicates that CLN3 belongs to a cluster of genes whose expression is activated by Swi5 ( Figure 5D ) . Analysis of ace2 versus ACE2* arrays ( Figure 5B ) shows that CLN3 behaves similarly to the rest of this Swi5 dependent cluster upon manipulation of ACE2 ( see Text S1 and Figure S1 for a complete list of Swi5 and Ace2 targets ) . Expression of these genes in ACE2* cells is lower than expression in ace2 at 5 min after anaphase , but higher from 15 min to 25 min ( Figure 5E ) ; that is , the genes appear to be repressed by Ace2 at early times , then activated by Ace2 at later times . This pattern is significantly different from a pattern assuming no regulation by Ace2 ( p<10−11 ) . CLN3 expression depends on Ace2 similarly to these other Swi5 targets ( probability that CLN3 is regulated as the other Swi5/Ace2 targets: p = 0 . 7 , Figure 5E; a model assuming that CLN3 is not affected by Ace2 can be excluded with p<0 . 03 , Figure 5F ) . Thus CLN3 and a class of Swi5 dependent genes follow a pattern consistent with early repression and late activation by Ace2 , and with early activation by Swi5 , likely acting in concert with ECB regulation [44] . We do not know the mechanism underlying this complex pattern . We speculate that Ace2 may be an intrinsically poorer activator than Swi5 , but it activates for a longer period due to its longer nuclear residence . Swi5 disappears from both mother and daughter nuclei a few minutes after anaphase , while Ace2 persists in daughter nuclei for about 20 min longer ( Figure 4A ) . Competition between Ace2 and Swi5 for the same binding site [35] could then contribute to the differential expression observed in these arrays . Alternatively , Ace2 could directly repress expression of these genes; however , no previous evidence suggests a directly repressive role for Ace2 . Microarray analysis for ash1 and ASH1* shows that CLN3 expression is repressed about 2-fold by Ash1 during the period from 10 min to 25 min after anaphase ( Figure 5C ) . During this interval Ash1 is present in the nucleus ( Figure 4A ) , suggesting that it could be a direct repressor of CLN3 expression . Many Swi5 and Ace2/Swi5 targets have moderately higher expression in the absence of Ash1 ( Figure S2 ) . The absolute repression of Swi5-dependent HO expression by Ash1 in daughter cells may thus be an enhancement of a common pattern of co-regulation . Our data suggest that Ace2 and Ash1 cooperate to repress CLN3 expression in daughters . Consistently , activation of the G1/S regulon controlled by Cln3 is delayed and/or happens at larger cell size in cdc20-synchronized cells containing these factors ( Figure S3 ) . We performed chromatin immuno-precipitation ( ChIP ) experiments in synchronized cell populations to ask if Ace2 , Swi5 , and Ash1 bind to the CLN3 promoter . Genome-wide localization data in asynchronous cell populations suggested binding of Ace2 and Swi5 to the CLN3 promoter but were statistically insufficient to definitively prove the association [42] , . We used synchronized cell populations to provide dynamic information on the possible binding of Ace2 , Swi5 , and Ash1 to the CLN3 promoter , providing a higher signal to noise ratio than can be obtained from asynchronous cells . Swi5 and Ace2 bound to regions in the CLN3 promoter around the time of anaphase , coincident with their nuclear entry ( Figure 5G , 5H ) . Swi5 is on the CLN3 promoter for only a few minutes ( Figure 5G ) , while Ace2 is on the CLN3 promoter for about 20 min ( Figure 5H ) , also consistent with the time of Swi5 and Ace2 nuclear localization ( Figure 4A and Text S1 ) . Thus , Ace2 and Swi5 might directly regulate CLN3 transcription by binding to multiple Ace2/Swi5 sites in the CLN3 promoter . A previous meta-analysis of multiple ChIP-chip experiments concluded that Swi5 and Ace2 both bound the CLN3 promoter with high probability ( data in Supp . Table S5 of Ref . [47] ) , consistent with our results . Ash1 binds the CLN3 promoter with kinetics similar to its nuclear localization ( Figure 5I and Figure 4A ) . In contrast , Ash1 residence at the HO promoter is much briefer , consistent with previous results [20] , despite persistence of Ash1 in the nucleus . We do not know the reason for this difference . We noted three candidate Ace2/Swi5 sites ( GCTGGS , consensus sequence: GCTGGT; [42] ) in the CLN3 promoter . The CLN3 promoter also contains two possible variant sites ( GCTGA ) ; such sites are over-represented in Ace2 and Swi5 targets ( B . F . , unpublished data ) . There are eight candidate Ash1-binding sites ( YTGAT ) [48] in the CLN3 promoter . We mutated these Ace2/Swi5 and/or Ash1 putative binding sites in the CLN3 promoter by exact gene replacement ( see Text S1 for details ) . To test if Ace2 , Swi5 , and Ash1 bind to these sites , we performed ChIP analysis in synchronized populations of heterozygous diploid strains containing a wild-type copy and a mutated copy of the CLN3 promoter ( Ace2/Swi5 and Ash1 putative binding sites mutated ) . Following immunoprecipitation , various regions of the CLN3 promoter were amplified by PCR and analyzed by sequencing to obtain an estimate of the ratio of wild-type promoter sequences to mutated sequences ( Figure 6 ) . These experiments are internally controlled ( as they do not require the comparison of two independent ChIP experiments ) . The measured ratio provides an indication of the preferential binding of Ace2 , Swi5 , and Ash1 to the identified putative binding sites . Ace2 , Swi5 , and Ash1 binding to the mutated CLN3 promoter was reduced to about 60% relative to the wild-type promoter , assaying multiple sequences from −1 , 183 to −998 ( ATG: +1 ) ( Table 2 and Table S6 ) . The binding of Ace2 , Swi5 , and Ash1 to sequences from −767 to −545 is not altered by mutation of the putative binding sites ( Table 2 and Table S6 ) . These results indicate that we have identified authentic Ace2 , Swi5 , and Ash1 binding sites in the 5′ region of the CLN3 promoter . The residual binding signal from the mutant is consistent with either a low level of background precipitation , or to genuine residual binding of the factors to non-consensus sites in the promoter . ( Due to uncertainties about such other sites , as well as variable shearing of the DNA in the ChIP procedure , we do not think we can use these data to reliably map which candidate site ( s ) might be directly bound by Ace2 , Swi5 , or Ash1 ) . As a test to see if we might have missed significant binding sites in the CLN3 promoter , we carried out a bioinformatics analysis ( Text S1 and Figure S11 ) looking for regulatory motifs in the promoters of the identified Ace2 and Swi5 targets in S . cerevisiae and three closely related yeasts . Interestingly , we found only two strongly conserved sites: one was one of the candidate Ace2/Swi5 sites we mutated , at position −701 , and the other was a similar but non-consensus site ( GCTTGG ) at position −569 , which we did not mutate since it did not meet the consensus we used in designing the mutagenesis ( see above ) . It is possible , although still untested , that this non-consensus site could account for residual binding of Ace2 to the mutant promoter . The absence of a cluster of Ash1-dependent genes and the low information content of the known Ash1 consensus site ( YTGAT ) does not allow us to perform similar bionformatics analysis; therefore , we cannot test the hypothesis that there are non-consensus Ash1 sites in the CLN3 promoter that we did not mutagenize . To test if the reduced binding of Ace2 , Swi5 , and Ash1 to the CLN3 promoter also has an effect on the regulation of Start , we analyzed the correlation between αT1 and ln ( Mbirth ) in strains carrying mutations of the identified Ace2/Swi5 and/or Ash1 putative sites . These plots show that these mutations reduce the T1 delay in daughters compared to similarly-sized mothers ( Figure 7 , Table 1 ) . The effect is easily detected and statistically significant in cells grown in glycerol-ethanol ( Figure 7 ) , a similar effect was observed in glucose , but this effect did not reach nominal statistical significance ( Figure S9 ) . In the Ace2/Swi5 sites mutant ( Figure 7B , 7D ) the duration of T1 in mothers is prolonged , consistent with the idea that Swi5 is an activator of CLN3 ( since mothers do not contain Ace2 ) . Simultaneous mutation of Ace2 and Ash1 sites did not significantly enhance the phenotype of mutation only of one or the other ( Figure 7 , Table 1 ) . Although these promoter mutations have significant effects , they are less potent than deletion of ACE2 and ASH1 ( compare Figure 1 with Figure 7 , see Table 1 ) . This may be in part due to the presence of additional non-consensus Ace2/Swi5 or Ash1 sites in the CLN3 promoter ( discussed above ) . Additionally , the comparison between mutating Ace2 sites and deleting ACE2 is not exact because removing Ace2 sites perforce also removes Swi5 sites , and on the other hand , deletion of Ace2 alters ASH1 expression . The promoter mutants could also be less effective than deletion of ACE2 and ASH1 because Ace2 has indirect effects on CLN3 expression . It was previously shown that “DDE” sites in the CLN3 promoter play an important role in Ace2-dependent asymmetric control of Start , but these sites did not appear to be bound by Ace2 , suggesting an indirect mechanism [28] . Interestingly , these sites are transcribed into the CLN3 mRNA , and the Whi3 RNA binding protein binds to a repeated RNA sequence at the center of the DDE [49] . Whi3 is a regulator of cell size control thought to work by regulation of CLN3 mRNA and protein [49] , [50] , [51] , [52] , [53] . At present , though , there is no information implicating Whi3 in mother-daughter asymmetry , nor is Ace2 known to regulate Whi3 . Finally , Ace2/Ash1 could regulate additional G1-regulatory genes at a level not detectable by our statistical analysis ( see above ) . The observation that reduced binding of Ace2 , Swi5 , and Ash1 to the CLN3 promoter results in a significant reduction of asymmetric control of Start by cell size in mothers and daughters supports the idea that Ace2 and Ash1 directly repress CLN3 expression in M/G1 , accounting for a significant part of the regulation of G1 length by these transcription factors . We suggest that direct regulation of CLN3 by Ace2 and Ash1 together with its indirect regulation by Ace2 through the DDE sites [28] can explain asymmetric control of Start by cell size in mothers and daughters . CLN3 expression in M/G1 is ∼2-fold higher in ash1 ace2 cells ( pseudo-mothers ) than in ASH1* ACE2* cells ( pseudo-daughters ) ( Figure 5A ) . While this change is small , CLN3 is a highly dosage-sensitive activator of Start . Previous measurements of cell sizes in cycling cell populations demonstrated effects on cell size upon 2-fold changes up or down in CLN3 gene dosage [10] , [44] . To increase the precision of this analysis , we analyzed the correlation between αT1 and ln ( Mbirth ) in cells carrying either two or six copies of CLN3 . This focuses the analysis on the critical interval , since Cln3 decreases cell size specifically by decreasing T1 in wild-type mothers and daughters [5] . Daughter cells with two copies of CLN3 exhibit efficient size control [high negative slope in αT1 versus ln ( Mbirth ) ] over a size range that is shifted to smaller cell size by ∼0 . 15 units of ln ( Mbirth ) , compared to wild-type; this shift is similar to that in ace2 ash1 daughter cells ( mother-like cells ) ( compare Figure 5J to Figure 3C ) . Thus , the observed ∼2-fold changes in CLN3 expression upon deletion versus mislocalization of Ace2 and Ash1 could account for the observed changes in cell size control in these mutants . Six copies of CLN3 almost eliminate size control even in very small daughter cells ( Figure 5J ) . Thus , size control is remarkably sensitive to CLN3 gene dosage; it can only be modulated by altering CLN3 expression in a narrow range before size control is lost . We analyzed the correlation between αT1 and ln ( Mbirth ) in cln3 cells and in cln3 cells expressing the CLN3 ORF ( without the upstream DDE sites ) from constitutive promoters . It is important for this analysis that the constitutive promoters provide expression levels of Cln3 similar to those in wild-type cells and that the promoter-CLN3 fusions complement the large-cell phenotype of cln3 mutants , without “overshoot” to a small-cell phenotype [8] , [10] , [54] . We screened a number of different constitutive promoters of different strengths [55] for these properties , examining both cell size and Cln3 protein levels using myc-tagged Cln3 , compared to wild-type ( including an approximate correction for cell cycle regulation of CLN3 expression from the endogenous promoter ) ( Table 3; [44] ) . The ACT1 and the ADH1 promoters result in over-expression of Cln3 and in a small cell-size phenotype for cells grown in glucose-containing media ( Table 3 ) . Expression of Cln3 from the CDC28 promoter is weaker than expression from the CLN3 promoter and results in cell sizes bigger than wild-type and only slightly smaller than cln3 cells ( Table 3 ) . Integration into the yeast genome of six copies of the CDC28pr-CLN3 construct results in a cell size distribution similar to that of wild-type cells . We also analyzed the effects of these constructs in glycerol-ethanol medium . Four tandemly integrated copies of CDC28pr-CLN3 results in an overall cell size distribution similar to that of wild-type cells in glycerol-ethanol ( Table 3 ) . As a result of decreased ADH1 expression in non-fermentable media [56] , the ADH1 promoter provides Cln3 levels similar to endogenous levels in glycerol-ethanol , resulting in a cell size distribution slightly ( ∼10% ) larger than wild-type ( Table 3 ) . Measurements of Cln3 protein levels show that Cln3 overexpressors were smaller than wild-type , and underexpressors larger ( Table 3 ) ; therefore , the measurements of Cln3 level were accurate over a physiologically relevant range . Based on results with a single copy of CDC28pr-CLN3-myc , four to six copies of CDC28pr-CLN3 should produce approximately wild-type levels of Cln3 in M/G1 , consistent with the observed cell size distributions ( Table 3 and Figure 8 ) . We therefore used strains containing 6xCDC28pr-CLN3 in glucose medium , and strains containing 4xCDC28pr-CLN3 or ADH1pr-CLN3 in glycerol-ethanol medium , to provide approximately endogenous levels of expression without mother-daughter asymmetry ( and presumably without regulation by the cell cycle , Ace2 , or Ash1; note that the 5′ DDE sites are not present in these constructs ) . In 6xCDC28pr-CLN3 cells the daughter-specific delay is almost entirely abolished ( Figure 8C , 8E and Table 1 ) . Similarly , in 4xCDC28pr-CLN3 and ADH1pr-CLN3 cells grown in glycerol/ethanol , the daughter-specific delay is almost entirely abolished , and small mothers and daughters have similar size control properties ( Figure 8D , 8F , and 8G and Table 1 ) . Thus , similarly to the results obtained by placing Ace2 and Ash1 in mother nuclei , size control in small mother cells can be detected by eliminating differential mother-daughter control of CLN3 expression . Small 4xCDC28pr-CLN3 and ADH1pr-CLN3 cells in glycerol/ethanol exhibit strong size control ( slopes of ∼−0 . 8 , compared to a theoretical expectation for perfect size control of −1 [5] , [30] ) ( Figure 8F , 8G ) , suggesting that while daughter-specific transcriptional regulation of CLN3 specifies the cell size domain over which size control is effective , the intrinsic mechanism of size control is not dependent on mother-daughter regulation of CLN3 transcription , or indeed on any transcriptional regulation acting through the CLN3 promoter . We speculate that an M/G1 burst of CLN3 expression from Mcm1 and/or Swi5 ( [44]; Figure 5 ) may be sufficient to drive cells rapidly through T1 , as is observed in wild-type mothers of all sizes ( Figure 1B , 1C; [5] ) ; in daughters , this burst may be suppressed by Ace2 and Ash1 . The daughter-specific delay of wild-type cells depends on CLN3 , since cln3 mother and daughter cells of similar size have similar αT1 . Remarkably , cells deleted for cln3 still exhibit strong effects of cell size on G1 duration , although these effects are symmetrical between mothers and daughters of similar size ( Figure 8C , 8D ) . Thus , while the cell size domain of effective size control in wild-type cells is set by CLN3 , there may be an underlying Cln3-independent or parallel program of cell size control that acts or becomes detectable only upon deletion of CLN3 [57] . In addition to loss of mother-daughter asymmetry , the response of cln3 cells to cell size is shifted to about 1 . 5-fold larger cell sizes as measured using ACT1pr-DsRed; this finding confirms that cln3 cells are larger in terms of protein content than wild-type , in contrast to the proposal that the increase in cell size of cln3 cells [8] , [10] is due primarily or entirely to increased vacuole size [58] . Our results are consistent with those of Laabs and collaborators , who reported that cln3 cells and cells expressing CLN3 from ectopic promoters lost mother-daughter asymmetry [28] . They also observed equal G1 durations for individual mother/daughter pairs [28] . In our analysis , in contrast , in almost all cln3 mother-daughter pairs , with or without ectopic expression of CLN3 , the daughters had a longer T1 period ( see Figure S6; p values<10−5 in glucose; p values<10−15 in glycerol/ethanol ) , although the daughter delay was reduced compared to wild-type , consistent with the results of Laabs and collaborators [28] . The symmetry that we observe in these mutants is restricted to mothers and daughters of similar size ( more precisely , in the mother and daughter plots of αT1 versus ln ( Mbirth ) , in regions where the domains of mothers and daughters overlap ) . We assume this discrepancy arises from the same reasons discussed above .
It was previously reported that asymmetric localization of Ace2 represses CLN3 in daughter cells [28] . Our results extend this finding by showing that Ace2 regulation of CLN3 is in part direct , mediated by Ace2 binding to the CLN3 promoter . In addition , our results implicate Swi5 and Ash1 as well as Ace2 in CLN3 regulation . Neither asymmetric expression of CLN3 , nor CLN3 itself , is essential for size control [57] ( Figure 8 ) . However , CLN3 sets the domain of cell sizes over which effective size control operates in wild-type cells . For this reason , negative control of CLN3 by Ace2 and Ash1 allows differential Start regulation in mothers and daughters . These findings provide empirical validation for one part of the theoretical cycle of transcriptional regulators proposed to account for a B-type cyclin-independent autonomous transcriptional oscillator [47] . In the budding mode of growth , cell mass produced after budding goes to the daughter , but all pre-budding mass is retained by the mother [3] . As a result , a daughter that “passes” size control will retain this size through all subsequent ( mother ) budding cycles . This cell could thus be accelerated through Start by the M/G1 CLN3 burst without a “need” for size checking . Given the high amount of noise in the mechanism of size control [5] , this could prevent unnecessary delays in already full-sized mothers . The M/G1 CLN3 burst , if experienced by daughters , would perturb the ability of daughters to effectively check their size ( Figure 5J ) . This could result in the requirement for daughter-specific blockage to the burst . Thus , mother/daughter-specific CLN3 regulation could simultaneously prevent unnecessary mother delays and prevent smaller daughters from passing Start prematurely . In addition to repressing initial expression of CLN3 in M/G1 , Ace2 also induces ASH1 expression; Ash1 represses later expression of CLN3 . This is an example of “feed-forward” regulation , which may be a common regulatory structure for providing delayed response [59]—in this case , prolonged CLN3 repression even after loss of Ace2 from the daughter nucleus . We speculate that this mechanism may allow daughter-specific delay over a broad range of timescales and growth rates . The CLN3 upstream region is unusually large ( 1 . 2 kb , compared to an average intergenic distance of 0 . 6 kb ) and contains six ECB sites [45] , multiple Ace2/Swi5 and Ash1 sites ( this work ) , and four DDE sites [28] . How all of these sites and the factors that bind to them cooperate combinatorially to properly regulate CLN3 is unknown . This is a large amount of regulatory machinery to provide a maximum peak-to-trough ratio only on the order of three [44]; however , since manipulation of CLN3 gene copy number up or down by only a factor of two yields significant cell size phenotypes ( Figure 5J ) [10] , [44] , this level of control is likely to be physiologically significant , perhaps for the reasons cited above . As noted above , our results can be interpreted in the classical framework of sizers and timers [18] , [34] by defining the point at which cells switch from efficient size control to a timer control as “critical size” ( Figure 3C ) : a precise size that cells must attain to transit Start . This point is marked by the intersection of the line of high negative slope with the line of low or zero negative slope in αT1 versus ln ( Mbirth ) plots . Using this framework , we can summarize our results by stating that Ace2/Ash1-containing cells have a larger “critical size” than cells lacking these factors ( normally , daughters and mothers , respectively ) . This formulation is inexact , primarily due to the evident non-zero slope in our data for the second component of the two-slope fit; remarkably , though , the effects of Ace2 and Ash1 shift not just the intersection point but the entire curve by 0 . 2 units of ln ( Mbirth ) . While increasing cell size and increasing Cln3 both decrease T1 ( i . e . , accelerate Whi5 exit from the nucleus after cytokinesis ) [5] , Ace2 exit from the daughter nucleus occurs about 15 min ( 15±6 min ) after cytokinesis , independent of cell size and CLN3 ( Text S1 and Figure S10 ) . Thus , overall , Start control may consist of three distinct modules: Ace2 and Ash1-dependent but cell-size independent setting of the domain of cell size control; size control itself , leading to initiation of Whi5 nuclear exit; and a final size-independent step driven by CLN1 , 2-dependent transcriptional positive feedback , which rapidly completes Whi5 exit and drives the downstream events of Start [5] , [7] . In wild-type homothallic budding yeast , only mother cells express the HO endonuclease and switch mating type , due to Ash1 repression of HO expression in daughters [26] , [27] . Phylogenetic analysis shows that in fungi , ASH1 appeared before HO . This suggests that Ash1 may have functions predating HO , which may be important for asymmetrical cell division . It would be interesting to test whether Ash1 functions in cell cycle control in other fungi that can divide asymmetrically , such as Candida albicans , which lacks a HO homolog but expresses an Ash1 homolog that localizes specifically to the daughter cells [60] , [61] . Ash1 also is found in A . gossypii , which undergoes asynchronous division in a multinucleate syncitium [62]; it would be interesting to evaluate the role of Ash1 in this asynchrony . Ace2 controls genes that confer diverse aspects of daughter cell biology [20] , [23] , [24]; here we elucidate how Ace2 also contributes to differential Start regulation in daughters [28] . There are interesting parallels and connections between HO control and CLN3 control . Both are activated by Swi5 and inhibited by Ash1 . Swi5 regulation of HO in mothers can be interpreted as feed-forward control , since Swi5 directly primes HO for expression [63] and also activates CLN3 expression , which later yields efficient activation of the SBF factor that drives HO transcription [7] , [63] . Cell cycle regulation and cell differentiation , often driven by asymmetric localization of cell fate determinants during cell division [64] , [65] , [66] , are inter-regulated in many systems [67] , [68] , [69] . As the decision of cells to differentiate is often made in G1 , cell differentiation and commitment to a stable G1 are often coregulated [67] , [69] , [70] . It would be interesting to examine cases in which stem cells produce one proliferating cell and one daughter that differentiates in G1 [65]; such cells might employ mechanisms similar to those we have uncovered in differential mother-daughter G1 control in budding yeast .
Standard methods were used throughout . All strains are W303-congenic . All integrated constructs were characterized by qPCR . Mutations of the Ace2/Swi5 and Ash1 binding sites on the CLN3 promoter were verified by sequencing . Preparation of cells for microscopy and time-lapse microscopy were performed as previously described [5] , [6] . Growth of microcolonies was observed with fluorescence time-lapse microscopy at 30°C using a Leica DMIRE2 inverted microscope with a Ludl motorized XY stage . Images were acquired every 3 min for cells grown in glucose and every 6 min for cells grown in glycerol/ethanol with a Hamamatsu Orca-ER camera . Custom Visual Basic software integrated with ImagePro Plus was used to automate image acquisition and microscope control . Automated image segmentation and fluorescence quantification of yeast grown under time-lapse conditions and semi-automated assignment of microcolony pedigrees were performed as previously described [6] . The nuclear residence of Whi5-GFP was scored by visual inspection of composite phase contrast-fluorescent movies . Cell size was measured as the total cell fluorescence from DsRed protein , expressed from the constitutively active ACT1pr , as previously described [5] . Cell size at every time point was extrapolated from a linear fit of the ln ( M ) as a function of time for cells grown in glucose and from a smoothing spline fit for cells grown in glycerol/ethanol . Individual cell growth in glycerol/ethanol appears to be intermediate between a linear and an exponential model ( unpublished data ) ; this deviation from exponentiality has very little effect on this analysis . Time-lapse fluorescence microscopy , microarray data , and sequencing data were analyzed with custom software written in MATLAB software ( see Text S1 for details on the analysis of the microarray data ) [5] . For cluster analysis , the log2 of the arrays data or of the subtracted arrays data were hierarchically clustered by the agglomerative algorithm [41] . Data were visually presented using JavaTreeView . For sequencing data , the area associated to every wild-type or mutated nucleotide was evaluated manually by using the MATLAB software . YEP medium was used for all cell cycle synchronization experiments , supplemented with the appropriate carbon source as indicated below . Cell cycle synchronization by the cdc20 GALL-CDC20 block release was achieved by growing cells to early log phase in YEP+galactose ( 3% ) , then filtering and growing them in YEP+glucose ( 2% ) for 3 h to arrest cells in metaphase . Cells were released from the block by filtering back into YEP+galactose ( 3% ) . GALL is a truncated version of the GAL1 promoter that shows inducible but significantly lower expression than the full-length GAL1 promoter [37] . Microarrays were performed as previously described [71] but using microarrays carrying PCR fragments from open reading frames of S . cerevisiae . Each array had each PCR fragment independently spotted four to eight times , leading to a high redundancy of data and small errors in expression ratios . RNA extraction , cDNA synthesis and labeling , and hybridization and scanning were carried out by the Stony Brook spotted microarray facility , as described previously [71] . Standard methods were used for ChIP experiments . Early log phase cells were fixed for 15 min in 1% formaldehyde at room temperature . Immunoprecipitations were performed with IgG Sepharose beads . Immunoprecipitated DNA was amplified by PCR .
|
Asymmetric cell division is a universal mechanism for generating differentiated cells . The progeny of such divisions can often display differential cell cycle regulation . This study addresses how differential regulation of gene expression in the progeny of a single division can alter cell cycle control . In budding yeast , asymmetric cell division yields a bigger ‘mother’ cell and a smaller ‘daughter’ cell . Regulation of gene expression is also asymmetric because two transcription factors , Ace2 and Ash1 , are specifically localized to the daughter . Cell size has long been proposed as important for the regulation of the cell cycle in yeast . Our work shows that Ace2 and Ash1 regulate size control in daughter cells: daughters ‘interpret’ their size as smaller , making size control more stringent and delaying cell cycle commitment relative to mother cells of the same size . This asymmetric interpretation of cell size is associated with differential regulation of the G1 cyclin CLN3 by Ace2 and Ash1 , at least in part via direct binding of these factors to the CLN3 promoter . CLN3 is the most upstream regulator of Start , the initiation point of the yeast cell cycle , and differential regulation of CLN3 accounts for most or all asymmetric regulation of Start in budding yeast mother and daughter cells .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology/cell",
"differentiation",
"cell",
"biology/cell",
"growth",
"and",
"division"
] |
2009
|
Daughter-Specific Transcription Factors Regulate Cell Size Control in Budding Yeast
|
The cellular endosomal sorting complex required for transport ( ESCRT ) machinery is involved in membrane budding processes , such as multivesicular biogenesis and cytokinesis . In HIV-infected cells , HIV-1 hijacks the ESCRT machinery to drive HIV release . Early in the HIV-1 assembly process , the ESCRT-I protein Tsg101 and the ESCRT-related protein ALIX are recruited to the assembly site . Further downstream , components such as the ESCRT-III proteins CHMP4 and CHMP2 form transient membrane associated lattices , which are involved in virus-host membrane fission . Although various geometries of ESCRT-III assemblies could be observed , the actual membrane constriction and fission mechanism is not fully understood . Fission might be driven from inside the HIV-1 budding neck by narrowing the membranes from the outside by larger lattices surrounding the neck , or from within the bud . Here , we use super-resolution fluorescence microscopy to elucidate the size and structure of the ESCRT components Tsg101 , ALIX , CHMP4B and CHMP2A during HIV-1 budding below the diffraction limit . To avoid the deleterious effects of using fusion proteins attached to ESCRT components , we performed measurements on the endogenous protein or , in the case of CHMP4B , constructs modified with the small HA tag . Due to the transient nature of the ESCRT interactions , the fraction of HIV-1 assembly sites with colocalizing ESCRT complexes was low ( 1 . 5%-3 . 4% ) . All colocalizing ESCRT clusters exhibited closed , circular structures with an average size ( full-width at half-maximum ) between 45 and 60 nm or a diameter ( determined using a Ripley’s L-function analysis ) of roughly 60 to 100 nm . The size distributions for colocalizing clusters were narrower than for non-colocalizing clusters , and significantly smaller than the HIV-1 bud . Hence , our results support a membrane scission process driven by ESCRT protein assemblies inside a confined structure , such as the bud neck , rather than by large lattices around the neck or in the bud lumen . In the case of ALIX , a cloud of individual molecules surrounding the central clusters was often observed , which we attribute to ALIX molecules incorporated into the nascent HIV-1 Gag shell . Experiments performed using YFP-tagged Tsg101 led to an over 10-fold increase in ESCRT structures colocalizing with HIV-1 budding sites indicating an influence of the fusion protein tag on the function of the ESCRT protein .
The budding of HIV-1 at the plasma membrane of a virus-producing cell relies on recruitment of , and interaction with , various host cell factors . Initial viral bud formation is primarily induced by assembly of plasma membrane associated Gag molecules into a hexagonal lattice that attains an outward curvature through the introduction of irregular lattice defects ( reviewed in [1–3] ) . However , for the final membrane remodeling steps leading to fission , HIV-1 relies on the cellular endosomal sorting complex required for transport ( ESCRT ) ( reviewed in [4–6] ) that is mechanistically involved in various cellular membrane bending and separation processes , including the formation of multivesicular bodies ( MVB ) or cytokinesis ( reviewed in [7–9] ) . ESCRT consists of four different sub-complexes ( ESCRT-0 to ESCRT-III ) and associated factors such as VPS4 and the ALG-2 interacting protein X ( ALIX ) ( reviewed in [9 , 10] ) , HIV-1 employs ESCRT-I , as well as components of ESCRT-III and the AAA ATPase Vps4 in the viral budding process [11–16] . Alternatively to ESCRT-I , ALIX may also serve to recruit ESCRT-III to HIV-1 budding sites [11 , 17 , 18] . The ESCRT machinery is recruited via the C-terminal p6-domain of HIV-1 Gag . This small protein comprises two so-called late domain ( L-domain ) motifs that bind to the central ESCRT-I component tumor susceptibility gene 101 ( Tsg101 ) [16 , 19–21] or to ALIX [11 , 17 , 22] , respectively . Both ESCRT-I and ALIX can serve to recruit ESCRT-III to the viral budding site . The Tsg101 interacting PT/SAP motif has been identified as the major determinant of HIV-1 release; its mutation affects particle release and leads to the arrest of late budding structures at the plasma membrane [16 , 23 , 24] . In contrast , the ALIX-interacting LYPXL motif apparently serves a partly redundant , auxiliary function in the case of HIV-1 [18] . Although mutation of the ALIX binding site has less pronounced effects on HIV-1 release compared to mutation of PT/SAP in model cell lines [25 , 26] , overexpression of ALIX fragments can exert a dominant-negative effect on virus budding [17 , 27] . Furthermore , the release of HIV-1 mutants unable to interact with Tsg101 can be rescued by overexpression of ALIX [25 , 28] . Thus , ESCRT-III recruitment in most cell systems studied is predominantly mediated by ESCRT-I via an ESCRT-II independent pathway , but an alternative route involving direct interaction of ALIX with the ESCRT-III component CHMP4 [17 , 25 , 29] is possible ( Fig . 1A ) . Besides their function in recruiting ESCRT-III , both ESCRT-I and ALIX may play a role in membrane bending or remodeling [30] . The formation and re-organization of membrane bound ESCRT-III multimers is presumed to be the central driving force for membrane remodeling during the fission process . In the past years , various in vivo and in vitro studies of cytokinesis [8 , 31] , as well as HIV-1 release [5 , 6 , 32] and multivesicular budding [7] , have led to different models for the mechanism of membrane constriction and fission ( Fig . 1B ) . One of the early proposed models was from Hanson et al . who observed circular and spiral-shaped structures of endogenous ESCRT-III protein membrane assemblies on the cytoplasmic side of the plasma membrane using deep-etch electron microscopy [33] . This and other studies [34 , 35] suggested models in which spiraling or constricting ESCRT-III filaments slide past each other resulting in membrane constriction ( Fig . 1B , model 1 ) . A second family of models presume that ESCRT-III filaments assemble within the budding neck [7] and , together with other ESCRT components , form dome- [36 , 37] , whorl- [38] , circular or spiral-shaped structures [31 , 34 , 39–43] . In these models , constriction of the plasma membrane occurs from within the neck ( Fig . 1B , Model 2 ) . Remodeling of multimeric CHMP4 complexes could eventually lead to narrowing of the bud neck from within the fission neck and spontaneously induce membrane fission when the neck diameter reaches a lower threshold [36 , 44]; alternatively , the ESCRT associated ATPase VPS4 has been proposed to be actively involved in membrane remodeling [13 , 45] . Two other ESCRT-III proteins , CHMP2 and CHMP3 , both of which stabilize the neck from within by forming a dome-structure , may also play a supportive role [36 , 37 , 39 , 42] . Electron micrographs of nascent HIV-1 particles arrested at a late stage of budding by depletion of CHMP2 showed ring-like striations situated within the bud neck[14] , favoring an internal constriction model . Further support for models of type 2 is provided by in vitro experiments using giant unilamellar vesicles ( GUVs ) and purified ESCRT-III proteins . These studies indicated a direct involvement of other ESCRT-III components in the initial membrane neck formation during HIV-1 assembly [34 , 46] . Studies from Lata et al . [37] showed the formation of large tubular structures by CHMP2-CHMP3 heteromeric complexes suggesting that CHMP2-CHMP3 clusters form a dome-like assembly inside HIV-1 budding neck , which is supported by computational thermodynamic calculations [36] . Recently , Hanson and coworkers used deep-etch electron microscopy to analyze membrane associated ESCRT-III assemblies arrested by VPS4 depletion , reporting spiral shaped arrangements with an average diameter of ~110 nm [40] . Nascent HIV-1 buds in Vps4 depleted cells were found to be encircled by ESCRT-III filaments . These findings suggested that Vps4 may be required for the transition of a ring-like structure encircling the bud neck to a growing spiral that constricts the bud neck from within . Finally , van Engelenburg et al . [47] recently proposed a new model predicting membrane constriction to be mediated from within the virus bud ( Fig . 1B , Model 3 ) . Using two-color 3D-iPALM super-resolution microscopy on eGFP protein fusions of Tsg101 and the ESCRT-III proteins CHMP2A and CHMP4B , they found that all factors localized in the center of a nascent Gag bud . This result contrasted all previously described models that suggested ESCRT-III proteins driving membrane fission from around or below the nascent virus-like particles ( VLPs ) , as well as biochemical analyses , which did not detect ESCRT-III proteins in purified HIV-1 particles . It thus remains to be clarified , whether the surprising findings might be related to altered localization or multimer arrangements of the GFP-tagged ESCRT components . To resolve this issue and investigate the mechanism of membrane scission under more physiological conditions , we measured the size and localization of endogenous ESCRT protein assemblies at assemblies of HIV-1 Gag expressed in the complete viral context [48] . To avoid potential alterations in ESCRT localization and binding caused by fluorescent protein ( FP ) tags , we used antibodies for immunodetection of endogenous ESCRT proteins or employed a small epitope tag where suitable antibodies were not available . Super-resolution fluorescence microscopy was employed to analyze complexes smaller than the diffraction limit of optical microscopy . This approach allowed us to monitor hundreds of μm2 in a single image and thereby obtain statistically relevant information on the comparatively low number of interactions observed with endogenous ESCRT proteins in the absence of an experimentally induced budding arrest . We investigated the early-acting factors Tsg101 and ALIX as well as the late-acting ESCRT factors CHMP4B and CHMP2A known to assemble into high-molecular structures at the membrane ( Fig . 1A ) . Super-resolution images of each of these ESCRT and ESCRT-related protein assemblies colocalizing with single HIV-1 budding sites showed a condensed circular spot with a full-width at half-maximum ( FWHM ) between ~ 45 nm – 60 nm or a diameter ( determined using a Ripley L-function analysis [49] ) of approximately ~ 60 – 100 nm . Our results are consistent with an internal membrane constriction mechanism driven by protein assemblies smaller than HIV-1 budding structures and more similar to the size of the budding neck . In addition , when performing experiments in cells overexpressing fluorescent protein fusions of the ESCRT protein Tsg101 , an increase of over 10-fold in the colocalization of ESCRT complexes with HIV-1 assembly sites was observed indicating that the function and potentially also the structure of the ESCRT complexes are altered by the large fusion protein label .
To put the super-resolution images of the various ESCRT factors in the context of the HIV assembly sites , we first measured the size of assembly sites at the plasma membrane with Photoactivation Localization Microscopy ( PALM ) in TIRF mode using mEos-labeled Gag expressed in the viral context ( See Supporting Information ) . The average size of the assembly sites was found to be 116 ± 36 nm full-width at half-maximum ( FWHM ) or a diameter of 141 ± 41 nm from the Ripley-L analysis ( See Supporting Results and S1 and S2 Figs . for details ) . This is similar to the size of released HIV-1 particles determined by cryo-electron microscopy [50 , 51] and previous super-resolution studies [52–54] . To verify that fixation and permeabilization of the HeLa cells does not lead to artifacts , we also performed live-cell PALM experiments with the mEos-labeled Gag constructs expressed in the viral context and obtained similar results ( a FWHM of 108 ± 35 nm or a diameter of 152 ± 37 nm from the Ripley-L analysis , S2 Fig . ) . As the relationship between results of the Ripley analysis and actual cluster size is very sensitive to labeling and cluster densities [55] , we chose to give the cluster sizes as the FWHM of our super-resolution analysis , which is slightly smaller than the actual diameter . The accuracy of measured super-resolution images depends on the labeling efficiency and effectivity of the antibody staining . To determine the efficiency of labeling protein structures within the bud , we performed experiments with an eGFP tagged version of the viral accessory protein Vpr . Vpr is recruited to the HIV-1 assembly sites by interacting with the C-terminal p6 domain of the polyprotein precursor Pr55Gag inside the bud [56–58] . eGFP . Vpr was co-expressed with HIV-1 Gag alone to avoid expression of endogenous Vpr; for this , we made use of a previously described Rev-independent Gag expression construct [59] , carrying an FP tag at the C-terminus . HeLa cells were transiently co-transfected with an equimolar ratio of plasmids encoding HIV-1 Gag and a mCherry-tagged derivative , respectively , together with peGFP . Vpr and fixed at 18–20 hpt . For super-resolution imaging , eGFP . Vpr was marked using a primary polyclonal anti-GFP antibody and fluorescently-labeled secondary-antibody . TIRF images of cells expressing both pEGFP . Vpr ( S3A Fig . , left panel ) , Gag:Gag . mCherry ( 1:1 ) ( S3A Fig . , middle panel ) and anti-GFP immunostaining ( S3A Fig , right panel ) revealed a high number of distinct eGFP . Vpr protein assemblies strongly colocalizing with HIVmCherry budding sites ( S3B–S3C Fig . ) as would be expected for functional recruitment of Vpr during Gag assembly . The primary/secondary antibody complexes labeled the majority of HIV-1 assembly sites displaying a detectable eGFP . Vpr signal , confirming that a significant fraction of proteins present within the nascent bud is accessible to immunostaining . Fitting a Gaussian function to the cross-section profile of 49 colocalizing spot-like super-resolution reconstructions of immunodetected Vpr protein assemblies yielded an average FWHM of 56 ± 12 nm ( S3D Fig . ) , smaller than the VLP diameter determined by PALM imaging of Gag . mEos , consistent with a central localization of eGFP . Vpr within the bud . A Ripley analysis of our super resolution data revealed similar results to the experiments performed by Lehman et al [53] with a diameter of 95 ± 53 nm . Having determined the diameter of HIV-1 budding sites in the context of our experiment , we characterized the size and structure of membrane associated endogenous Tsg101 protein assemblies acting as recruitment factors during the early stages of HIV-1 assembly ( Fig . 1A ) using Stochastical Optical Reconstruction Microscopy ( STORM ) . HeLa cells were transiently transfected with the previously characterized full viral construct for HIVmCherry [48] at a ratio of 1:1 with unlabeled HIV construct to mark the position of HIV-1 budding sites . Endogenous Tsg101 proteins were directly visualized 14–15 h post transfection by adding primary monoclonal anti-Tsg101 antibodies and appropriate secondary antibodies that were labeled with the activator-reporter dye pair Alexa Fluor 488-Cy5 . TIRF microscopy images of fixed cells revealed distinct HIV-1 budding sites and Tsg101 protein assemblies at the membrane ( Fig . 2A-2B ) that were classified into colocalizing and non-colocalizing structures . Large fluorescent Gag assemblies whose appearance did not correspond to individual budding sites were excluded from the analysis . In 14 cells analyzed , 18 Tsg101 clusters were found to colocalize with HIV-1 assembly sites , corresponding to a colocalization percentage of 1 . 8% with respect to all detected assembly sites . To verify that the low number of colocalizations is not due to inefficient labeling of the epitope , we performed experiments with YFP-tagged Tsg101 ( discussed in detail below ) . Over 60% of the Tsg101 complexes observable via the YFP signal were also detected using an anti-Tsg101/secondary antibody construct . Hence , the low number of colocalizations can be attributed to the transient , dynamic nature of the HIV-1 budding process and ESCRT recruitment [13 , 60] . Reconstructions based on 2D Gaussian localization revealed that all colocalizing Tsg101 structures appeared as small , condensed , circular spots ( Fig . 2B ) . The size ( FWHM ) of the 18 colocalizing spot-like Tsg101 clusters was between 40 and 75 nm ( Fig . 2C ) with a mean value of 58 ± 7 nm . This is significantly smaller than the previously determined mean HIV-1 bud size of 116 ± 36 nm ( Reference ) . Untransfected HeLa cells stained with the same set of antibodies showed condensed , roundish Tsg101 protein assemblies at the cell membrane ( S4A Fig . ) resembling those observed in HIV-1 transfected cells ( Fig . 2 ) . This is in agreement with Welsch et al . who used quantitative immuno-EM to show that ~15% of all Tsg101 proteins were present at the plasma membrane of untransfected immune cells and that the distribution of Tsg101 between plasma and intracellular membranes did not significantly change upon HIV-1 infection [61] . Our findings confirm the capability of Tsg101 to assemble at the plasma membrane due to processes unrelated to HIV-1 budding . It also highlights the necessity to verify the colocalization of ESCRT structures with HIV assembly sites and thus to exclude structures in the analysis caused by other cellular events . We also performed experiments with an HIVmCherry late- variant of HIVmCherry comprising a mutation of the PT/SAP motif . Consistent with what is expected , we did not observe a single structure that colocalized with the more than 210 HIV-1 assembly sites detected . If the neck of the budding structure was limiting the size of the Tsg101 containing protein assemblies at budding sites , one might expect a more narrow size distribution than for non-colocalizing structures . Indeed , whereas the average diameter of non-colocalizing clusters was similar ( 60 ± 15 nm ) , the STORM analysis revealed a broader size distribution than for the colocalizing Tsg101 clusters , ranging from 35 to 135 nm ( Fig . 2D ) . Likewise , a broader size distribution was also observed for 83 non-colocalizing Tsg101 clusters detected at the membrane of recruitment defective HIVmCherry ( late- ) expressing cells ( Fig . 2E and S4C Fig . ) . This observation suggests that the dimensions of the Tsg101 complexes specifically recruited to HIV-1 budding sites may be confined . We also investigated the ESCRT-related protein ALIX , another early-acting factor in HIV-1 budding ( Fig . 1A ) , because of its key function in the ESCRT-recruitment process and its potential role in stabilizing the CHMP4-scaffold . HeLa cells were transiently transfected with pCHIV:pCHIVmCherry and endogenous ALIX was directly stained 14–15 h post transfection using primary anti-ALIX antibodies labeled with the activator-reporter dye pair Alexa Fluor 488-Cy5 . TIRF images of the fixed samples revealed several individual HIV-1 budding sites at the cell membrane as well as immunostained ALIX protein assemblies , which were manually classified into colocalizing and non-colocalizing clusters ( Fig . 3A ) . Three distinct classes of colocalizing ALIX structures were identified from super-resolution data , namely condensed spots ( Fig . 3A , crop ( 1 ) ) , condensed spots with a surrounding , diffuse cloud-like structure ( Fig . 3A , crop ( 2 ) ) and diffuse cloud-like structures without a central spot ( Fig . 3A , crop ( 3 ) For the non-colocalizing clusters , only condensed spots were observed . Super-resolution STORM reconstructions of membrane associated ALIX structures in untransfected cells exclusively exhibited condensed spots of ALIX ( S5A Fig . ) , verifying that the cloud-like ALIX structures were budding site specific . Thus , the non-colocalizing round ALIX structures in cells expressing HIVmCherry ( Fig . 3A , circles ) likely represent ALIX localized with the plasma membrane in the course of other cellular processes , which is again supported by quantitative EM analyses showing ALIX localized with the plasma membrane of uninfected cells [61] . ALIX is known to associate with so-called exosomes [4 , 62] , intraluminal vesicles of multivesicular bodies ( MVBs ) , that fuse with the plasma membrane to release their content for cell-cell communication [63] . Thus , the spot-like assemblies of ALIX in cells not expressing HIV or non-colocalizing with HIV-1 budding sites might potentially be attributed to exosome-related ALIX structures captured during the process of membrane secretion . In the 10 cells analyzed , a total of 18 ALIX clusters with or without cloud structures were found to colocalize with HIVmCherry . In one case , it was not possible to clearly separate the central spot from the cloud structure; this site was excluded from further analysis . With the remaining 17 ALIX colocalizing structures , 3 . 4% of the HIV-1 assembly sites had an associated ALIX cluster , which is similar to the colocalization percentage observed for Tsg101 . The FWHM of the condensed central ALIX spot ranged from 34 to 94 nm ( Fig . 3B , S5B Fig . ) with an average value of 64 ± 18 nm . Again , this is significantly smaller than the mean size of the HIV-1 bud . Both early-acting factors Tsg101 and ALIX appeared to form compact protein assemblies inside the confining structure of the budding neck . The localization of many ALIX proteins inside the budding neck would be necessary for any kind of neck stabilizing mechanism during HIV-1 budding or membrane scission . This is in good agreement with the proposed stabilizing function of ALIX in addition to its involvement in recruitment of downstream ESCRT factors . Analysis of the non-colocalizing ALIX clusters showed a broader size distribution ranging from 30 nm up to more than 130 nm ( Fig . 3C ) Thirteen out of the 17 ( 76% ) detected ALIX membrane clusters colocalizing with HIV-1 budding sites ( S1 Table ) displayed an additional diffuse cloud-like distribution of ALIX proteins surrounding the central condensed spot . This suggests that ALIX molecules , which directly interact with the LYPXnL or LXnLF L-domains of the NC domain of Gag [17 , 64] , are taken up into the HIV-1 bud . Incorporation of ALIX molecules in released virus like particles has already been shown using biochemical analyses [11 , 65–68] . Any unspecific membrane-association of ALIX molecules can be excluded due to the presence of the auto-inhibitory domain in ALIX [69–71] . When ALIX molecules are taken up into the bud , the distribution of ALIX molecules should be consistent with the size of a nascent HI virion . Since the number of ALIX specific signals detected in the area of the clouds was too low to apply a 2D-Gaussian fit , we performed a Ripley’s L-function analysis [49] . This analysis yielded cluster sizes of 95 to 210 nm with a mean value of 164 ±31 nm for the colocalizing ALIX cloud-like structures ( Fig . 3D , S5C Fig . ) . This is comparable to the distribution of HIV-1 bud sizes ranging from 80 to 180 nm with an average value of 141 ± 41 nm , when applying the same Ripley-L analysis . In addition , the size of the cloud structure shows a relatively narrow size distribution with no strong outliers . From our data , we also made a rough estimation of the minimum number of ALIX molecules incorporated into the virus Gag shell . Only the molecules were counted that were distinct from the central cluster and fell within the size of the ALIX cloud estimated by the Ripley’s L-function . As the same fluorophore can be photoactivated several times , reoccurring molecules that localized within our resolution of 40 nm of each other were counted as a single molecule . The distribution of ALIX molecules ranged from 2 up to 15 per cloud ( Fig . 3E ) with a mean value of 7 ± 5 molecules . Since multiple ALIX molecules may exist within a distance of 40 nm , this represents a minimum estimation of the number of ALIX molecules within the bud . In view of the different models for membrane fission , it was of particular interest to investigate the downstream ESCRT factors directly involved in the membrane scission process . CHMP4B , in contrast to the CHMP4A or CHMP4C isoforms , has a direct impact on viral infectivity and release as established from siRNA knockdown experiments [14] . Based on structural data , CHMP4B can be directly recruited to the HIV-1 assembly site via the Bro-domain of the early-acting factor ALIX ( Fig . 1A ) [29] . Therefore , we extend our studies on the membrane scission machinery to CHMP4B . As no antibodies against CHMP4B suitable for super-resolution imaging were available , we employed a version of the human isoform CHMP4B tagged with the small hemagglutinin ( HA ) tag that was subsequently immunostained with primary monoclonal anti-HA antibodies labeled with Alexa Fluor 488 and Cy5 . With use of the HA tag , we minimize the dominant negative effects on ESCRT recruitment that were reported for CHMP4 isoforms fused to fluorescent proteins [11 , 33] . In HeLa cells expressing CHMP4B-HA and HIV:HIVmCherry , both colocalizing and non-colocalizing CHMP4B protein clusters were observed around 14–15 h post transfection ( Fig . 4A ) . Considering the relatively high density of detected CHMP4 protein clusters on the plasma membrane ( Fig . 4A ) , the overall colocalization percentage of 1 . 5% for CHMP4B with respect to HIV-1 assembly sites is low . The low proportion of colocalizations detected in steady state is consistent with the transient recruitment of CHMP4B to the membrane for a period of approximately 2–3 min during the complete HIV-1 budding process , as shown by Jouvenet et al . [13] in live-cell imaging studies . Analysis of the 23 CHMP4B-HA clusters colocalizing with HIV-1 assembly sites showed a narrow cluster size ( FWHM ) distribution ranging from 35 to 85 nm with a mean value of 56 ± 12 nm ( Fig . 4B ) . Analysis of the non-colocalizing CHMP4 clusters showed a much broader size distribution ranging from 20 nm up to more than 160 nm ( Fig . 4C ) . This observation strongly suggests that colocalizing CHMP4 protein assemblies form inside a restricting structure such as the budding neck whereas non-colocalizing CHMP4 clusters , which could not be correlated to nascent HIV-1 buds , have larger freedom in spatial spreading . Some non-colocalizing CHMP4 protein assemblies formed diffuse clusters at the membrane that were significantly larger than the spatial resolution obtained under our conditions ( circles , Fig . 4A ) . These large clusters were also observed in cells transfected with CHMP4B-HA in the absence of HIV-1 ( S6B Fig . ) but not in control experiments where the HeLa cells were not expressing CHMP4B-HA ( S6A Fig . ) . These clusters were not considered in further analyses . Again , we performed experiments in HeLa cells co-expressing the HIV:HIVmCherry late- mutant ( 1:1 ) together with CHMP4B-HA ( S6C Fig . ) . We observed a reduction in the total number of CHMP4B clusters as well as in the number of clusters colocalizing with the HIVmCherry late- mutant ( only 1 colocalization was observed for over 200 HIV-1 late- assembly sites ) . This is consistent with what is expected for a partially recruitment defective Gag variant containing a disrupted PTAP late- motif , supporting that CHMP4B-HA behaves similar to the untagged protein under our conditions . Analysis of the size distribution of the non-colocalizing CHMP4B structures in the late- experiments showed a similar distribution to non-colocalizing structures in experiments with wildtype HIV-1 , which is broader than observed for the structures that colocalized with wildtype HIV-1 ( Fig . 4B-4D ) . We also performed super-resolution microscopy with CHMP2A , which is the last ESCRT protein recruited to the HIV-1 budding site and , in turn , recruits VPS4 . Endogenous CHMP2A proteins were directly visualized using primary polyclonal anti-CHMP2A antibodies and secondary anti-bodies fluorescently labeled for dSTORM imaging . TIRF images of fixed cells revealed , similar to the previous experiments , distinct single HIV-1 budding sites and compact CHMP2A protein assemblies at the membrane that either colocalized with the HIV-1 assembly sites or not ( Fig . 5A ) . The colocalization of CHMP2A with respect to HIV-1 assembly sites was 2 . 0% , similar to the values determined for Tsg101 and CHMP4B but smaller than for ALIX . Super-resolution STORM images of CHMP2A clusters at HIV-1 assembly sizes revealed small , condensed spots ( Fig . 5A ) that showed a size distribution ( FWHM ) smaller than what we observed for Tsg101 , the central spot of ALIX clusters and CHMP4B . It ranged from 38 to 82 nm with a mean value of 56 ± 12 nm ( Fig . 5B ) . This value is in good agreement with the size of CHMP2 structures found in in vitro EM measurements by Effantin et al . [72] . In control experiments using untransfected HeLa cells , STORM imaging of the CHMP2A clusters showed condensed , circular structures at the cell membrane ( S7A Fig . ) . Analysis of the size distribution of the non-colocalizing CHMP2A clusters reveals a broader size distribution than determined from the colocalizing structures ranging from 36 nm up to 138 nm ( Fig . 5C ) with a mean size of 59 ± 20 nm . Again , a narrower distribution of colocalizing CHMP2A clusters suggests that the cluster size is limited by the HIV-1 budding neck . The cluster sizes determined here for all cellular proteins investigated differ significantly from those reported by Van Engelenburg et al . [47] and Bleck et al . [73] . In order to define the reason behind the apparent discrepancy , we performed control experiments employing FP-tagged variants of Tsg-101 and either labeled Gag in the full viral context ( HIV:HIVmCherry ) or labeled Gag alone ( Gag:Gag . mCherry ) . Similar to what we observed for endogenous Tsg101 , microscopy of cells co-transfected with YFP-Tsg101 fusions in the context of the full viral construct HIV:HIVmCherry ( 1:1 ) [48] revealed condensed circular structures ( S8A–S8B Fig . ) . Most strikingly , we detected a much higher colocalization rate between HIVmCherry and YFP-Tsg101 , detected using anti-Tsg101 antibodies , increasing from 1 . 8% up to ~18% . This effect was even stronger when HIV:HIVmCherry was replaced by Gag:Gag . mCherry alone . The number of colocalizations between Gag . mCherry and immunostained YFP-Tsg101 ( S8D–S8E Fig . ) increased by over a factor of 10 up to 30% compared to results obtained for endogenous Tsg101 . Though the increase in colocalization is mostly like induced by changes in the dynamics of the ESCRT interaction , it is conceivable that the higher fraction of colocalizations is due to better accessibility of the antibody epitope in the FP-Tsg101 clusters . In either case , there is a clear impact of the FP labeling . The dSTORM images of the YFP-Tsg101 clusters when coexpressed with either HIV:HIVmCherry or Gag:Gag . mCherry showed condensed circular structures of 60 ± 19 nm and 60 ± 10 nm , respectively ( S8C and S8F Fig . ) . The value for YFP-Tsp101 in the presence of Gag:Gag . mCherry is similar to what we observed for endogenous Tsg101 clusters in the context of full viral construct HIVmCherry ( Fig . 2C ) . To ensure that the increase in colocalization is not an artifact due to over expression of Tsg101 , we also transiently expressed Tsg101-FLAG-IRES-GFP in HeLa cells . The small FLAG tag has been shown not to influence HIV-1 budding [74] and the IRES-GFP tag allowed us to identify the cells that were overexpressing Tsg101 . The percentage of HIV-1 assembly sites with a colocalizing Tsg101 cluster ( S4D Fig . ) was similar to the value for endogenous Tsg101 ( 2 . 9% ) and there was no detected difference in the size of the Tsg101 clusters ( S4E–S4F Fig . ) . For confirmation , we directly compared the size of the Gag shell of the HIV assembly site and the CHMP4B structure in the same bud by performing a combined STORM/PALM experiment in cells expressing CHMP4B-HA and HIV:HIVmEos . In the super-resolution image reconstruction of the CHMP4B and HIVmEos TIRF images ( S9A Fig . ) , the spot-like CHMP4B cluster in this case was situated on the rim of the HIV bud . The overlay of both super-resolution images ( S9B Fig . ) clearly illustrates the factor of two size difference between the CHMP4B protein assemblies and the HIV-1 bud as supported by the differences in the corresponding FWHM of the Gaussian fits ( S9C Fig . ) . We have used super-resolution imaging of immunostained endogenous ESCRT components to investigate the architecture of ESCRT complexes involved in HIV budding and release . The super-resolution images of Tsg101 , ALIX punctae , CHMP4B and CHMP2A membrane clusters that colocalized with HIV-1 budding sites all revealed a similar size distribution with an average cluster diameter that was significantly smaller than the average diameter of an HIV-1 bud . This is consistent with the observation that ALIX can function as a scaffold for stabilizing CHMP4 filaments within the neck as previously reported from in vitro experiments [30] . The most striking difference between our experiments with endogenous ESCRT proteins or CHMP4B-HA and other measurements using fluorescent protein fusions of the ESCRT machinery is the very low fraction of HIV-1 assembly sites with colocalizing ESCRT complexes . The transient interaction of ESCRT proteins on the timescale of a few minutes [13 , 60] with nascent assembly sites is expected to lead to the low percentage of colocalizations . Using deep-etch electron microscopy of endogenous ESCRT proteins in VPS4 depleted cells , Cashikar et al . [40] observed flat rings or conical spirals with an average diameter of 108 ± 30 nm surrounding either a central membrane protrusion of ~ 50 nm diameter or Gag assembly sites . In our studies , we did not observe any ESCRT I ( Tsg101 ) , ALIX or ESCRT-III ( CHMP4B and CHMP2A ) protein structures colocalizing with HIV-1 assemblies sites with average sizes of 100 nm or larger . The larger cluster size may be attributable to the depletion of VPS4 used by Cashikar et al . , which was necessary to increase the number of colocalizations between ESCRT factors and HIV assembly sites for electron microscopy . To check this possibility , we performed dSTORM experiments with CHMP2A in the presence of dominant negative VPS4 mutant . As expected , the number of colocalizations increased from 2 . 0% to 8 . 5% . In addition , we observed occasional colocalizing structures with sizes above 100 nm ( S7C–S7E Fig . ) similar to what we observed for non-colocalizing protein clusters at the membrane . The non-colocalizing structures are most likely linked to other cellular processes such as exosome activity . Therefore , the confined size distribution ( FWHM ) of less than ~60 nm for the various ESCRT protein clusters we measured is most likely determined by the constriction into scaffold structures smaller than the bud , such as the budding neck . Due to the low colocalization percentages in our experiments , the statistics are low and we cannot conclude with certainty that each ESCRT component is restricted to the neck region of the HIV-1 assembly site . However , the fact that no larger complexes were observed for all the ESCRT proteins investigated is a much stronger observation and supports an internal constriction mechanism where most of the ESCRT components are situated close to or within the HIV-1 budding neck ( Fig . 1B , Model 2 ) [7 , 38] rather than protein structures restricting the bud neck from outside ( Model 1 ) . A recent study using iPALM reported ESCRT protein clusters with diameters > 100 nm inside the Gag shell [47] which points towards a membrane constriction coming from within the bud ( Model 3 ) . As the structures we measured are significantly smaller than those from Van Engelenburg et al . , the difference cannot be attributed to our two-dimensional measurements or differences in lateral resolution of 38 nm for our measurements versus their calculated iPALM resolution of 25 nm . The difference cannot be attributed to insufficient labeling as we have demonstrated our ability to label structures inside the bud using eGFP . vpr and antibodies against GFP ( S3 Fig . ) . The major differences between our experiments and the results from Van Engelenburg et al . is their use of exogenous expressed FP-tagged versions of the ESCRT proteins , while we have , in most cases , performed immunostaining of endogenous proteins in the context of the full virus genome . It has been shown that ESCRT factors are very sensitive to the incorporation of additional tags that can affect the structure [33 , 40] and functionality [11] of the complexes . The results from our control experiments using YFP-tagged Tsg101 demonstrated an over 10-fold increase in the number of ESCRT complexes colocalizing with the Gag assembly sites ( S1 Table ) . However , we observed no significant difference in cluster sizes for the YFP-Tsg101 construct with Gag . mCherry ( 60 ± 10 nm FWHM , 89 ± 26 nm for the Ripley’s analysis ) compared with endogeneous Tsg101 colocalizing with HIVmCherry ( 60 ± 19 nm FWHM , 85 ± 27 nm for the Ripley’s analysis ) . The increase in the colocalization of ESCRT factors with HIV-1 assembly sites when using fluorescent protein-tagged ESCRT conjugates demonstrates that there may be differences in recruitment dynamics and potentially structure even when it was reported that the constructs showed no detectable difference in phenotype . In addition , Van Engelenburg et al . detected significant amounts of eGFP-Tsg101 , eGFP-CHMP2A and eGFP-CHMP4B proteins inside released viral particles by their eGFP fluorescence intensity and by immune-gold staining of cryo-TEM images using anti-eGFP antibodies . These results are inconsistent with the low amount of ESCRT proteins detected in released HIV-1 virus like particles as determined using western blot and mass spectrometry [11 , 17] . In recent experiments by Bleck et al [73] co-expressing Gag alone with FP-tagged ESCRT proteins , ESCRT proteins were observed to be significantly shifted with respect to the Gag assembly , inconsistent with ESCRT proteins being located in the lumen of the VLP ( model 3 ) . Hence , we also analyzed the position of the ESCRT complexes with respect to the center of the HIV-Gag signal . Consistent with the findings of Bleck et al [73] , we often observed the ESCRT complexes on the edge , rather than the center , of the HIV bud ( S10 Fig . ) . Hence , our data also support the model of ESCRT proteins being localized in the neck region of the HIV bud ( model 2 ) rather than being incorporated into the bud itself ( model 3 ) . As the cluster sizes we are measuring are not much larger than our localization precision of ~ 40 nm ( S2 Table ) , the question arises regarding what impact our measurements could have on the actual sizes of the clusters . In super-resolution microscopy , as in normal optical microscopy , the final image is a convolution of the actual structure with the point-spread function of the optical system . For super-resolution microscopy , the point-spread function is determined by the localization precision . Therefore , the actual size of the ESCRT clusters will be smaller than the measured value . In addition , we have used primary and secondary full antibodies for labeling ESCRT proteins for STORM imaging . This can contribute up to an additional 30 nm ( 15 nm per antibody ) to actual size of the ESCRT cluster [75] . Thus , the actual size of the ESCRT protein clusters at HIV-1 assembly sites is significantly smaller than our measured 60 nm , which would be closer to the late bud neck diameter and certainly lies below the diameter of ~50 nm for the central membrane protrusions observed by Cashiker et al . [40] . To distinguish between different models for ESCRT-III arrangement within the neck , 3D images with resolution of much better than ~ 40 nm will be necessary . As discussed above for the ESCRT machinery , the actual size of the HIV-1 clusters is also convoluted with the localization precision of the PALM experiments . Even though the PALM experiments had a lower resolution ( ~70 nm ) than the STORM experiments , the correction factor for HIV-1 is similar or smaller than for the ESCRT complexes due to the larger size of the HIV-1 buds . In addition , the size of the FP tag is much smaller than the size added to the detected ESCRT complexes when using the full primary and secondary antibodies for labeling . A simple estimation of the corrected HIV-1 particle size yields a FWHM of ~ 95 nm rather than 116 nm , which is still significantly larger than the size of the observed ESCRT clusters . In the case of ALIX , additional molecules were often ( 76% ) detected in a cloud surrounding the central dense cluster . These cloud-like structures were exclusively found in colocalizations with HIV-1 assembly sites . HeLa cells not expressing HIV-1 exhibited only condensed spot-like clusters with a size and shape comparable to the central ALIX spots of colocalizing constructs in pCHIV-transfected cells . These results support the previous observation that ALIX is also located at the plasma membrane of uninfected cells [61 , 63] . A Ripley L-function cluster analysis of these cloud distributions revealed a clear correlation of the ALIX cloud to the size of a nascent virus particle , indicating that the diffuse , cloud-like ALIX distribution is based upon individual ALIX proteins being incorporated into budding Gag shells . This is consistent with previous studies that have shown that ALIX is present in released virus particles using a western blot assay [11 , 17] . Accumulation of ALIX at the budding site during virus assembly has been previously visualized in live-cell experiments with equine infectious anemia virus [13] . In contrast , transient recruitment of FP-tagged ALIX was at the end of HIV-1 bud assembly has recently been described [60] . Less than 20% of the recruited ALIX proteins were found remaining in the released viruses after fission [60] , consistent with the relatively low number of ALIX proteins we detect in the cloud-like structures . Even though both Tsg101 and ALIX are known to bind to the p6-domain of Gag [17 , 19] , we did not observe any cloud-like structures for Tsg101 . Thus , the persistence of ALIX at the budding site even after fission may suggest that ALIX has an additional role beyond recruitment of ESCRT factors to the budding site of HIV-1 . In summary , by using super-resolution fluorescence imaging of endogenous CHMP2 and HA-tagged CHMP4 , we provide evidence for a membrane scission process driven from inside the HIV-1 budding neck by ESCRT-III protein assemblies including CHMP4B and CHMP2A . Protein recruitment factors acting early in the HIV-1 budding process such as Tsg101 and ALIX were also located in condensed clusters similar to the dimension of the neck and significantly smaller the HIV-1 bud . In addition , ALIX showed diffuse localizations surrounding the central cluster within the neck that did not extend the calculated dimension of the bud , indicating the internalization of individual ALIX proteins into the nascent HIV-1 particle . The low number of colocalizing events of recruited endogenous Tsg101 , ALIX , CHMP2 and even of overexpressed HA-tagged CHMP4 with HIV-1 assembly sites supports the previously reported transient recruitment of those factors during virus particle membrane assembly rather than the accumulation of ESCRT proteins inside the bud . This stands in contrast to the much higher numbers of colocalization that was observed for the larger eGFP-Tsg101 protein fusion constructs and may lead to a higher accumulation probability of ESCRT proteins into the HIV-1 assembly sites . The resolution of STORM imaging is not yet high enough to resolve details below the diameter of the budding neck . Hence , the question is still open whether CHMP4 together with other ESCRT-III components form a spiral , whorl or a dome-like structure . The ongoing development of new labeling strategies and imaging techniques has the potential to answer this question and eventually unveil the membrane scission mechanism in the near future .
HIV encoding plasmid pCHIV and its labelled derivatives HIVeGFP , HIVmCherry and HIVmEos have been previously described [48 , 76] . The Vps4A-E228Q-mCherry plasmid was also described previously [45] . Plasmid synGag was obtained from Ralf Wagner ( University of Regensburg , Germany ) [77] , pGag . eGFP was contributed by Marilyn Resh [59] , and peGFP . Vpr was kindly provided by Tom Hope [78] . pGag . mCherry was derived from pGag . eGFP by replacing a BamHI/BsrGI fragment comprising the eGFP coding sequence with a corresponding restriction fragment comprising the mCherry ORF generated by PCR . The pCHMP4B-HA plasmid encoding CHMP4B fused to an HA-tag was kindly provided by H . Göttlinger [79] . The YFP-Tsg101 plasmid was a kind gift of Wesley Sundquist ( University of Utah , Salt Lake City , USA ) . HeLa cells ( Japanese Collection of Research Bioresources Cell Bank , Osaka Japan ) were grown in Dulbecco’s modified Eagle’s medium ( DMEM ) , supplemented with 10% fetal calf serum ( FCS ) . HeLa cells were seeded into LabTek II 8-well chamber slides at a cell density of 2*104 cells/well and transfected on the following day using X-tremeGENE transfection reagent ( Roche ) according to the manufacturer’s instructions . For the CHMP4B-HA/ HIV:HIVmCherry experiments , 100 ng of CHMP4B-HA , 50 ng of pCHIV ( wildtype ) and 50 ng of pCHIVmCherry were transfected per well . For the study of endogenous ESCRT proteins Tsg101 , CHMP2A and ALIX , 50 ng pCHIV and 50 ng pCHIVmCherry were used or , for PALM imaging of the HIV-1 assembly sites , 50ng pCHIV and 50 ng pCHIVmEos were used . After transfection , cells were incubated for 14–15 hours at 37°C and 5 . 0% CO2 . For validating immunostaining efficiency within the viral bud , 75 ng of pEGFP . vpr and 50 ng of synGag ( wildtype ) and 50 ng of Gag . mCherry or Gag . eGFP were used for transfection , respectively . For experiments with Tsg101-fusion proteins , 50 ng of YFP-Tsg101 and 50 ng of pCHIVmCherry and pCHIV were used for transfection , respectively . In the case of Gag . mCherry , the pCHIV plasmids were replaced by 50 ng synGag and 50 ng of Gag . mCherry , respectively . For Vps4A-depletion experiments , 70 ng pCHIVeGFP , 70 ng pCHIV and 35 ng Vps4A-E228Q-mCherry were used for imaging of CHMP2A and cells were transfected with 50 ng CHMP4B-HA , 35 ng pCHIVeGFP , 35 ng pCHIV and 35 ng Vps4A-E228Q-mCherry for imaging CHMP4B-HA . For STORM imaging , immunostaining of Tsg101 was performed using primary mouse monoclonal anti-Tsg101 ( clone 4A10 , #GTX70255; Genetex ) and secondary donkey anti-mouse IgG ( #ABIN336468 , purchased via antibodies-online . com ) antibodies . For labeling of secondary anti-mouse antibodies , the unconjugated antibody was mixed with Alexa Fluor 488 succinimidyl ester ( Invitrogen ) and Cy5 bis-NHS-ester ( GE Healthcare ) in a molar ratio of 1:4:1 ( antibody: Alexa Fluor 488: Cy5 ) in 150 mM NaHCO3 buffer ( pH 8 . 2 ) and incubated overnight . Unreacted dye molecules were subsequently removed by gel permeabilization chromatography ( Performa DTR Gel Filtration Cartridges from Edge BioSystems ) . For dSTORM experiments , secondary antibodies were only labeled with Cy5-bis-NHS-ester in a molar ratio dye:protein = 1:4 . Monoclonal primary mouse anti-ALIX antibodies ( clone 3A9 , #634502 , BioLegend ) were directly labeled for immunostaining of ALIX according to the protocol above as was also done for monoclonal primary anti-HA antibodies ( clone 3F10 , #11867423001; Roche ) that were used to stain CHMP4B-HA . For immunostaining of CHMP2A , we used unlabeled primary polyclonal rabbit anti-CHMP2A antibodies ( #ab76335 , abcam ) combined with secondary donkey anti-rabbit IgG ( #ABIN376979 , purchased via antibodies-online . com ) that were labeled according to the protocol above . Anti-GFP primary polyclonal antibodies from rabbit ( #ABIN121945 , purchased via antibodies-online . com ) were used for labeling of peGFP . Vpr fusion protein together with the same secondary anti-rabbit IgG antibodies than were used for CHMP2A . STORM sample preparation followed a slightly modified version of the protocol for microtubule immunostaining for STORM imaging described in [80] . HeLa cells were fixed not earlier than 14–15 h post transfection using 4% ( v/v ) paraformaldehyde solution ( PFA , Electron Microscopy Sciences ) for 15 min . Cells were permeabilized with 1% ( v/v ) Triton X-100 for 2 min . The cells were again rinsed twice with PBS . Unspecific antibody binding was blocked by incubating the sample with a buffer containing 3% bovine serum albumin ( BSA ) and 0 . 2% Triton X-100 for 30 min followed by immunostaining with primary antibodies for 60 min . In case of additional labeling with secondary antibodies , the sample was then stained with labeled secondary antibodies for 45 min . Finally , post-fixation was done by treating the cells with 4% ( v/v ) paraformaldehyde solution in PBS for 10 min . STORM and dSTORM measurements were carried out using a glucose oxidase based oxygen scavenging buffer with mercaptoethylamine ( Fluka ) as reducing agent as described in [80] . PALM measurements were carried out using PBS buffer . TIRF imaging and all PALM , STORM and dSTORM measurements were carried out on a combined TIRF and wide-field ( WF ) microscope as depicted in S1 Fig . A 561-nm diode-pumped solid-state laser ( CrystaLaser , Reno , NV , USA ) was used to excite Gag . mCherry and the red state of Gag . mEos . For photoconversion of mEosFP , we used a 405-nm diode laser ( LuxX 405–120 , Omicron Laserage , Rodgau , Germany ) . Excitation of GFP- and YFP-fusion proteins and activation of Cy5 for STORM imaging was achieved by exciting Alexa488 with a 488-nm diode laser ( LuxX 488–60 , Omicron Laserage , Rodgau , Germany ) . Activated Cy5 was excited by a 642-nm diode laser ( PhoxX 642 , Omicron Laserage , Rodgau , Germany ) . We performed wavelength selection by coupling the different laser lines into an acousto-optic tunable filter ( TF525-250-6-3-GH18A , Gooch & Housego , Ilminster , UK ) . A set of two lenses then expanded the excitation beam in order to increase the field-of-view illuminated by TIRF excitation , before the light was focused on the back focal plane of a 100x N . A . 1 . 49 Apo TIRF oil immersion lens ( Nikon , Tokyo , Japan ) . A rectangular glass prism introduced into the beam path after the focusing lens allowed switching between wide-field to TIRF imaging by changing the displacement of the beam path relative to the optical axis of the objective . A dichroic mirror ( Di01-R405/488/561/635-25x36 , Semrock , Rochester , NY , USA ) then directed the excitation light into the microscope body , which consisted of a home-built microscope stage built out of Inwar to reduce thermal drift . After passing the dichroic mirror , the collected fluorescence signal passed a set of emission filters to select a specific fluorescence wavelength: a 670/30 bandpass filter was used for Cy5 ( Laser 2000 ) , a 535/22 bandpass filter for GFP and YFP ( FF01-535/22-25 , Semrock , Rochester , NY , USA ) and a 593/40 bandpass filter ( FF01-593/40-25 , Semrock , Rochester , NY , USA ) for the red state of Gag . mEos and Gag . mCherry . Finally , the fluorescence signal was projected onto an EMCCD camera ( DU860D-CS0-BV , Andor , Belfast , UK ) . The resulting pixel size was 120 nm . ESCRT proteins were measured with either STORM or dSTORM , where a movie stack comprising 10 , 000 frames was acquired at a frame rate of 20 Hz , where each activation cycle consisted of one activation frame ( λexc = 488 nm ) , followed by 9 imaging frames ( λexc = 642 nm ) . Laser power at the output of the objective was 50 mW for imaging . For activation , the power of the blue laser was ~0 . 5 μW at the beginning of data acquisition for STORM and 5 . 0 μW for dSTORM and successively increased over the course of the experiment up to ~ 2 mW . This was necessary to counteract gradual bleaching of the activator dye . For PALM measurements of viral Gag protein , an analogue protocol was used with λexc = 405 nm for activation ( increasing power starting at 5 . 0 μW to 2 mW ) and λexc = 561 nm ( 1 . 1 mW ) for imaging . The data analysis protocol is based on the procedure described by Rust et al . [81] and by Bates et al . [82] and implemented into a self-programmed analysis software using MATLAB ( MathWorks , Natick , MA , USA ) . The same algorithms were used for PALM , STORM and dSTORM experiments . First , structures in the fluorescent image representing local maxima were isolated in a square 11 x 11 pixel window . The point spread function of the fluorescent molecule was evaluated by fitting these regions to a continuous ellipsoidal 2D Gaussian using Levenberg-Marquard’s nonlinear least-squares algorithm . The intensity I of the Gaussian distribution at coordinates ( x , y ) is given by: Ix , y = A+ I0e0 . 5* - ( x-x0σx2- ( y-y0σy2 where A is the background intensity , I0 the maximum amplitude of the distribution , x0 and y0 the coordinates of the centroid and σx and σy the standard deviations in x and y-direction , respectively . Sample drift was corrected by pixel-wise cross-correlation of each frame n of the image stack with the first frame of this stack . The normalized cross-correlation function Gnx , y for each frame is given by: Gnx , y = ∑i∑jI1i , jIni+x , j+y∑i∑jI1i , j2∑i∑jIni+x , j+y20 . 5 The intensity I1 from the first frame at coordinates ( i , j ) is correlated with the intensity In from the nth frame at coordinates ( i+x , j+y ) . The respective drift in x- and y is given by the position in ( x , y ) of the maximum of Gnx , y . In order to reduce fluctuations caused by fluorophore blinking , the obtained drift function was fitted to a polynomial and the fit function used for drift correction . We rejected all molecules where the Gaussian function fitted to the point spread function showed an ellipticity , E , higher than 15% ( S11A Fig . ) . E depends on the Gaussian standard deviations in x and y-direction ( σx and σy ) and is defined as: E=| σ x − σ y σ x + σ y | Points that appeared in two or more consecutive frames within a distance smaller than 1 px ( 120 nm ) were considered as originating from the same fluorescent molecule . The positions determined from individual frames where the same molecule was observed were averaged for rendering of the final super-resolution image . Diffraction limited spots that appeared for only one frame or molecules with less than 300 detected photons were discarded . Localization displacements of single fluorophores molecules were used to estimate the actual resolution of our system by means of the FWHM value of the Gaussian function fit to the resulting displacement histograms in the x- and y-direction respectively ( S11B Fig . ) . The average STORM image resolutions for the different analyzed proteins are summarized in S2 Table . For the final STORM image rendering with a pixel size of 12 nm in the super-resolution images , each localized molecule was represented by a 2D Gaussian function with a fixed amplitude I0 = 1000 counts and a fixed standard deviation σx = σy = 1 . 2 px = 14 . 4 nm . Diffraction limited TIRF images of ( d ) STORM/PALM measurements , which were used to identify colocalizations of ESCRT and HIV-1 buds , were emulated by building the average time projection of the acquired image stack . An self-written image analysis algorithm was developed in ImageJ Macro language [83] and used to consistently assess the number of HIV-1 assembly sites and clusters and determine their colocalization . Images were individually analyzed as follows: First , a convolution filter ( Gaussian blur ) followed by background subtraction ( “rolling ball” algorithm [84] ) were applied . Next , point objects were selected based on their intensities ( local maxima ) and a multi-point selection was created . Objects corresponding to either assembly sites or clusters were then segmented by a watershed approach . As a last step , center of brightness , distribution of intensities and area of each object were measured . These readouts were collected , analyzed accordingly and generated the statistical information presented in Figs . 2–5 . Gag assemblies with an area > 0 . 860 μm² were excluded from further evaluation . ESCRT clusters were counted by rendering a STORM image with a pixel size equal to the widefield pixel size of 120 nm and all clusters with an intensity lower than 5 , 000 counts were discarded as well as all clusters where no distinct structure could be obtained in the final STORM image with a pixel size of 12 nm . Object sizes in the final STORM or PALM images were either estimated by means of the average full-width at half-maximum ( FWHM ) of 1-D Gaussians fitted to two orthogonal 1-D cross-sections through the middle of the respective cluster as demonstrated in S2B , S4B , S5B , S6D and S7B Figs . , or alternatively using Ripley’s L-test [49] . In the case of estimation of the size of the cloud surrounding ALIX clusters , the points contributing to the central clusters were excluded from Ripley’s analysis to achieve more accurate results for the size of the cloud ( see S5C Fig . ) .
|
Viruses hijack the cellular machinery to complete their life cycle . In the case of HIV-1 , the endosomal sorting complex required for transport ( ESCRT ) is recruited by nascent viruses to release themselves from infected cells . Currently , there has been an intense amount of research on how the ESCRT machinery induces viral release . Using super-resolution imaging with endogenous ESCRT proteins or ESCRT proteins containing a small tag , we are able to provide insight into how ESCRT leads to budding of HIV-1 . Super resolution imaging of the early ESCRT factors Tsg101 and ALIX , as well as later factors CHMP4B and CHMP2A , also showed condensed , circular structures with diameters of roughly 60 to 90 nm . The cluster sizes were significantly smaller than that of the HIV-1 bud and the distribution of cluster sizes that colocalized with nascent HIV-1 assembly sites were narrower than for non-colocalizing structures . This indicates that the point of interaction between the ESCRT machinery and the HIV-1 assembly site is in the bud neck .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Super-Resolution Imaging of ESCRT-Proteins at HIV-1 Assembly Sites
|
The development of rapid serodiagnostic tests for sleeping sickness and other diseases caused by kinetoplastids relies on the affordable production of parasite-specific recombinant antigens . Here , we describe the production of recombinant antigens from Trypanosoma brucei gambiense ( T . b . gambiense ) in the related species Leishmania tarentolae ( L . tarentolae ) , and compare their diagnostic sensitivity and specificity to native antigens currently used in diagnostic kits against a panel of human sera . A number of T . b . gambiense protein antigen candidates were chosen for recombinant expression in L . tarentolae based on current diagnostics in field use and recent findings on immunodiagnostic antigens found by proteomic profiling . In particular , the extracellular domains of invariant surface glycoprotein 65 ( ISG65 ) , variant surface glycoproteins VSG LiTat 1 . 3 and VSG LiTat 1 . 5 were fused with C-terminal histidine tags and expressed as soluble proteins in the medium of cultured , recombinant L . tarentolae . Using affinity chromatography , on average 10 mg/L of recombinant protein was purified from cultures and subsequently tested against a panel of sera from sleeping sickness patients from controls , i . e . persons without sleeping sickness living in HAT endemic countries . The evaluation on sera from 172 T . b . gambiense human African trypanosomiasis ( HAT ) patients and from 119 controls showed very high diagnostic potential of the two recombinant VSG and the rISG65 fragments with areas under the curve between 0 . 97 and 0 . 98 compared to 0 . 98 and 0 . 99 with native VSG LiTat 1 . 3 and VSG LiTat 1 . 5 ( statistically not different ) . Evaluation on sera from 78 T . b . rhodesiense HAT patients and from 100 controls showed an acceptable diagnostic potential of rISG65 with an area under the curve of 0 . 83 . These results indicate that a combination of these recombinant antigens has the potential to be used in next generation rapid serodiagnostic tests . In addition , the L . tarentolae expression system enables simple , cheap and efficient production of recombinant kinetoplatid proteins for use in diagnostic , vaccine and drug discovery research that does not rely on animal use to generate materials .
Human African Trypanosomiasis ( HAT ) , also known as African sleeping sickness , is usually a fatal disease caused by the parasites Trypanosoma brucei gambiense ( T . b . gambiense ) or T . b . rhodesiense [1–3] . The parasite is transmitted by the bite of infected tsetse flies in sub-Saharan Africa with T . b . gambiense being responsible for 95% of the cases in West and Central Africa . The remaining infections are caused by T . b . rhodesiense in East and Southern Africa . Although the number of infections currently reported at less than 5 , 000 cases per year are not at the level reported during the last century ( 300 , 000 per year ) , this disease still causes considerable suffering and burden on communities in terms of disability-adjusted life years [4 , 5] . The disease follows two stages where the trypanosomes are limited to the blood and lymphatic systems initially , but in most cases will invade the central nervous system . This second stage causes sleep cycle disruption and neurological damage leading to coma and death if not treated [6] . Diagnosis of sleeping sickness is complex as detection of the parasite in a patient is very difficult and , in the case of a positive result , involves a subsequent lumbar puncture to assess the disease stage [7] . Following an infective bite , the patient suffers peaks of parasitaemia as the host mounts a humoral response to the changing surface antigens on the trypanosome , in particular the variant surface glycoproteins ( VSGs ) . Parasites attempt to evade this humoral attack by expressing alternative VSGs ( antigenic variation ) and continue to multiply until the new variable antigen type ( VAT ) is recognised by the host and attacked [8] . One of the most common VAT is considered to be LiTat 1 . 3 and its VSG is the major antigen present in the Card Agglutination Test for Trypanosomiasis ( CATT ) which is currently used in large scale population screening [9] . In the field , this is followed by a series of time-consuming parasitological tests that often are used to reach a diagnosis by the application of a complex algorithm [10] . Recent efforts to improve diagnosis have focussed on developing rapid immunochromatography-based serodiagnostic tests ( RDT’s ) , which as point of care tests ( POCT ) , should follow the WHO ASSURED criteria; affordable , sensitive , specific , user-friendly , rapid , equipment-free and deliverable to the people at need [11] . First generation RDTs look very promising but have lower specificity than expected when tested on a wider target area [12–14] . Current RDTs for HAT rely on the infection of rodents and the purification of the native VSGs LiTat 1 . 3 and LiTat 1 . 5 , which do not represent all the variants encountered in the field . Other approaches to identifying immunodiagnostic antigens have involved soluble fractions of other native VSGs ( sVSG117 ) and the proteomic identification and selection of other surface antigens [15 , 16] . The latter has identified a number of antigens that were recognised by the sera of HAT infected patients and in particular the invariant surface glycoprotein 65 ( ISG65 ) showed potential , either on its own or in combination with sVSG117 . Other potential antigens could not be expressed as recombinant proteins in E . coli and were not investigated further . However , unlike prokaryotic expression systems , eukaryotes have the ability to carry out more complex folding and post-translational modifications ( PTMs ) such as glycosylation and Pichia pastoris has recently been used successfully to express soluble VSGs [17] . Sugar moieties are known to be major antigenic structures by themselves and /or in combination with their folded protein partner [18] . For recombinant expression the main aim is to reproducibly produce material as close to the native type as possible without having to handle highly virulent parasites , and in an economic system . In this study we have examined the potential application of the related trypanosomatid organism Leishmania tarentolae ( L . tarentolae ) as a recombinant antigen production system [19] . It is a Biosafety handling level one system , which can be genetically engineered and scaled up to bioreactor production using readily available medium and equipment . We have expressed , purified and compared recombinant VSG fragments from L . tarentolae , with VSG’s purified from their native counterparts . We have also examined the potential of rISG65 from L . tarentolae to act as an antigen in serodiagnostic tests for sleeping sickness caused by both T . b . gambiense and T . b . rhodesiense . Testing was carried out on serum from confirmed HAT patients and controls from a wide range of endemic areas . We show that the L . tarentolae system can express recombinant antigens that are as effective as the native antigens for use in RDTs for HAT , and suggest this may have a wider application as a production platform of recombinant proteins for diagnostic and vaccine generation applications .
Part of the serum samples used in this study were collected within a diagnostic study ( SeroStrip ) carried out in the Democratic Republic of the Congo ( DRC ) [13] . Permission for this study was obtained from the national ethics committee of DRC ( CNG/M . D . /111/2012 ) and from the ethics committee of the University of Antwerp ( 11435795 ) . The other sera were obtained from the World Health Organization HAT Specimen Bank [20] . The specimen collection and banking was approved by the WHO Ethical Review Committee and the different national ethical committees in each country where specimens were collected . The National Ministries of Health also gave their approval . All individuals gave their written informed consent for the use of their plasma specimen in HAT research before providing blood . All specimens were anonymised . Table 1 represents the serum collection used in this study . Serum donors were classified as HAT patients when trypanosomes were detected in any body fluid ( blood , lymph , cerebrospinal fluid ) . Non-HAT controls were persons from HAT endemic regions but without history of HAT and without clinical , serological or parasitological evidence of infection with T . b . gambiense or T . b . rhodesiense . One hundred forty one sera from gambiense HAT ( g-HAT ) patients and from non-g-HAT controls were collected within the SeroStrip study conducted in the DRC [13] . Additional sera from g-HAT patients and non-g-HAT controls and from rhodesiense HAT ( r-HAT ) patients and non-r-HAT controls were received from the World Health Organization HAT Specimen Bank [20] and originated from Guinea , Tchad and DRC ( T . b . gambiense ) and from Malawi and Tanzania ( T . b . rhodesiense ) . Peptides were identified using data from proteomic studies on proteins recognised by g-HAT patient sera that were matched to an ISG65 from T . b . brucei ( UniProt reference Q26712 and Q58F5 ) [16] . Using protein-protein BLAST a 98% identical T . b . gambiense homologue ( UniProt reference C9ZJ77 ) was selected for expression and investigation in this project . The DNA coding for amino acid residues 19–385 of this homologue was commercially synthesized by Gene Art ( LifeTechnologies ) . The DNA for recombinant VSG LiTat 1 . 3 ( UniProt reference X5GEX5 ) and VSG LiTat 1 . 5 ( UniProt reference E7EDN2 ) were kindly donated by S . Rogé , ITM , Belgium [17] . It had previously been shown that N-terminal peptides contain more specific epitopes [17] so the native DNA sequences coding for amino acid residues 24–372 for VSG LiTat1 . 3 and amino acids 33–426 for VSG LiTat 1 . 5 were used . All constructs were cloned into the LEXSYS vector pLEX hyg2 ( Jena Bioscience ) in frame with the signal peptide of secreted acid phosphatase of L . mexicana and a carboxy-terminal hexa-histidine tag within the vector . Following electroporation of LEXSYS host strain p10 , clonal isolates were selected as previously described [19] . Production of recombinant protein was carried out in 1 litre baffled Erlenmeyer flasks in BHI medium ( supplemented with antibiotics and hemin ) and the medium was harvested when the OD600 reached 4 ( approx . 70 h post inoculation , 108 cells/ml ) . All media components were from Jena Bioscience . Clarified medium was concentrated twenty fold on a Pellicon XL 50 Ultrafiltration cassette ( 10 kDa MWCO ) and diluted four times in binding buffer ( 20 mM Phosphate , 500 mM NaCl , 10 mM Imidazole ) before addition to an equilibrated His Trap ( GE Healthcare ) FPLC column . Bound proteins were eluted using an increasing Imidazole gradient ( 20 mM Phosphate , 500 mM NaCl , 500 mM Imidazole ) . Peak protein containing fractions , as determined by A280 nm measurement , were combined , desalted and concentrated by centrifugation in Amicon Ultra-15 device ( 30 kDa MWCO ) . The final protein was stored at 1 mg/ml in solution in PBS , 15% glycerol at -20°C . Following SDS-PAGE of purified recombinant proteins ( Fig 1 ) each band was excised from the gel , digested with Trypsin ( Sigma–Aldrich , P7367 ) and the resulting peptides subjected to matrix-assisted laser desorption ionization time of flight mass spectrometry ( MALDI-TOF-MS . Briefly , excised bands were reduced and alkylated with iodoacetamide and digested in-gel with trypsin ( Promega V5111 ) as previously described [21] . The tryptic peptides ( 0 . 5 μl ) were spotted onto an MTP AnchorChip 384 T F target plate ( Bruker part number 209514 ) and allowed to air dry . Peptides spots were then overlaid with an equal volume of α-cyano-4-hydroxycinnamic acid matrix ( 0 . 7 mg/ml in 85% acetonitrile , 15% water , 0 . 1% TFA and 1 mM NH4H2PO4 , Bruker ) . The same volume of peptide calibration standard II mix ( Bruker , made according to the manufacturer’s instructions ) was spotted in the appropriate positions and allowed to air dry . Mass spectra were collected on an UltrafleXtreme ( Bruker ) instrument and Mascot software used to match peptides to the SwissProt database . For gel filtration experiments , purified rISG 65 ( 100 μl of 1 mg/ml ) was applied to a gel filtration Superdex 200 10/300 GL column ( GE Healthcare LifeSciences ) in sodium phosphate buffer ( pH7 . 5 ) , 0 . 5 M NaCl previously calibrated with molecular weight markers ( Sigma-Aldrich MWGF100 ) and eluted with at a flow rate of 1 ml/min . In order to examine the potential presence of N-linked sugars , rISG 65 ( 1 μg ) was treated with PNGase F ( 1 . 5 U ) ( Sigma-Aldrich , P7367 ) for two hours at 37°C using manufacturers instructions before being analysed by SDS-PAGE to determine if this resulted in a shift in molecular weight . Native variant surface glycoprotein ( VSG ) was prepared following standard procedures from cloned populations of T . b . gambiense variant antigen types ( VATs ) LiTat 1 . 3 and LiTat 1 . 5 [22] . After purification , the native VSG LiTat 1 . 3 ( nLiTat 1 . 3 ) and native VSG LiTat 1 . 5 ( nLiTat 1 . 5 ) antigens were lyophilised in aliquots of 1 mg and stored at -80°C prior to use . The ELISA protocol was based on the procedure according to Lejon et al . [23–25] . Microplates ( Maxisorp , Nunc ) were coated overnight at 4°C with 100 μl/well of purified recombinant protein at 4 μg/ml or with native antigen at 2 μg/ml . All antigens were diluted in phosphate buffer ( PB ) ( 10 mM sodium phosphate , pH 6 . 5 ) . To correct for aspecific reactions , caused by contaminating L . tarentolae proteins that were not eliminated from the protein mixture by the one-step affinity purification , control wells were coated with the supernatant of a culture of untransfected L . tarentolae cells at 4 μg/ml . Further manipulations were undertaken at ambient temperature . After coating , the wells were blocked with PBS-Blotto ( 0 . 01 M sodium phosphate , 0 . 2 M sodium chloride , 0 . 05% NaN3 , 1% skimmed milk powder , pH 7 . 4 ) for 1 hour . Before addition to the microplate the sera were diluted at 1:150 in PBS-Blotto . Antibody binding was visualised with goat anti-human IgG ( H+L ) conjugated with horseradish peroxidase ( 1:40000; Jackson ImmunoResearch ) and the chromogen ABTS ( 2 , 2’-azinobis[3-ethylbenzothiazonline-6-sulfonic acid]-diammonium salt; Roche ) . The optical densities ( ODs ) were read at 414 nm ( Multiskan RC Version 6 . 0; Labsystems ) . Corrected optical density ( ODcorr ) values were calculated by subtracting for each serum the OD reading in the control well from the OD reading in the antigen coated well . ELISA results were captured in a Microsoft Excel 2010 database . The accuracy of the different antigens for diagnosis was determined in SigmaPlot 12 . 5 by calculation of the area under the receiver operator characteristics ( ROC ) curve ( AUC ) [26] . Confidence intervals ( CI ) were determined according to DeLong [27] . Sensitivities and specificities with 95% binomial Wilson confidence intervals and the Youden index were calculated using SigmaPlot 12 . 5 [28] . The McNemar Chi2 test was used to test differences in the AUCs .
Soluble forms of the T . b . gambiense invariant surface glycoprotein 65 ( ISG65 ) and the variant surface glycoproteins VSG LiTat 1 . 3 and VSG LiTat 1 . 5 ( with C-terminal histidine tags remaining intact ) were expressed as soluble proteins in the medium of the recombinant LEXSYS host strain p10 and purified on Ni-NTA resin . Protein purity was examined by SDS-PAGE analysis and subsequent Coomassie blue staining of the resulting gel to visualise those protein bands present ( Fig 1 ) . The identity of the recombinant proteins were confirmed by tryptic digestion of the excised bands ( Fig 1 ) followed by mass spectrometry and matching the peptides to the protein sequence databases [29] . Recombinant constructs for the VSG LiTat1 . 3 and LiTat1 . 5 had been previously expressed in Pichia pastoris as a secreted product with evidence of post-translational processing [17] . rLiTat1 . 3 has a predicted molecular mass of 38 kDa from its amino acid sequence with 2 potential N-glycosylation sites . rLiTat1 . 5 is predicted to be 40 . 5 kDa with no potential N-glycosylation sites . The presence of two bands in the purified material for the rVSGs is likely to arise from either different post-translational modifications of the two species present or , possibly , be due to proteolysis as observed by Rogé et al . [17] when expressed in Pichia pastoris . All bands were positively identified as corresponding to the respective VSG by mass spectrometry but this data and not allow us to determine conclusively the difference between the two bands observed . rISG65 is predicted from its amino acid sequence to be 41 . 8 kDa with 2 potential N-glycosylation sites . Further analysis of the purified rISG65 by gel chromatography confirmed the presence of a monomeric molecule of 50–60 kDa as shown in Fig 2 . Following PNGase F treatment to remove potential N-glycans , the apparent molecular mass of rISG65 was reduced , as determined using SDS-PAGE analysis , consistent with the presence of N-linked sugars ( Fig 3 ) . The sera from 172 g-HAT patients , 119 non-g-HAT controls and 50 non-r-HAT controls were tested by ELISA with rISG65 , rLiTat 1 . 3 , rLiTat 1 . 5 , nLiTat 1 . 3 and nLiTat 1 . 5 . With the parasitological status as reference and the cut-off for each antigen set at its highest Youden index ( sensitivity + specificity -100 ) , each antigen displayed a sensitivity > 92 . 4% and a specificity > 94% ( Table 2 ) . The diagnostic potential of each antigen in function of varying ODcorr cut-off is represented in the Receiver Operating Curve ( ROC ) plots in Fig 4 showing areas under the curve ( AUC ) ranging from 0 . 97 to 0 . 98 for the recombinant antigens and from 0 . 98 to 0 . 99 for the native antigens ( Fig 4 and Table 2 ) . Pairwise comparison of the AUC obtained with the different antigens showed no statistically significant differences except between the AUC of rLiTat 1 . 3 and nLiTat 1 . 3 ( Table 3 ) . In a similar way , the sera from 78 r-HAT patients , 50 non-r-HAT controls and 50 non-g-HAT controls were tested in ELISA with all the antigens . The results are represented in Fig 5 as ROC plots and in Table 4 . Sensitivities and specificities ranged respectively from 37 . 2 to 79 . 5% and from 76 to 90% due to the poor reactivity of the antigens with r-sera . Significant differences in the AUC of the antigens were observed in more than half of the pairwise comparisons ( Table 5 ) .
As the numbers of people infected with sleeping sickness continue to fall , the targeted elimination of the disease by 2030 , as suggested by WHO , is conceivable [30] . However , many obstacles still need to be overcome including the development and availability of a reliable point of care test ( POCT ) and the development of oral delivery drugs which can be used in the remote areas where foci still exist . Such an assay will become even more essential for the monitoring and surveillance of remote areas in view of the targeted continuing decline of the prevalence . A recent examination of the currently available RDTs for serodiagnosis of g-HAT in West Africa demonstrated a lower specificity than expected ( 88% ) but suggested a parallel use of both available tests could increase specificity and sensitivity [14] . These first generation RDTs contain native ( n ) nLiTat 1 . 3 and nLiTat 1 . 5 as antigens but research is ongoing to replace them by second and future generation RDTs containing recombinant antigens . A recombinant fragment of ISG65 has been expressed in Escherichia coli and used as an antigen in a prototype lateral flow device on its own [16] and in combination with native VSGs , VSG117 [15] or VSG 117 purified from T . b . brucei [31] . Although these tests showed a good sensitivity ( 88% to 98% for the single antigen device ) , the specificities only ranged between 65% and 93% . In the dual antigen prototypes , the specificities varied from 83% to 97% [15 , 16 , 30] . Recombinant antigens expressed in E . coli are not glycosylated and therefore may miss some critical epitopes with diagnostic potential that may be present on glycosylated and correctly folded native glycoproteins , such as ISG65 . Correct glycosylation is often a requirement for proper folding of proteins and the absence of this in E . coli expression systems is likely to lead to both incorrect folding and absence of glycan containing epitopes . In contrast , eukaryotic expression systems , including insect and mammalian cells , yeast and protozoa , can yield glycosylated recombinants that can be engineered to be secreted into the culture medium [32] . The collection/harvesting of the medium can then be used as a first step in the purification process for such recombinant molecules . In this study we describe the use of the L . tarentolae expression system ( LEXSYS ) to express T b gambiense surface antigens as secreted recombinant molecules . The native and recombinant VSG LiTat 1 . 5 showed comparable diagnostic accuracy when tested with g-HAT sera and non-HAT controls in the ELISA format as did the recombinant and native VSG LiTat 1 . 3 . This suggests that recombinant VSG fragments could eventually replace the native forms currently used in RDT tests thus eliminating the complex and dangerous process of purifying native antigens from living , highly virulent infective parasites grown in laboratory rodents . Our results are similar to those reported for the expression of recombinant VSG LiTat 1 . 3 and VSG LiTat 1 . 5 fragments in Pichia pastoris [17] . Although VSGs LiTat 1 . 3 and LiTat 1 . 5 are T . b . gambiense specific and are not expressed in T . b . rhodesiense infections , both the recombinant and native forms of VSGs LiTat 1 . 3 and LiTat 1 . 5 reacted with a considerable number of sera from r-HAT patients yielding a sensitivity of up to 80% for native VSG LiTat 1 . 5 . This could be due to epitopes on these VSG fragments that are not variant-specific but common to other variants , including variants expressed by other subspecies of T . brucei [18] . Among the recombinant antigens investigated here , it was rISG65 that showed the highest diagnostic potential for r-HAT with an AUC of 0 . 83 . As ISG65 belongs to the invariant surface glycoprotein set of T . brucei the antigen should also be recognised by sera from g-HAT that do not contain antibodies against VSG LiTat 1 . 3 and LiTat 1 . 5 . However , most sera from g-HAT patients used in the current study have been pre-selected by using CATT/T . b . gambiense as screening test during active case detection in the field . Therefore , a carefully designed prospective study with unbiased inclusion of participants is needed to confirm the hypothesis that rISG65 could detect patients that are not reactive with recombinant or native VSG LiTat 1 . 3 or VSG LiTat 1 . 5 . We conclude that rLiTat 1 . 3 , rLiTat 1 . 5 and rISG65 can be expressed and post-translationally processed and secreted by L . tarentolae in a manner similar to the related kinetoplastid Trypanosome species . Moreover , the ease of engineering and production of recombinant proteins by this system allows for the development of tests for single or multiple diseases by combining different antigens on a single lateral flow test . We expect that rISG65 , in combination with one or another recombinant VSG , will lead to the development of a powerful serodiagnostic test for g-HAT that will meet the ASSURED criteria for NTD diagnostics [11] . In addition L tarentolae has the potential to be used for the expression of antigens for other neglected tropical diseases caused by kinetidoplastid organisms as it appears to process the recombinant proteins in the same manner as the native antigens .
|
The development of rapid serodiagnostic tests for African sleeping sickness and other diseases caused by kinetoplastids relies in part on the affordable production of parasite-specific recombinant antigens . The majority of cases of sleeping sickness are caused by the parasite Trypanosoma brucei gambiense ( T . b . gambiense ) which is transmitted when bitten by an infected tsetse fly . Existing tests rely on the utilisation of extracts from the parasite or use antigens raised in animal models . In this study we have shown that using a cell culture system devised from a parasite similar to T . b . gambiense recombinant antigens can be produced that are as effective in rapid diagnostic tests as the native antigens purified from T . b . gambiense parasites grown in laboratory rodents . We compared the diagnostic sensitivity and specificity of the antigens we produced recombinantly to native antigens currently used in diagnostic kits against a panel of human sera . The evaluation on sera from 172 T . b . gambiense patients and from 119 controls without sleeping sickness showed very high diagnostic potential of two recombinant antigens where the response was not significantly different to that from the native antigens . These results indicate that a combination of these recombinant antigens has the potential to be used in next generation rapid serodiagnostic tests .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Expression of Trypanosoma brucei gambiense Antigens in Leishmania tarentolae. Potential for Use in Rapid Serodiagnostic Tests (RDTs)
|
Examining the fundamental structure and processes of living cells at the nanoscale poses a unique analytical challenge , as cells are dynamic , chemically diverse , and fragile . A case in point is the cell membrane , which is too small to be seen directly with optical microscopy and provides little observational contrast for other methods . As a consequence , nanoscale characterization of the membrane has been performed ex vivo or in the presence of exogenous labels used to enhance contrast and impart specificity . Here , we introduce an isotopic labeling strategy in the gram-positive bacterium Bacillus subtilis to investigate the nanoscale structure and organization of its plasma membrane in vivo . Through genetic and chemical manipulation of the organism , we labeled the cell and its membrane independently with specific amounts of hydrogen ( H ) and deuterium ( D ) . These isotopes have different neutron scattering properties without altering the chemical composition of the cells . From neutron scattering spectra , we confirmed that the B . subtilis cell membrane is lamellar and determined that its average hydrophobic thickness is 24 . 3 ± 0 . 9 Ångstroms ( Å ) . Furthermore , by creating neutron contrast within the plane of the membrane using a mixture of H- and D-fatty acids , we detected lateral features smaller than 40 nm that are consistent with the notion of lipid rafts . These experiments—performed under biologically relevant conditions—answer long-standing questions in membrane biology and illustrate a fundamentally new approach for systematic in vivo investigations of cell membrane structure .
Because of the inherent complexity of living cells , high-resolution analytical methods are generally used ex vivo , and many depend on labels to differentiate the species of interest . In the case of cellular membranes , their nanoscopic structure ( ultrastructure ) has been elucidated primarily through electron microscopy ( EM ) with heavy-atom labels[1] , supplemented with in vitro biophysical studies of model membranes . [2–12] In vivo membrane studies at lower resolution , notably with fluorescence microscopy , have provided additional details of lateral structure as well as important information on diffusion and other dynamic processes . [13] Through these techniques , a vivid picture of the membrane has emerged . [14–19] Nonetheless , they have significant limitations , and fundamental questions about the ultrastructure of the membrane in vivo remain unresolved . [20] For example , membrane hydrophobic thickness is a basic structural parameter with implications for membrane transport , as well as membrane protein folding and function . Yet the hydrophobic thickness of membranes has never been determined in vivo . Another outstanding question concerns the existence of nanoscopic domains , or lipid rafts . The lipid raft hypothesis[21] invokes lateral organization of membrane lipids and proteins into distinct domains in the plane of the membrane to facilitate the assembly and regulation of multi-molecular complexes . This hypothesis provides a compelling rationale for numerous observations relating to membrane trafficking , endocytosis , signal transduction , and other processes . [22–25] However , evidence for lipid domains has been largely inferential , and it is now widely believed that these domains are nanoscopic as well as transient[26 , 27] , making them difficult to detect . New methods to interrogate lateral structure in vivo—ideally without the use of extrinsic probes—would help us to understand the role of lateral heterogeneity in the many cellular processes where it has been implicated . In this context , small-angle neutron scattering ( SANS ) has emerged as a uniquely powerful tool for the study of lipid bilayer structure in vitro . Neutrons have wavelengths on the order of Ångstroms ( Å ) and are thus inherently nanoscopic probes , well-matched to the dimensions of membrane structure . Using appropriately designed experiments , both transverse ( normal to the membrane plane ) [7] and lateral ( within the membrane plane ) [28 , 29] structure can be accurately determined . Importantly , neutron-scattering techniques do not require extrinsic molecules or heavy atoms as labels and rely instead on hydrogen ( H ) /deuterium ( D ) isotopic substitution . Cold and thermal neutrons are also nondestructive[30–32] , making them ideal for the study of living systems . However , applications of powerful neutron-based ultrastructural methods have been restricted to in vitro membrane models due to the complex scattering signal arising from cells as a result of their diverse biomolecular composition . Scattering and imaging experiments require that the feature ( s ) of interest have an observable signal , contrasted from the background . Where contrast is insufficient in the native system , it must be imparted through the use of labels . In the case of fluorescence imaging , contrast arises from the distinctive excitation and emission properties of native , or more commonly , introduced fluorophores . With X-rays and electrons , scattering contrast arises from differences in electron density , which scale with atomic number . Neutrons , on the other hand , are scattered by atomic nuclei , and the key contrast parameter—analogous to electron density—is the scattering length density ( ρ ) . Neutron scattering lengths ( b ) are unrelated to atomic number and , in fact , vary among isotopes of the same element ( S1 Fig ) . Most importantly , the scattering lengths of hydrogen ( bH = −3 . 74 fm ) and deuterium ( bD = 6 . 67 fm ) are substantially different . [31] Thus , neutron contrast is unique in that it can be varied in hydrogen-rich biological systems with H/D isotopic labels , as opposed to the heavy-atom labels used in X-ray and electron scattering , or fluorescent labels used in microscopy . Different classes of biomolecules ( i . e . , proteins , lipids , carbohydrates , and nucleic acids ) have different elemental compositions , which gives each class a different value of ρ ( Fig 1a and 1b; Table A in S1 Text ) . Multicomponent systems can be designed such that 2 or more components have the same ρ , for example , protein in approximately 42% D2O/H2O . In this case , the protein and solvent are said to be contrast-matched , and the protein does not generate a distinct scattering signal—i . e . , it is effectively invisible to neutrons . [30] If a third component is introduced having a different ρ ( e . g . , DNA ) , it then becomes the only contributor to the net scattering . Through judicious H/D-labeling of the sample and the aqueous medium , ρ for the different biomolecules can be tuned to enhance or attenuate their scattering . Thus , contrast can be varied without changing the chemical composition of the system . In this report , we describe a chemical–biological approach to controlling neutron contrast variation in vivo and its application to the study of cell membrane ultrastructure . As a platform , we chose the gram-positive bacterium B . subtilis . This organism has a number of attractive features , such as being genetically tractable , having a well-characterized lipid metabolism , and growing readily in deuterated media . Importantly , it has a single membrane and uses only saturated fatty acids ( FAs ) that can be prepared readily in deuterated form to allow tuning of membrane contrast . After suppressing cellular contrast through global D-labeling ( i . e . , replacing most of the H with D throughout the cell ) , we selectively reintroduced contrast to the membrane by supplying H-FAs , which the cell incorporated . In this way , we were able to isolate the SANS spectrum of the membrane , which showed it to be a lamellar structure with a hydrophobic thickness of 24 . 3± 0 . 9 Å . A modification of the contrast-labeling scheme led to the observation of nanoscopic lipid structures of <40 nm within the plane of the cell membrane , providing experimental evidence for the existence of nanoscopic lipid domains ( lipid rafts ) in an active biological membrane .
The neutron scattering signal from a cell reflects the sum of contributions from all cellular components , here , taken to be water , protein , RNA , DNA , carbohydrate , and lipid ( Fig 1a ) . Each of these has a characteristic ρ ( Fig 1b ) , and from their relative abundances in B . subtilis , [33] we estimated the total scattering from the organism as a function of D2O concentration in the aqueous environment . Water accounts for approximately 85% of the cellular volume and exchanges rapidly across the membrane . Therefore , immersion of cells in deuterated buffer changes the neutron contrast , ergo the net scattering , with the relative contributions of the different molecular species varying as a function of D2O concentration ( Fig 1c; S2 Fig ) . We then measured the total scattering from B . subtilis cultured in standard M9 minimal medium ( H-M9 ) and resuspended in phosphate buffered saline ( PBS ) at different D2O concentrations . The observed scattering corresponded closely to predictions of the simple compositional model , as shown in Fig 1c . Note that the cells scatter strongly over the entire range of percent D2O , and the total signal at any concentration of D2O reflects contributions from the multiple classes of biomolecules present in the cell . Structural analysis based on the overlapping signals is impractical , if not impossible . To isolate a meaningful scattering signal from the membrane , it was first necessary to suppress neutron contrast from the entire cell by manipulating its H/D composition . This objective was achieved by culturing B . subtilis in deuterium-enriched M9 minimal medium ( D-M9 , prepared in 90% D2O with H-glucose as the carbon source ) . Due to metabolic H/D exchange , deuterium from the growth medium becomes permanently incorporated into the carbon skeletons of biosynthetic molecules . Overall , skeletal deuteration was approximately 70% ( S4 Fig , Tables B and C in S1 Text ) , which was predicted to create a near-contrast matched condition for all biomolecules when the cells were immersed in approximately 85% D2O buffer ( Fig 2a ) . This expectation was borne out in the scattering experiment , where a strong reduction in the total scattering was observed at approximately 85% D2O ( Fig 2b , open circles; S2 Fig ) . With contrast suppressed , the next step was to reintroduce contrast specifically into the membrane by providing exogenous FAs for incorporation into the membrane phospholipids . Membrane FAs , after conversion to FA methyl esters ( FAMEs ) , are readily analyzed for structure and isotope substitution by gas chromatography/mass spectrometry ( GC/MS ) . As shown in Fig 2c ( top ) , B . subtilis uses a mixture of 7 main FAs , all of which are saturated ( have no double bonds ) , and all but one of which are branched . [34 , 35] Initial supplementation experiments with wild-type B . subtilis showed that exogenous FAs were not incorporated into the membrane lipids intact , so the native pathways for both catabolism and anabolism of FAs had to be blocked ( Fig 2d ) . Catabolism was blocked genetically by the deletion of yusL , the gene encoding a critical enzyme in β-oxidation , enoyl-CoA hydratase . [36] Anabolism was blocked chemically using cerulenin , an irreversible inhibitor of β-ketoacyl-ACP synthase that suppresses de novo FA biosynthesis . [37] This combination resulted in conditional FA-dependent growth ( demonstrated in S5 Fig ) . Cerulenin-induced growth inhibition can be rescued with a mixture of just 2 of the native complements of FAs—palmitic acid ( normal-hexadecanoic acid [n16:0] ) and 12-methyltetradecanoic acid ( anteiso-pentadecanoic acid [a15:0] ) . [38] These 2 constitute a minimal set of 1 high-melting ( n16:0 ) and 1 low-melting ( a15:0 ) FA , with which the cell can regulate the fluidity and structure of its membrane . Unlabeled ( H ) and perdeuterated ( D ) forms of these 2 FAs were then used in various mixtures to tune neutron contrast in the membrane . Growth of B . subtilis ΔyusL in D-M9 medium did not alter the FA composition of the membrane ( Fig 2c , upper and middle panels; see S4 Fig for peak assignments and Table B in S1 Text for peak areas and deuteration analysis ) . However , when cerulenin-treated ΔyusL cells were grown in the same D-M9 medium , supplemented with H-n16:0 and H-a15:0 , their membrane FAs were found to consist exclusively of these 2 FAs , with no deuterium incorporated from the medium ( Fig 2c , bottom panel ) . The total neutron scattering from these cells showed a large increase in scattering at 85% D2O ( marked by the blue arrow in Fig 2b ) , which can be attributed entirely to the H-FAs incorporated into the membrane phospholipids . Having imparted contrast to the membrane ( Fig 3a ) , we used SANS to determine the transverse membrane structure by recording the scattered intensity , I ( q ) , as a function of the scattering wavevector , q . Cells for this experiment were cultured as described above in D-M9 medium , supplemented with cerulenin and H-n16:0 and H-a15:0 to label the membrane , then transferred to 85% D2O buffer . Because these are more time-consuming experiments , the cells were resuspended in an 85% D2O buffer supplemented with glucose , Mg2+ , and cerulenin . These additives prevent autolysis and preserve the membrane potential , [39 , 40] such that cells in suspension remained >90% viable over a period of 4 h at 25°C , as determined by direct cell counts , optical density measurements , and live/dead staining ( S6 and S7 Figs ) . The residual background was recorded using cerulenin-treated ΔyusL cells , which were fed a mixture of FAs contrast-matched to 85% D2O ( a15:0 and n16:0 , each 30% H and 70% D ) , and which do not contribute substantially to the net neutron scattering signal . In SANS , the shape of the I ( q ) versus q data describes the structure of the sample on the order of Ångstroms to approximately 100 nm . Formally this data is modeled using a lamellar form factor representing the acyl core of the bilayer , with the 2 identical head group regions on either side , contrast-matched to the scattering density of the surrounding cellular environment . The lamellar form factor is characterized by a q−2 dependence at low-q , arising from the 2D membrane surface of the entire bacterium , transitioning to a q−4 dependence typical of 3D objects with smooth surfaces and sharp interfaces—more detail is available in the literature[41–44] and the materials and methods . Subtraction of the background from the sample scattering revealed a pure membrane spectrum , which displayed a lamellar form factor characteristic of a lipid bilayer ( Fig 3b ) , with the expected q−2 dependence at low-q ( dashed line ) . A fit of the data ( solid red line ) revealed the average membrane hydrophobic thickness ( 2DC ) to be 24 . 3 ± 0 . 9 Å at 25°C ( S8 Fig ) . The hydrophobic thickness of the living B . subtilis membrane is thus comparable to that of synthetic phosphatidylcholine ( PC ) bilayers , such as dimyristoyl PC ( 2DC = 25 . 7 Å at 30°C ) or 1-palmitoyl-2-oleoyl PC ( 2DC = 28 . 8 Å at 30°C ) . [46] It is also comparable to that of purified basolateral plasma membranes from rat hepatocytes ( approximately 26 Å ) , as measured by small-angle X-ray scattering , [6] highlighting the conservation of membrane structure across animal and eubacterial kingdoms . Finally , we sought to examine the lateral structure within the membrane to determine whether the lipids in the B . subtilis membrane are uniformly mixed or display nanoscopic organization , as predicted by the lipid raft hypothesis . [21] Canonical mammalian lipid domains are believed to be enriched in the high-melting lipids , cholesterol and sphingomyelin . Bacteria generally lack these lipids , but are nonetheless believed to have lipid domains[47–53] formed from lipid species playing analogous roles . [54 , 55] The expected hallmark of lipid domains , in any system , is the lateral separation of higher- and lower-melting lipids with associated proteins into distinct phases . We recently introduced a contrast-variation strategy to detect lateral lipid organization in synthetic vesicles using SANS[28 , 29] . This strategy relies on differential H/D labeling of the lipid phases to control contrast in the plane of the membrane . With a suitable labeling strategy , the separated phases can be either contrasted or matched with respect to each other and the buffer . A particularly important case is where a mixture of lipids has an average contrast matching that of the buffer , but partitions into H- and D-enriched phases that contrast each other and the buffer . Experimentally , uniform mixing creates a null-scattering condition , whereas phase separation in the plane of the membrane ( domain formation ) induces neutron contrast and an observable signal in the SANS spectrum . In S9 Fig and Table F in S1 Text we present a series of similar SANS experiments which show how neutrons are sensitive to contrast in the plane of the bilayer and how neutron contrast can be controlled by thermally induced mixing of the 2 phases or by selecting specific isotopic mixtures which result in contrast-matched phases . The in vivo adaptation of this strategy compares scattering from cells with 2 different H- and D-FA mixtures that are , on average , contrast-matched to the medium ( Fig 4a ) . In the control mixture , there can be no contrast regardless of whether or not the membrane lipids are uniformly mixed , because all species are present at the same H/D ratio . However , in the experimental mixture , de-mixing among lipids creates contrast as described above , producing a measurable increase in the scattered intensity as a result of local inhomogeneities in the H/D distribution . We implemented this strategy in cerulenin-treated B . subtilis ΔyusL cells using an experimental mixture of 100% D n16:0 ( high-melting ) and 40/60 H/D a15:0 ( low-melting ) . When corrected for the relative abundances of these FAs in the cell ( S10 Fig ) , this mixture creates an average membrane contrast matching the 30/70 H/D FA ratio ( a15:0 and n16:0 , each 30% H and 70% D ) used for the control mixture ( inset to Fig 4b ) as well as the rest of the cellular components , including the 85% D2O buffer . SANS spectra from B . subtilis fed the experimental mixture reproducibly displayed an excess scattering over the q range 0 . 015–0 . 2 Å−1 compared to cells fed the control mixture ( results from a repeat experiment are shown in S11 Fig ) . As the only source of contrast in the experiment was the membrane H/D FA pool , this result supports the notion that there is lateral de-mixing of lipids containing the high- and low-melting FAs on a length scale ℓ of 3–40 nm ( ℓ = 2π/q ) ( Fig 4c ) , consistent with the lipid raft hypothesis .
Understanding how the nanoscale structure of biological systems relates to function is a challenging , ongoing pursuit . One difficulty is that only a few probes are capable of directly interrogating structure at this scale: electrons , X-rays , and neutrons . Electrons have enjoyed the widest application in cell biology , and EM remains the single most powerful tool for studying cellular ultrastructure . With regard to the B . subtilis membrane , a cryo-EM study of sectioned , freeze-substituted cells provided a striking picture of the cellular architecture , including the cell wall . [56] However , the structure of the plasma membrane was not well determined , and its thickness was estimated at 66 ± 8 Å , a surprisingly large value . Our results complement this EM picture of the cell envelope by providing a high-resolution hydrophobic thickness determination obtained under physiological conditions . X-ray and neutron scattering have been widely used for studying the structure of model membranes composed of defined lipid mixtures[3–5 , 28 , 29 , 46] or natural lipid extracts . [6–12] X-ray scattering has also been used recently for the ex vivo study of cellular membranes , [6] but its application to intact cells is confounded by the issue of background scattering from water and biomolecules . Neutron scattering uniquely provides a solution to the background problem in the form of isotopic contrast variation . We have shown here that in vivo contrast variation through metabolic labeling can effectively suppress scattering from the complex cellular milieu , while highlighting specific features of interest , even when they arise from minor components such as lipids ( approximately 1% of cellular wet mass ) . Furthermore , the cold neutrons used for scattering experiments are well suited for studies on living cells because of their low kinetic energy ( <0 . 025 eV ) and their nonionizing character—in contrast to high-energy X-ray and electron beams ( >5 , 000 eV ) . Prior applications of neutron scattering in vivo have relied upon external solvent contrast , only , which in some cases has been sufficient to observe Bragg scattering from repeat structures in thylakoid membranes[57 , 58] and mitochondria . [59] These studies have not revealed membrane structure per se , but have provided information on the arrangement of closely packed membranes . By creating internal contrast , i . e . , by differentially labeling specific cellular components , we have , for the first time enabled high-resolution ultrastructural measurements on a single membrane . In this work , we demonstrated the power of a chemical-biology–based approach to create selective internal contrast , thereby enabling high-resolution measurements of the in vivo membrane thickness . Our in vivo contrast variation approach also provides a new tool to study lateral membrane structure in living cells . Neutron-based structural methods offer distinct advantages in that they report nanoscopic lipid structure directly , without the need for models or extrinsic probes . Indications of nonuniform mixing[60 , 61] within the plasma membrane emerged contemporaneously with the landmark fluid–mosaic model proposed in 1972 , [16] the concept of membrane domains became well established by the mid-1970s , [62] and the lipid raft hypothesis was formalized in 1997 . [21] Nonetheless , the existence of lipid domains has remained controversial , [63] and because they are believed to be both transient and smaller than the diffraction limit of light ( 200 nm ) , they eluded observation by conventional microscopic techniques . Recently , however , ultra-resolution fluorescence microscopy was used to identify diffusionally restricted islands on the scale of 20 nm in the plasma membrane of rat kidney epithelial cells . [6] Our observation of lipid segregation on a comparable scale in the plasma membrane of B . subtilis is consistent with the existence of analogous lipid domains in bacteria and supports the notion that nanoscopic lipid assemblies are an integral feature of biological membranes . The critical barrier that has prevented application of high-resolution neutron scattering techniques in vivo was lack of a means to create internal contrast . In this work , we overcame the barrier and showed that B . subtilis is an ideal in vivo model system for the application of neutron contrast variation strategies . Through specific growth conditions and select genotypes , we were able to attenuate cellular contrast globally and precisely reintroduce contrast into the membrane . With the ability to control both the chemical and isotopic properties of the membrane lipids , we were able to interrogate both transverse and lateral membrane structure . The same general approach to selective contrast can potentially be extended to other biomolecules and model organisms for applications outside the membrane arena . More immediately , the in vivo experimental platform can be used to investigate the response of the plasma membrane to a diverse range of physical , chemical , genetic , and environmental stimuli . We anticipate that this capability will therefore prove valuable in many areas , such as antibiotic development , biofuel production , membrane protein function , and understanding the interplay between the membrane , cytoskeleton and cell wall in creating a protective , adaptable , multifunctional interface .
Deuterium oxide ( 99 . 9% D ) and algal amino acids ( unlabeled and uniformly D-labeled , with an isotopic purity of 98% ) were obtained from Cambridge Isotope Laboratories . Palmitic acid ( H-n16:0 ) and palmitic acid-d31 ( D-n16:0 ) were obtained from Sigma–Aldrich . 12-Methyltetradecanoic acid ( H-a15:0 ) was purchased from Sigma–Aldrich or prepared from s-butyl magnesium chloride ( Sigma–Aldrich ) and 11-bromoundecanoic acid ( Sigma–Aldrich ) according to the method of Baer and Carney[64] and purified by vacuum distillation . Perdeuterated D-a15:0 was prepared from H-a15:0 through 3 cycles of H/D exchange with D2O catalyzed by 10% Pt/C at 220°C as described by Yepuri et al . , [65] followed by chromatography on silica gel and vacuum distillation . The final product was chemically homogeneous and had an isotopic purity of 99% , as determined by analysis of its derived methyl ester by GC/MS . Cerulenin was obtained from Alfa Aesar ( Tewskbury , MA ) and stored in the dark at –80°C as a solid . FA-free bovine serum albumin ( BSA , catalog number A8806 ) was obtained from Sigma–Aldrich ( St . Louis , MO ) . All other materials were obtained from commercial suppliers and used as received . B . subtilis 168 ( parent strain ) and a ΔyusL mutant ( strain BKE32840 ) were obtained from the Bacillus Genetic Stock Center ( The Ohio State University , Columbus , OH ) . The ΔyusL strain lacks enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase activity and is severely deficient in its ability to degrade exogenous long-chain FAs;[36] therefore , this strain was used for all experiments where exogenous FAs were supplied in the cultivation medium . General growth and maintenance of B . subtilis was performed in either Luria–Bertani ( LB ) rich medium or M9 minimal medium with 2% glucose supplemented with 5 mM l-tryptophan . Solid media were prepared by the addition of 1 . 5% Bacto or Noble Agar ( Difco ) to LB or M9 medium , respectively . Erythromycin was added to 0 . 5 μg/mL for routine maintenance of BKE32840 . Cultures were incubated at 37°C , and liquid cultures were aerated by shaking at 250 rpm . Where applicable , supplemental FAs were added to a final concentration of 8 mg/L each of a15:0 and n16:0 from 25 mg/mL stock solutions in ethanol , along with 10 g/L of FA-free BSA as a carrier to aid solubility . Cerulenin ( 10 mg/mL in ethanol stock solution ) was prepared fresh and added immediately prior to inoculation . The required final concentration was determined empirically and varied by supplier and batch . For the work described here , cerulenin from Alfa Aesar was used at a final concentration of 50 μg/mL , which was sufficient to fully suppress endogenous FA synthesis , as judged by inhibition of growth , and rescued by exogenous FA ( S5 Fig ) . Partially deuterated cells were grown in M9 , prepared using 90% v/v D2O and H-glucose . This medium produced approximately 60%–70% deuteration of the carbon skeletons in biosynthetic molecules ( analysis described below ) . B . subtilis 168 and BKE32840 were adapted to growth in M9-Gluc + 5 mM l-tryptophan prepared with 90% ( v/v ) D2O by serial passage ( 1:100 inoculum ) in media prepared with increasing concentrations of D2O ( H2O:D2O 100:0 , 50:50 , and 10:90 ) . Cerulenin-treated , FA-supplemented cells were grown starting from a culture of untreated , unsupplemented cells . The starter culture ( OD600 0 . 8–1 . 0 ) was diluted 1:20 in cerulenin/FA medium to an OD600 of ca . 0 . 05 , incubated for growth to an OD600 of 0 . 8–1 . 0 , then passaged by dilution ( 1:20 ) into fresh medium . FA composition was monitored at each passage by GC/MS; 5–8 passages were required to clear the native FAs and adapt the cells to the FA-dependent condition . Control cultures without supplied FA ( no-growth controls ) were monitored in parallel for each experiment to ensure that the cerulenin remained active during the incubation period . Optical densities were determined at 600 nm using a Synergy Mx plate reader ( BioTek , Winooski , VT , USA ) , using a 96-well microtiter plate and 300-μL well volumes . For contrast experiments ( Figs 1 and 2 , and S2 Fig ) , B . subtilis 168 cells were resuspended in PBS ( 10 mM Na2HPO4 , 1 . 8 mM KH2PO4 , 137 mM NaCl , and 2 . 7 mM KCl ) , prepared either with H2O or D2O ( H-PBS or D-PBS , respectively ) , then mixed in appropriate proportions . For all H cells ( Fig 1c ) , 250 mL of fresh culture ( OD600 = 1 . 5 ) in H-M9 medium was split into 2 parts , which were processed in parallel . Cells were harvested by centrifugation at 6000 × g for 20 min at 4°C , then washed by centrifugation/resuspension with 3 × 10 mL of H- or D-PBS , allowing 5 min for equilibration at each step . The washed cell pellets were then resuspended in 11 mL of the same buffer to provide a cell concentration approximately 50 mg/mL wet weight , equivalent to approximately 10 mg/mL dry weight . For deuterated cells ( Fig 2 ) , the same procedure was followed except that a 125-mL culture grown in D-M9 prepared with 90% D2O was used , providing a final cell concentration of approximately 5 mg/mL dry weight . Cell suspensions were loaded into quartz “banjo” cells ( diameter 22 mm , path length 1 mm ) for study by SANS . Measurements were made at 37°C , using a single detector position , and the total scattering was analyzed as described below . For membrane structural analyses ( Figs 3 and 4 , S7 and S11 Figs ) , FA-fed B . subtilis 168 ΔyusL cells were grown in M9-Gluc prepared with 90% D2O . Cells from a 35-mL culture were harvested at an OD600 of 0 . 5 , washed with 3 × 3 mL of PBS , prepared with 85% D2O , and resuspended in 1 mL of the same buffer , providing a final cell concentration of approximately 5 mg/mL dry weight . Because of the long data collection times required ( up to 4 h ) , glucose ( 0 . 1% w/v ) , MgSO4 ( 10 mM ) , and cerulenin ( 50 μg/mL ) were added to the final resuspension buffer , the pH was reduced to 6 . 8 , and the measurements were made at 25°C to prolong cell viability and minimize autolysis . [39 , 40] As discussed below under cellular viability and integrity and shown in S6 and S7 Figs , cells remained >90% viable and displayed consistent SANS spectra over a period of 4 h under these conditions . Lipids were extracted from cells and characterized for FA content by GC/MS of derived FAMEs ( schematic description provided in S4 Fig ) . Total lipid extracts were prepared using a modification[66] of the method of Bligh and Dyer . [67] In brief , cells were pelleted by centrifugation at 6000 × g for 15 min , followed by 3 washes in 1% ( w/v ) NaCl . Samples were lyophilized in 10 mL glass test tubes with Teflon-faced screw caps , to each of which was sequentially added 0 . 5 mL of chloroform , 1 mL of methanol , and 0 . 4 mL of water , with vigorous agitation at each stage . This mixture forms a single phase and was left to stand for 18 h at room temperature with occasional agitation . After 18 h , phase separation was induced by the addition of 0 . 5 mL of chloroform and 0 . 5 mL water . The lipids were recovered from the lower chloroform phase by evaporation of the solvent in a new 10-mL glass test tube under an argon stream . FAMEs were prepared by acidic methanolysis of dried lipid extracts or intact cells . [68] Solvents , if present , were evaporated under a stream of argon prior to the addition of 1 mL of concentrated HCl/methanol ( 10% v/v ) . The test tube was then purged with argon , sealed , and heated to 85°C for 2 h . After cooling , 1 mL of water and 1 mL of hexane were added , and the contents were vortex-mixed . After phase separation , a portion ( approximately 700 μL ) of the upper phase was taken out for GC/MS analysis . Cellular amino acids were analyzed by GC/MS , as described by Dauner and Sauer . [69] Cells from 10 mL of culture at OD600 ≈ 0 . 7 were harvested by centrifugation , washed 3 times with water by resuspension/centrifugation , and stored frozen . For analysis , cells were resuspended in 1 mL of water in a microcentrifuge tube and lysed by sonication . After centrifugation at 14 , 000 × g for 15 min , a 500-μL portion of the supernatant was transferred to a glass vial and mixed with 1 . 5 mL of 6 M HCl . Standard solutions were prepared from H- or D-algal amino acid mixtures and processed in the same manner . The vials were sealed with Teflon-lined caps and heated at 110°C for 24 h to hydrolyze protein , after which the volatiles were removed by rotary evaporation . The hydrolysates were resuspended in tetrahydrofuran ( 100 μL ) and N-tert-butyldimethylsilyl-N-methyltrfluoroacetamide ( 100 μL ) and then heated to 60°C for 1 h to produce tert-butyldimethylsilyl amino acid derivatives suitable for GC/MS analysis . Samples were diluted with hexane 1:10 ( v/v ) prior to analysis . GC/MS analysis was performed using an Agilent 5890A gas chromatograph with a 5975C mass-sensitive detector operating in electron-impact mode ( Agilent Technologies , Santa Clara , CA ) . The instrument was equipped with an HP-5ms capillary column ( 30 m long , 0 . 25 mm outside diameter , and 0 . 25 μm coating thickness ) using helium at 1 mL/min as the carrier gas . Samples of 1 μL were introduced using splitless injection at an inlet temperature of 270°C . FAMEs were eluted using a temperature program of 2 min at 60°C; 20°C/min to 170°C; 5°C/min to 240°C; and 30°C/min to 300°C for 2 min . Derivatized amino acids were eluted with a temperature program of 2 min at 100°C; 10°C/min to 280°C for 2 min . ; and 25°C/min to 325°C for 2 min . Peak assignment , integration , and mass spectral analysis were performed using the instrument's ChemStation Enhanced Data Analysis software and the NIST mass spectral database . Peaks for deuterated compounds were identified on the basis of retention times and spectral comparison with nondeuterated compounds . The extent of deuteration was assessed by determining the gain in molecular mass for parent ions of FAMEs ( Table B in S1 Text ) or of selected fragment ions for amino acids ( Table C in S1 Text ) . SANS data were collected on Bio-SANS[70] and the Extended Q-range Small Angle Neutron Spectrometer ( EQ-SANS ) [71] at the at the High Flux Isotope Reactor and the Spallation Neutron Source , respectively , both located at Oak Ridge National Laboratory . Two-dimensional scattering data from both instruments were reduced using the Mantid[72] software and normalized to a porous silica standard to establish an absolute scale , and corrected for pixel sensitivity , dark current , and sample transmission . Background scattering was subtracted from the 1D intensity versus q , which is defined as: q= 4πsin ( θ ) /λ ( 1 ) where λ is the neutron wavelength and 2θ is the scattering angle relative to the incident beam . At EQ-SANS , the data were collected in 60 Hz mode with 2 instrumental configurations: 1 . 3-m sample-to-detector distance with 4–7 Å neutrons ( q = 0 . 050–0 . 4 Å−1 ) and 4 . 0 m sample-to-detector distance with 10–13 . 4 Å neutrons ( q = 0 . 009–0 . 07 Å−1 ) , yielding a total q-range from approximately 0 . 009 to 0 . 4 Å−1 . At BioSANS , 6-Å neutrons were used at 2 sample-to-detector distances , 2 . 5 m ( q = 0 . 050–0 . 30 Å−1 ) and 15 . 3 m ( q = 0 . 005–0 . 060 Å−1 ) , yielding a total q-range from 0 . 005 to 0 . 30 Å−1 . Collection times did not exceed 4 h , at which point cells were determined to be better than 90% viable , as shown below in cell viability and sample integrity ( S6 and S7 Figs ) . The scattering data from the cells demonstrate no statistically significant change in scattered intensity after 4 h ( S7 Fig ) . Model lipid mixtures ( S9 Fig ) were produced using synthetic lipids from Avanti Polar Lipids ( Alabaster , AL ) and prepared as follows: lipids were dissolved in chloroform , dispersed as a film by evaporation in a 20-mL scintillation vial , and dried overnight under vacuum ( >6 h ) . The lipid films were then rehydrated in the isotopically appropriate solvent ( H2O , D2O , or a combination thereof ) and subjected to 5 freeze-thaw cycles prior to extrusion through 100-nm pore-diameter track-etched polycarbonate membrane filters . The final concentration of vesicles for scattering measurements was 10 mg/mL . Data used for the contrast experiments ( Figs 1 and 2 ) were collected on EQ-SANS for 0 . 009 Å−1 < q < 0 . 06 Å−1 . The data were evaluated as the Porod invariant: Q*=∫q2I ( q ) dq=2πI ( 0 ) Vp ( 2 ) The proportionality of Q* to the total scattering intensity I ( 0 ) makes it a useful metric for comparison to structure-independent estimates of I ( 0 ) based only on composition and scattering length density , ρ . Vp is the Porod volume . To analyze the acyl thickness of the bilayer , the data were modeled using the expression:[73] I ( q ) =NVVs2 ( ρm−ρs ) 2〈|F ( q ) |2〉 ( 3 ) for an arbitrary number of bilayers , N , of volume , Vs , with scattering length density ρ s , in a solvent of ρm , where F ( q ) is the form factor describing the lamellar shape , and ( ρm − ρs ) , is the contrast term . The scattering law used to model the data was the lamellar form factor representing the hydrocarbon core of the bilayer , with 2 identical head group regions on either side , with scattering densities matching that of the surrounding cellular environment . Fitting of the SANS data was performed in SASview . [74] Estimates of total scattering from B . subtilis ( Figs 1 and 2 ) were made as follows . A neutron scattering signal arises when there is a difference in neutron scattering length density , ρ , between a species , s , and the medium in which it resides in , m . The difference , ( ρm−ρs ) , is called contrast , and the scattered intensity is proportional to its square . For any species , ρ= ( Σinbi ) /V ( 4 ) where n is the number of atoms , b is the coherent neutron scattering length for each atom , and V is the volume of the species . Because each class of biomolecule ( i . e . , protein , lipid , carbohydrate , RNA , or DNA ) has a nearly constant chemical composition and density , ρ is well approximated as a single value for each class ( Table A in S1 Text ) . [30] Deuterium labeling increases the ρ of an unlabeled molecule ( ρH ) because the scattering length b is greater for D than H ( 6 . 67 versus –3 . 74 fm ) . [31] In biomolecules , the hydrogen atoms can be assigned to 2 categories , those bound to carbon ( CH ) , which do not exchange with water , and those bound to heteroatoms ( XH , X = N , O , or S ) , which do exchange with water . The fractions in the 2 categories are consistent within each class of biomolecule , which allows for the definition of the terms , ΔρCD and ΔρXD , that represent the increase in ρ resulting from the complete deuteration of each category . For a deuterated biomolecule: ρ=ρH+fXDΔρXD+fCDΔρCD ( 5 ) where ρH is ρ for the unlabeled ( all-H ) molecule , and fXD and fCD are the fractions of deuterium substitution in the XH and CH categories , respectively . Estimates ( Figs 1b and 2a ) were generated using Eq 5 and the values for ρH , ΔρXD , and ΔρCD in Table A in S1 Text . The differing slopes for each class of biomolecule reflect differences in ΔXD . The total scattering from a cell , Icell ( 0 ) , is the sum of the scattering contributions from all of the cell’s biomolecular species , given by the relation Icell ( 0 ) ∝Σsχs ( ρm−ρs ) 2 ( 6 ) where χs is the volume fraction of each species s , ρs is its neutron scattering length density , and ρm is the average neutron scattering length density of the medium ( all species s , including intracellular water ) . When ρs = ρm , the species is contrast-matched and thus effectively invisible to neutrons , provided the medium is uniform . The scattering attributable to an individual species j is given by: Ij ( 0 ) ∝0 . 5χjΣsχs ( ρj−ρs ) 2 ( 7 ) In an undeuterated cell , the χj for biomolecules is sufficiently high and the ρj is sufficiently different that the surrounding medium is not truly uniform . As a result , the total scattering for a given species j reflects interspecies contrast and is not completely nulled at the nominal contrast matchpoint defined by the average of the medium ( χj = χm ) . However , judicious deuteration can be used to converge ρ for water and most biomolecules ( cf . Figs 1b and 2a ) , effectively suppressing the interspecies contributions . Knowledge of the cell’s composition ( Table A in S1 Text ) allows one to estimate the total cellular scattering—broken down by biomolecular species as a function of deuteration in the solvent and in the CH skeletons of the various biomolecules ( Figs 1c and 2d ) . Approximately 80% of the dry mass of a B . subtilis cell is made up of protein ( 53% ) , RNA ( 18% ) , DNA ( 2 . 6% ) , lipid ( 5 . 2% ) , and carbohydrate ( 2 . 8% ) . [33] The remaining mass—which was neglected in our estimates—consists of small organic molecules , such as amino acids , cofactors , and nucleotides plus inorganic material . These data were used with Eqs 6 and 7 to calculate the predicted total scattering in Fig 1c . Estimated scattering for cells grown in D-M9 ( Fig 2b ) relied on analyses of lipids and proteins extracted from B . subtilis under relevant conditions to adjust ρ according to Eq 5 . The net deuteration of the FA pool in the absence of cerulenin and supplemental FAs was 68 . 5% , with deuteration of individual FAs ranging from 66% to 71% ( Table B in S1 Text ) . The deuteration of the amino acid pool was more variable and had a lower average deuteration of 60% ( Table C in S1 Text ) , which is similar to what would be expected in Escherichia coli . [75] A value of 70% was assumed for other biomolecules , and the extent of deuteration at water-exchangeable positions was assumed to match that of the medium . Cell viability was evaluated using optical density measurements , manual cell counts with a hemocytometer , and fluorescent live/dead staining with the BacLight Bacterial Viability Kit ( Catalog number L7012 , Molecular Probes , Eugene , OR ) . The live/dead assay uses a mixture of 2 fluorescent nucleic acid stains ( SYTO 9 and propidium iodide ) , which stain live cells green and dead cells with compromised membranes , red . Fluorescence micrographs of stained cells were acquired with a Zeiss Axioskop 2 Plus and analyzed with ImageJ[76] to count cells using green and red fluorescence channels . Nonirradiated suspensions of the ΔyusL strain prepared identically to the samples used for neutron beam experiments were used as controls . Cell suspensions were incubated in a sealed cuvette at 25°C , and optical density ( OD600 ) was recorded at 1-min intervals over a period of 24 h ( S6a Fig ) . Direct cell counts and live/dead staining of the nonirradiated samples were performed immediately after processing and at 4 h and 24 h ( S6b Fig ) . These experiments showed close correspondence among the 3 measures of viability and that cells remained 90% viable over 4 h and 50%–60% viable over the course of 24 h under the conditions of the experiment . Irradiated cell suspensions were taken from the neutron beam after data collection and allowed to decay for approximately 30 min , at which time they were subjected to a radiological survey . Radiological safety protocols do not permit timely removal of cell suspensions for microscopic examination . Instead , optical density measurements of the cells at 4 h and 24 h were carried out using a Shimadzu UV-2700 UV–Vis spectrophotometer ( S6a Fig ) . FA stability was monitored over the course of the 4 h it took to collect a complete SANS data set using GC/MS as described above for nonirradiated samples ( S7 Fig ) . Finally , cell stability in the neutron beam was assessed by repeating the SANS measurement 3 . 5 h after the cells were put in the beam showing no change in the scattered intensity ( S7 Fig ) .
|
The structure and organization of the cell membrane are central to many biological functions , and although they have been extensively studied , there is still much that we don’t understand . A wealth of detailed information has been obtained from studies of model lipid membranes . However , these systems are often highly simplified in composition and always lack the active processes found in cells . The major obstacle to studying membranes in vivo in living cells has been that high-resolution physical methods used to investigate membrane structure are not compatible with living organisms . To overcome this obstacle , we employed a new in vivo approach that allows membrane structure to be observed directly in the gram-positive bacterium B . subtilis . This approach relies on tuning the isotopic content of hydrogen within the membrane , and other parts of the bacterium , to generate neutron scattering spectra exclusively from the membrane . Using this approach , we confirmed that the structure of the B . subtilis membrane is lamellar and has an average hydrophobic thickness of 23 . 9 ± 0 . 9 Ångstroms ( Å ) . We were also able to observe nanoscopic lateral membrane structures , consistent with the notion of lipid rafts . This experimental approach will allow for a wide range of future structural studies of the cell membrane ( and possibly other classes of biomolecules ) without the need for extrinsic probes or labels . In this way , it fundamentally changes the scope of nanoscale structural questions that can be addressed in vivo .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"neutron",
"scattering",
"pathology",
"and",
"laboratory",
"medicine",
"particle",
"physics",
"pathogens",
"bacillus",
"microbiology",
"prokaryotic",
"models",
"experimental",
"organism",
"systems",
"lipid",
"structure",
"molecular",
"biology",
"techniques",
"cellular",
"structures",
"and",
"organelles",
"bacteria",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"amino",
"acid",
"analysis",
"lipids",
"medical",
"microbiology",
"microbial",
"pathogens",
"chemistry",
"nucleons",
"molecular",
"biology",
"cell",
"membranes",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"physics",
"biochemistry",
"hydrogen",
"chemical",
"elements",
"scattering",
"cell",
"biology",
"nuclear",
"physics",
"bacillus",
"subtilis",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"organisms"
] |
2017
|
The in vivo structure of biological membranes and evidence for lipid domains
|
CD8+ T cells play an important role in controlling of HIV and SIV infections . However , these cells are largely excluded from B cell follicles where HIV and SIV producing cells concentrate during chronic infection . It is not known , however , if antigen-specific CD8+ T cells are excluded gradually as pathogenesis progresses from early to chronic phase , or this phenomenon occurs from the beginning infection . In this study we determined that SIV-specific CD8+ T cells were largely excluded from follicles during early infection , we also found that within follicles , they were entirely absent in 60% of the germinal centers ( GCs ) examined . Furthermore , levels of SIV-specific CD8+ T cells in follicular but not extrafollicular areas significantly correlated inversely with levels of viral RNA+ cells . In addition , subsets of follicular SIV-specific CD8+ T cells were activated and proliferating and expressed the cytolytic protein perforin . These studies suggest that a paucity of SIV-specific CD8+ T cells in follicles and complete absence within GCs during early infection may set the stage for the establishment of persistent chronic infection .
Most human immunodeficiency virus ( HIV ) -infected individuals fail to adequately control persistent high-level viral replication that results in gradual loss of CD4 T cells and ultimately AIDS in the absence of antiretroviral therapy ( ART ) . B cell follicles in secondary lymphoid tissues have been identified as important sanctuaries that contain large amounts of virus-producing cells during chronic HIV and simian immunodeficiency virus ( SIV ) infection [1–5] . CD4+ T follicular helper ( TFH ) cells , a population that mainly resides in B cell follicles , serve as a major site of productive HIV and SIV replication during the chronic phase of infection [1 , 2 , 4 , 6–8] . In SIV-infected rhesus macaques that control viral replication , either via a natural highly effective immune response or receiving long-term , fully suppressive ART , residual productive SIV infection is strikingly restricted to TFH cells [9] . In HIV infected aviremic individuals treated with long-term ART , TFH also serves as a major reservoir for active and persistent virus transcription [10] . Therefore , understanding the immune activity needed to kill virus-infected TFH cells in B cell follicles is necessary for developing novel therapies to fully eradicate HIV or SIV infection . Antigen-specific CD8+ T cells have a key role in controlling HIV and SIV infections . Their emergence during the acute phase of infection is associated with a decline in plasma viremia [11–13] . Moreover , the transient depletion of CD8+ T cells during SIV or SHIV infections induces high levels of plasma viremia which are reduced upon reconstitution of CD8+ lymphocytes [14–16] . Strong HIV-specific CD8+ T cell activity is directly associated with long-term elite control of infection [17–19] . Furthermore , we previously showed a significant inverse relationship between SIV-specific CD8+ T cell frequency and SIV-producing cell levels in lymphoid compartments during chronic SIV infection [3] . However , in spite of the notable anti-viral effect , HIV- and SIV-specific CD8+ T cells fail to fully eliminate viral replication and the vast majority of HIV and SIV-infected individuals eventually develop disease in the absence of ART . We and others previously showed that HIV- and SIV-specific CD8+ T cells are largely excluded from B cell follicles in lymph node and spleen tissues during chronic infection [2 , 3 , 20 , 21] . The paucity of virus-specific CD8+ T cells inside B cell follicles , where HIV- and SIV-producing cells are highly concentrated , creates an immune privileged site and an important mechanism of immune evasion by HIV and SIV . This mechanism may , at least partially , account for the failure of CD8+ T cells to fully eradicate HIV and SIV infections . The exclusion of anti-viral CD8+ T cells from B cell follicles during chronic infection is not absolute . Studies indicate that there are populations of functional CD8+ T cells expressing CXCR5 in B cell follicles in chronic LCMV , HIV and SIV infections [20 , 22 , 23] , and levels of follicular virus-specific CD8+ T cells correlate with reductions of plasma viral loads and tissue viral replication [3 , 20 , 24 , 25] . Thus , while typically relatively low in numbers , virus-specific CD8+ T cells in follicles appear capable of suppressing viral replication . Because HIV and SIV replication is concentrated within lymphoid follicles during chronic infection , studies of the location , abundance , and phenotype of follicular SIV-specific CD8 T cells during early stages of infection are warranted . Whether virus-specific CD8+ T cells migrate into B cell follicles during early HIV and SIV infections remains to be determined . Our hypothesis is that a paucity of SIV-specific T cells in lymphoid follicles contributes to the establishment of the follicular reservoir of SIV during early SIV infection . To test this hypothesis , in this study , we determined the abundance , distribution and phenotype of SIV-specific T cells in lymph nodes from a cohort of SIV infected rhesus macaques during the early stages of infection .
To determine whether SIV-specific CD8+ T cells accumulate within B cell follicles during early infection , we evaluated the distribution and quantity of SIV-specific CD8+ T cells at 14 ( n = 7 ) and 21 ( n = 9 ) days post-infection ( dpi ) in the lymph nodes from SIV-infected rhesus macaques . The age , viral loads , and CD4 T cell levels of the study animals are indicated in Table 1 . All of the animals in this study expressed the MHC-I allele Mamu-A1*001 , and immunodominant Gag specific CD8+ T cells were identified using in situ tetramer staining with Mamu-A1*001 tetramers loaded with Gag CM9 peptides . In combination with MHC tetramer staining of SIV-specific CD8+ T cells , antibodies against IgM were used to label B cell follicles cells in situ in lymph node sections . We detected tetramer+ SIV-specific CD8+ T cells in B cell follicles at both 14 and 21 days post-infection ( Fig 1A ) . Similar to chronic infection [3] , the number of SIV-specific tetramer+ CD8+ T cells/mm2 inside B cell follicles was significantly lower than in extrafollicular regions at 14 dpi ( Fig 1B ) and at 21 dpi ( Fig 1C ) . However , the extent of this difference was greater during chronic infection which showed a median of 4X lower cells/mm2 in follicles [3] , compared to a median of 2X lower found here at 14 and 21 dpi . We previously showed that there is a positive correlation between numbers of follicular and extrafollicular SIV-specific tetramer+ CD8+ T cells/mm2 [3 , 20] . In this study , we investigated the relationship between levels of SIV-specific tetramer+ CD8+ T cells inside and outside B cell follicles during early infection at 21 dpi . A highly significant positive correlation between levels of follicular and extrafollicular SIV-specific tetramer+ CD8+ T cells in lymph node was observed on 21 dpi ( Fig 1D ) . A similar correlation was seen at 14 . These data demonstrated that a small number of SIV-specific CD8+ T cells enter the B cell follicles during early SIV infection and that levels of follicular SIV-specific CD8+ T cells positively correlate to levels of SIV-specific CD8+ T cells in extrafollicular regions . We next analyzed whether SIV-specific tetramer+ CD8+ T cells accumulate within the germinal center ( GC ) area of the follicles , where follicular dendritic cells ( FDCs ) retain infectious virus trapped in immune complexes [26] . To address this question , we used Gag CM9 tetramers to detect SIV-specific CD8+ T cells , IgM-specific antibodies to identify B cell follicles and Ki67-specific antibodies to identify GCs . At 14 dpi , very few lymph node follicles with GCs were detected . At 21 dpi follicles with GC were abundant . Strikingly , at 21 dpi , 60 . 3% ( 44/73 ) of the GCs were completely devoid of SIV-specific tetramer+ CD8+ T cells ( Fig 2A ) . Furthermore , quantitative analysis showed that the frequency of SIV-specific tetramer+ CD8+ T cells in GCs was significantly lower than in non-GC follicular areas ( Fig 2B ) . In addition , we had longitudinal data from two of the study animals . Both showed increases in SIV-specific tetramer+ cells from 21 to 42 dpi in GC ( rh2516 from 42 to 98 , and rh2520 from 13 to 87 cells/mm2 ) . PD-1 is a marker of functional exhaustion [27 , 28] as well as a marker of CD8+ T cells that have recently been exposed to antigenic stimulation [29] . PD-1 is markedly upregulated on the surface of dysfunctional virus-specific CD8+ T cells during chronic HIV and SIV infections [30 , 31] , and blockade of PD-1 in vivo enhanced SIV-specific CD8+ T cells responses [31] . Moreover , recent studies found that high percentages of follicular CD8+ T cells in chronic HIV and SIV infection express inhibitory molecule PD-1 [20 , 32] . However , the degree to which follicular and extrafollicular SIV-specific CD8+ T cells in early infection express PD-1 has not yet been investigated . To determine the percentage of SIV-specific CD8+ T cells expressing PD-1 in follicular and extrafollicular compartments , we stained lymph node tissue sections from SIV infected rhesus macaques with MHC-class I tetramers , antibodies directed against PD-1 , and antibodies directed against CD20 to label B cell follicles . We found PD-1+ SIV-specific tetramer+ CD8+ T cells in both follicular and extrafollicular areas in early SIV infection at 21 dpi ( Fig 3A ) . Quantitative analysis showed a broad range of both follicular ( median 62%; range 12–67% ) and extrafollicular ( median 65%; range 8–68% ) SIV-specific tetramer+ CD8+ T cells expressing PD-1 . No significant difference was observed between the percentage of PD-1+ SIV-specific tetramer+ CD8+ T cells inside and outside B cell follicles ( Fig 3B ) . We further compared the percentage of SIV-specific tetramer+ CD8+ T cells that express PD-1 between early and our previously published data [20] from chronic SIV infection in follicular and extrafollicular regions respectively . Again , no significant differences were observed ( Fig 3C and 3D ) . Regulatory T cells ( Tregs ) play a crucial role in maintaining immunological self-tolerance and controlling autoimmune diseases [33] . They also are involved in suppressing immune activation in viral infection [34 , 35] . A large proportion of Tregs are characterized by the expression of the transcription factor Foxp3 [36 , 37] . While most Tregs are CD4+ , there exist a small population of CD8+ Tregs [38 , 39] . Directed contact is an important mechanism mediates suppression by Tregs [40] . We next investigated whether Foxp3+ cells were in contact with and potentially inhibiting the function of follicular SIV-specific CD8+ T cells in early infection . We stained lymph node tissue sections from rhesus macaques in early SIV infection , at 21 dpi , with MHC-class I tetramers to label SIV-specific CD8+ T cells , anti-Foxp3 antibodies to label Foxp3+ Tregs , and anti-IgM antibodies to label B cell follicles . We found follicular SIV-specific tetramer+ CD8+ T cells that directly contact Foxp3+ cells ( Fig 4A ) and follicular SIV-specific tetramer+ CD8+ T cells that expressed Foxp3+ ( Fig 4B ) . An average of 12 . 4% ( range 7–20% ) follicular SIV-specific tetramer+ CD8+ T cells were in direct contact with Foxp3+ cells and the corresponding number in extrafollicular was 18 . 6% ( rang 9–26% ) ( Fig 4C ) . No significant difference was shown between the percentage of SIV-specific tetramer+ CD8+ T cells that contact Foxp3+ cells inside and outside B cell follicles in early SIV infection ( Fig 4C ) . The percentages of Foxp3+ SIV-specific tetramer+ CD8+ T cells inside and outside B cell follicles also showed no significant difference ( Fig 4D ) . Similar to SIV-specific CD8+ T cells , Foxp3+ cells levels were also significantly lower in follicular than extrafollicular regions ( Fig 4E and 4F ) . In addition , there was no significant difference between the ratios of Tet+ cells: Foxp3+ cells in follicular and extrafollicular regions . These findings suggest that contact mediated suppression of Foxp3+ Tregs on SIV-specific CD8+ T cells is similar in follicular and extrafollicular regions in early SIV infection . We next evaluated whether the relationship of Foxp3+ Tregs and SIV-specific CD8+ T cells changed from early stages of infection to chronic infection . Interestingly , we first found significantly higher levels of Foxp3+ cells in early infection compared to chronic infection [20] in follicular areas ( Fig 5A ) , but not in extrafollicular areas ( Fig 5B ) . Second , the percentage of follicular SIV-specific tetramer+ CD8+ T cells in contact with Foxp3+ cells in early infection was significantly higher than chronic infection ( Fig 5C ) , and this difference was not observed in extrafollicular SIV-specific CD8+ T cells ( Fig 5D ) . Third , the percentage of SIV-specific tetramer+ CD8+ T cells in follicular areas that expressed Foxp3 was significantly higher during early infection than chronic infection in both follicular ( Fig 5E ) and extrafollicular areas ( Fig 5F ) . Lastly , the ratios of Tet+ cells: Foxp3+ cells in both follicular and extrafollicular regions were significantly lower in early infection compared to chronic infection ( Fig 5G and 5H ) . We next assessed the levels SIV-specific CD8+ T cells that were activated and proliferating in follicular and extrafollicular regions of lymph nodes during early SIV infection , at 21 dpi . Ki67 is an activation and proliferation marker of T cells [41 , 42] . Ki67+ SIV-specific tetramer+ CD8+ T cells were found in follicular as well as in extrafollicular regions ( Fig 6A ) . Quantitative image analysis showed that the percentage of follicular Ki67+ SIV-specific tetramer+ CD8+ T cells was significantly lower than in extrafollicular counterparts ( Fig 6B ) . An average of 40% ( range 7–61% ) of SIV-specific tetramer+ CD8+ T cells were Ki67+ in follicular area compared to 54% ( range 19–76% ) in extrafollicular areas . To evaluate whether there was a change in Ki67 expressing virus-specific CD8+ T cells during early compared to chronic infection , we compared these findings to our previously published data from the chronic stage of infection [20] . We found that significantly fewer SIV-specific tetramer+ CD8+ T cells in both follicular and extrafollicular regions during chronic SIV infection expressed Ki67 ( Fig 6C and 6D ) . Perforin is a crucial factor for cytolytic function in virus-specific CD8+ T cells . We previously showed that during chronic infection approximately 35% of follicular SIV-specific CD8+ T cells express perforin and that most expressed another cytolytic effector molecule , granzyme B , which typically works in concert with perforin to lyse infected cells [3 , 20] . Here we determined the expression of perforin in SIV-specific CD8+ T cells during early infection . Perforin expression levels within SIV-specific CD8+ T cells were divided into four categories ( negative , low , medium , and high ) ( Fig 7A ) . We found that most follicular SIV-specific tetramer+ CD8+ T cells expressed perforin ( mean 73%; range 35–94% ) . Similar levels of low , medium and high perforin expressing SIV-specific tetramer+ CD8+ T cells were observed in follicular and extrafollicular areas ( Fig 7B ) . Comparison of perforin expression in SIV-specific tetramer+ CD8+ T cells in early and chronic infection [20] , showed significantly higher levels of perforin+ SIV-specific tetramer+ CD8+ T cells in early SIV infection in both follicular ( Fig 7C ) and extrafollicular regions ( Fig 7D ) . In particular , the percentage of perforinhigh SIV-specific tetramer+ CD8+ T cells were higher during early compared to chronic stages of infection ( Fig 7E and 7F ) . A hypothesis yet to be tested regarding the relative low abundance of virus-specific CD8+ T cells in follicles is that follicular virus-specific CD8+ T cells die via apoptosis at greater rates than extrafollicular virus-specific CD8+ T cells . Here we tested that hypothesis and determined the levels of SIV-specific CD8+ T cells in follicular and extrafollicular levels that expressed Poly ( ADP-ribose ) polymerase ( PARP ) in early chronic SIV infection ( 50–60 dpi ) . PARP is involved in cells death by promoting release of apoptosis-inducing factor ( AIF ) , and can be used as a marker of cells undergoing apoptosis [43] . We found small subsets of SIV-specific tetramer+ CD8+ T cells that expressed PARP in both follicular as well as extrafollicular areas ( Fig 8A ) . Quantitative analysis showed no significant difference between the levels of PARP+ SIV-specific tetramer+ CD8+ T cells inside and outside B cell follicles ( Fig 8B ) . In order to gain evidence as to whether levels of SIV-specific CD8+ T cells were impacting viral replication in follicular and extrafollicular compartments at 21 dpi , we compared levels of SIV-specific tetramer+ CD8+ T cells and SIV RNA+ cells in follicular and extrafollicular regions of lymph nodes . In situ hybridization was used to localize and quantify SIV RNA+ cells and in situ tetramer staining used to localize and quantify SIV-specific CD8 T cells . We detected a median of 2 . 2 ( range 0 . 8–7 . 2 ) viral RNA+ cells/mm2 in follicles , and a median of 1 . 5 ( range 1 . 0–3 . 3 ) in extrafollicular areas ( Table 1 ) . We found that SIV-specific tetramer+ CD8+ T cells were significantly inversely correlated with SIV RNA+ cells in follicular regions ( Fig 9A ) , but not in extrafollicular regions ( Fig 9B ) . We also determined the in vivo effector tetramer+ SIV-specific CD8 T cell to target SIV vRNA+ target cell ratios ( E:T ) . We found that follicles showed a median E:T of 20 ( range 7–82 ) and in extrafollicular areas a median E:T of 31 ( range 15–114 ) ( Table 1 ) . No apparent association was detected between levels of levels of SIV-specific CD8 T cells and the age of animals , viral loads , or CD4 T cell levels . However , we can’t rule out that an association might be detected with a greater sample size .
Eradication of HIV-infected cells in vivo remains a critical obstacle to cure HIV infection . B cell follicles are major anatomical reservoirs that permit active HIV and SIV replication during chronic infection [2 , 3 , 9] . However , virus-specific CD8+ T cells generally fail to accumulate in high frequency in B cell follicles [2 , 3 , 20 , 21] . The ongoing viral replication in B cell follicles during chronic infection is likely due , at least partially , to the paucity of follicular anti-viral CD8+ T cell responses [1–3 , 9 , 44] . It is not known whether this phenomenon also occurs during early stages of infection . Here , we determined the location , abundance , and phenotype of SIV-specific CD8+ T cells in follicular and extrafollicular regions during early SIV infection . We found that few follicles had GCs at 14 dpi and many follicles had GCs at 21 dpi , suggesting that SIV infection induced the formation of GCs as part of the immune response to combat SIV infection . At 21 dpi , SIV-specific CD8+ T cells were largely excluded from B cell follicles , similar to what we reported during chronic disease [3] . Strikingly , and in contrast to chronic infection , we also discovered that follicular SIV-specific CD8+ T cells were entirely absent in 60% of GC areas examined . A complete absence of SIV-specific CD8+ T cells in most GCs at 21 dpi whilst during chronic infection there are equal numbers of SIV-specific CD8+ T cells in GC and non-GC follicular areas [20] suggests that during early infection , there is a temporal delay from the formation of nascent follicles and the entry of SIV-specific CD8+ T cells . Importantly , by 21 days dpi , the time in which we found most GC areas examined to be entirely devoid of SIV-specific CD8+ T cells , SIV has already seeded FDCs in GCs [45] . These findings indicate that the arrival of SIV-specific CD8+ T cells in GCs occurs after SIV has already seeded FDCs and infected follicular CD4 T cells . Thus , low levels or a complete absence of SIV-specific CD8 T cells in nascent GCs during early infection likely contributes mechanistically to the ability of SIV to become concentrated in lymphoid follicles during chronic stages of infection , and sets the stage for the establishment of persistent chronic infection . As mentioned above , during chronic HIV and SIV infections , viral RNA+ cells are concentrated in lymphoid follicles . Interestingly , during early infection at 14 dpi , SIV RNA+ cells are in similar concentrations in follicular and extrafollicular areas [3] . The adaptive virus-specific CTL response becomes detectible during the second week post-infection and by the third and fourth weeks post-infection most SIV RNA+ cells in non-lymphoid tissues are cleared [45 , 46] , and as we found here , at 21 dpi SIV RNA+ cells begin to become concentrated within lymphoid follicles . In addition , at 21 dpi , we found that the in vivo E:T ratios in follicles showed a median of 20 ( range 7–82 ) and in extrafollicular areas a median of 31 ( range 15–114 ) . This disparity in E:T ratios in follicles and extrafollicular regions becomes greater during chronic infection . During chronic stages of infection , the in vivo effector to target ratios average less than 10 in follicles and over 100 in extrafollicular areas . These findings suggest that over time , follicular virus-specific CD8 T cells become increasingly less able to control viral replication , whereas control of viral replication is fairly well maintained in extrafollicular areas . In addition to determining the quantity of SIV-specific CD8+ T cells in follicular and extrafollicular compartment of lymph nodes during early infection , we also determined the phenotype of these cells . We found a broad range of SIV-specific CD8+ T cells expressing PD-1 during early infection . The PD-1 expression on SIV-specific CD8+ T cells during early infection likely reflects cells that were recently were exposed to antigenic stimulation , however , exhaustion cannot be ruled out . We found subsets of SIV-specific CD8+ T cells were in direct contact with FoxP3+ cells and some SIV-specific CD8+ T cells were FoxP3+ . We detected an increase in FoxP3+ cells in follicles and increased contact of tetramer+ SIV-specific CD8 T cells during early compared to chronic infection . These data suggest that the function of some follicular as well as extrafollicular SIV-specific CD8+ T cells may be inhibited by Foxp3+ cells during early SIV infection , and that this inhibition in follicles may be greater during early than chronic stages of infection . However , since FoxP3 is transiently induced on T cells upon activation in vitro [47] , additional studies are needed to confirm inhibition . We found subsets of SIV-specific CD8+ T cells expressing the activation and proliferation marker Ki67 in both follicular as well as extrafollicular areas , with higher levels in extrafollicular areas . These data indicate that during early infection , virus-specific CD8+ T cells were activated and proliferating in both follicular as well as extrafollicular compartments; and that most activated and proliferating cells were located in the extrafollicular areas . Levels of SIV-specific CD8+ T cells expressing Ki67 were higher in early infection compared to those we found during chronic stages of infection , and may be reflective of the reduction of virus and antigen for the SIV-specific CD8+ T cells during the transition of early to chronic stages of infection . Most SIV-specific CD8+ T cells expressed perforin in both follicular as well as extrafollicular areas of lymph nodes during early infection indicating that these subsets were likely were capable of immediately killing an SIV infected cell upon contact . We found similar levels of tetramer+ virus-specific CD8+ T cells expressing perforin low , medium , and high levels of perforin or not expressing perforin in follicular and extrafollicular areas . With perforin expression levels being a marker of effector and memory subsets , this suggests that levels of effector and memory subsets are maintained at similar levels in follicular and extrafollicular compartments . Decreases in tetramer+ SIV specific T cells expressing high levels of perforin from early to chronic infection likely reflected the transition from the effector dominant CD8+ T cell response transitioning into a primarily memory pool of cells . In this study , we also found similar levels of SIV-specific CD8+ T cells in follicular and extrafollicular areas of lymph nodes expressing the apoptosis marker PARP . These findings suggest that the relative low abundance of virus-specific CD8+ T cells in follicular relative to extrafollicular regions of lymph nodes is not likely due to follicular virus-specific CD8+ T cells dying via apoptosis at increased rates compared to extrafollicular virus-specific CD8+ T cells . This study was done with Mamu-A1*001 animals because in these animals CM9/Gag is a consistent immunodominant SIV epitope that CD8 T cells respond to during acute infection and SIV does not mutate and escape CM9/Gag responses during chronic infection [48] . While this study was limited to the examination of Mamu-A01/Gag-specific CD8 T cells , we would expect to see similar findings with other virus-specific CD8 T cells . We have examined animals with other MHC genotypes including Mamu-A1*002A and Mamu-B*008:01 in previous studies of chronic SIV infection [3 , 20] , as well as HIV infected individuals with different MHC genotypes [2] . We found in every case , similar patterns of localization of virus-specific CD8 T cells . We also consistently find that the location and relative abundance of tetramer+ SIV-specific CD8 T cells is similar to that of total CD8 T cells [2 , 3 , 20 , 25 , 46 , 48–53] . Importantly , during early SIV infection , levels of follicular virus-specific CD8+ T cells inversely correlated with levels of follicular SIV replication ( SIV RNA+ cells ) . These findings suggest that follicular virus-specific CD8+ T cells are not only important in controlling SIV replication during early infection , but also that numbers matter . These findings are consistent with our previous studies that similarly showed that higher levels of virus-specific CD8 T cells were associated with lower levels of viral RNA+ cells [3 , 20 , 24] , indicating that levels of virus-specific CD8 T cells influence viral clearance . While no significant correlation of tetramer+ SIV-specific CD8 T cells and viral RNA+ cells was detected in the extrafollicular areas , the sample size was small . A greater sample size might reveal a correlation in the extrafollicular areas . Notably , the concentration of SIV-specific CD8 T cells were significantly higher in extrafollicular areas compared to follicular areas , and the median and range of viral RNA+ cells were lower compared to follicular areas , indicating that the overall control of viral replication was a bit better in the extrafollicular areas .
Lymph nodes were obtained from captive-bred rhesus macaques of Indian origin infected with SIVmac239 . Animals rh2515 , rh2516 and rh2520 were infected with 1 , 000 TCID50 SIVmac239 intravenously , and animals rh2583 , rh2584 , rh2578 , rh2579 , rh2587 and rh2588 were infected with 500 TCID50 SIVmac239 intravenously . Portions of fresh lymphoid tissues were immediately snap frozen in OCT and/or formalin fixed and embedded in paraffin . Simultaneously , portions of fresh lymphoid tissues were also collected in RPMI 1640 medium with sodium heparin ( 18 . 7 U/ml ) and shipped overnight to the University of Minnesota for in situ tetramer staining . In situ tetramer staining combined with immunohistochemistry was performed on fresh lymph tissue specimens shipped overnight , sectioned with a compresstome [54] and stained essentially as previously described [51 , 55] . Biotinylated MHC-class I monomers were loaded with peptides ( National Institute of Health Tetramer Core Facility , Emory University , Atlanta GA ) and converted to MHC-class I tetramers . Mamu-A1*001 molecules loaded with SIV Gag CM9 ( CTPYDINQM ) peptides [56] or irrelevant negative control peptides FV10 ( FLPSDYFPSV ) from the hepatitis B virus core protein . Fresh lymph node sections were incubated with MHC-class I tetramers ( 0 . 5 μg/ml ) alone or along with goat-anti-human PD-1 Abs ( 1 μg/mL , polyclonal , R&D Systems ) . For secondary incubations , sections were incubated with rabbit-anti-FITC Abs ( 0 . 5 μg/mL , BioDesign , Saco , ME ) and mouse-anti-human Ki67 Abs ( 1:500 dilution , clone MM1 , Vector ) , or mouse-anti-human perforin Abs ( 0 . 1 μg/mL , clone 5B10 , Novacastra ) , or mouse-anti-human Foxp3 Abs ( 2 . 5 μg/mL , clone 206D , BioLegend ) or mouse-anti-human CD20 Abs ( 0 . 19 μg/mL , clone L26 , Novocastra ) , or mouse-anti-human PARP Abs ( 3 . 5 μg/mL , clone Asp214 , Cell signaling ) . For the tertiary incubations , the sections stained with goat-anti-human PD-1 Abs were incubated with Cy3-conjugated donkey-anti-rabbit Abs ( 0 . 3 μg/mL , Jackson ImmunoResearch Laboratories , West Grove , PA ) , Alexa 488-conjugated donkey-anti-goat Abs ( 0 . 75 μg/mL , Jackson ImmunoResearch Laboratories ) , and Cy5-conjugated donkey-anti-mouse Abs ( 0 . 3 μg/mL , Jackson ImmunoResearch Laboratories ) . All other sections were incubated with Cy3-conjugated goat-anti-rabbit Abs ( 0 . 3 μg/mL , Jackson ImmunoResearch Laboratories ) , Alexa 488-conjugated goat-anti-mouse Abs ( 0 . 75 μg/mL , Molecular probes ) , and Dylight 649-conjugated goat anti-human IgM ( 0 . 3 μg/mL , Jackson ImmunoResearch Laboratories ) . Sections were imaged using a Zeiss LSM 800 confocal microscope . Montage images of multiple 512 × 512 pixels were created and used for analysis . For the determination of levels of SIV-specific CD8+ T cells and percentages of SIV-specific CD8+ T cells that co-expressed specific molecules , follicular areas were identified morphologically as clusters of brightly stained closely aggregated IgM+ or CD20+ cells . Follicular and extrafollicular areas were delineated using ImageJ software . Areas that showed loosely aggregated B cells that were ambiguous as to whether the area was a follicle were not included . For PD-1 expression analysis , an average of 112 tetramer+ cells ( range , 67–190 ) was analyzed in follicular regions and 213 ( range , 117–272 ) in extrafollicular regions . For quantification of tetramer+ cells that were in contact with Foxp3+ cells and express Foxp3+ , an average of 102 tetramer+ cells ( range , 57–193 ) was analyzed in follicular regions and 298 ( range , 168–560 ) in extrafollicular regions . For Ki67 expression analysis , an average of 133 tetramer+ cells ( range , 30–246 ) was analyzed in follicular regions and 307 ( range , 130–464 ) in extrafollicular regions . For perforin expression level analysis , an average of 97 tetramer+ cells ( range , 22–193 ) was analyzed in follicular region and 276 ( range , 98–530 ) in extrafollicular region . To determine levels of perforin expression , tetramer+ cells were scored using the following objective criteria . Tetramer+ cells with no detectable perforin staining above background levels were scored as perforin negative . Tetramer+ cells with perforin staining 2-3X greater than background were scored as perforin low , with perforin staining 4-9X higher than background as perforin medium , and those with 10X or greater than background levels and with perforin staining detectable throughout much of the cytoplasm were scored as perforin high . Cell counts were done on single z-scans . While doing the cells counts , we demarcated cells using a software tool to avoid counting the same cell twice . All quantitative image analyses were done with lymph node tissues . An average of 7 . 42 mm2 ( range , 5 . 63–10 . 08 mm2 ) was analyzed for each lymph node . In situ hybridization for SIV RNA was performed as previously described [3] . This technique identifies cells that are actively transcribing SIV , but not extracellular virions encapsulated in envelope glycoprotein and bound to FDC . Briefly , 6 μm frozen sections were fixed in 4% paraformaldehyde ( Sigma-Aldrich , St . Louis , MO ) , hybridized overnight with digoxygenin labeled SIVmac239 antisense probes ( Lofstrand Labs , Gaithersburg , MD ) and visualized using NBT/5-bromo-4-chloro-3-indolyl phosphate ( Roche , Nutley , NJ ) . Immunohistochemistry staining for B cells was performed in the same tissues using mouse-anti-human CD20 ( clone 7D1; AbD Serotec , Raleigh , NC ) and detected using HRP-labeled polymer anti-mouse IgG ( ImmPressKit; Vector Laboratories , Burlingame , CA ) and Vector NovaRed substrate ( Vector Laboratories ) . SIV RNA+ cells were counted by visual inspection and classified as either inside or outside of B cell follicles which were identified morphologically as a cluster of CD20+ cells as previously described [55] . Total tissue area and area of follicles was determined by quantitative image analysis ( Qwin Pro version 3 . 4 . 0; Leica , Cambridge , U . K . ) and used to calculate the frequency of SIV+ cells per mm2 . An average of 12 . 5 mm2 ( 7 . 1 mm2–87 . 2 mm2 ) was analyzed . To compare differences in cells/mm2 , values were log-transformed and either paired t-tests or independent t-tests with unequal variance were performed , as appropriate; for reporting , estimates and 95% confidence intervals were back-transformed and reported as percent differences . For relationships between two measures of cells/mm2 , values were also log-transformed and Pearson's correlation and linear regressions were performed . To compare differences in percentages , paired t-tests or Wilcoxon rank-sum tests were performed , as appropriate . All tests were two-sided and p<0 . 05 was considered statistically significant . All calculations performed in R 3 . 4 . 3 [57] . All rhesus macaques used in this study were cared for by the staff at the Wisconsin National Primate Research Center ( WNPRC ) in accordance with the principles described in the National Research Council's Guide for the Care and Use of Laboratory Animals and with the recommendations of the Weatherall report ( https://royalsociety . org/topics-policy/publications/2006/weatherall-report/ ) . This study was approved by the University of Wisconsin-Madison Graduate School Institutional Animal Care and Use Committee ( Animal Care and Use Protocol Number G005529 ) . The trained employees of the Animal Care Unit of the Wisconsin Primate Research Center provided daily care for the animals included in these studies . The Wisconsin Primate Research Center is AAALAC certified . Animals are cared for by a staff of veterinarians under the direction of Saverio Capuano , DVM who is the Attending Veterinarian , four staff veterinarians and trained vet techs . At least twice daily animals were evaluated for signs of pain , illness and stress observing appetite , stool , typical behavior etc . by the staff of Animal Care Unit . If any of the above parameters were found unacceptable a WPRC veterinarian was contacted and appropriate treatments started . After a procedure , the affected site was observed for potential complications 24 hours later , or as directed by the veterinarian . An animal was euthanized at the recommendation of the attending WPRC veterinarian if any of the preceding conditions were observed and the veterinarians deem euthanasia necessary . Any deteriorating condition deemed particularly distressful to the animal ( as assessed by research and/or veterinary staff ) will be a condition for euthanasia . Animals were treated for pain and discomfort for the effects of SIV infection on the advice of the veterinarian . Blood draws , experimental SIV infections , and biopsies were done under anesthesia . Animals were anesthetized using ketamine ( up to 15 mg/kg i . m . ) to be reversed at the conclusion of a procedure by up to 0 . 25 mg/kg atipamizole ( i . v . or i . m . ) . Monitoring of anesthesia recovery was documented every 15 minutes until the animal is sitting upright , then every 30 minutes until the animal was fully recovered from the anesthesia . Animals were monitored for disease progression . Disease progression is variable , thus we monitored and used multiple factors in the consideration of euthanasia . These factors include: weight loss of 20% of total body weight ( to determine weight loss , animals were weighed at least every 90 days , according to Primate Center SOP 3 . 01 , and more frequently if deemed necessary by research or veterinary staff ) , infection with an opportunistic pathogen and no response to treatment , progressive decline in condition regardless of treatment , chronic diarrhea and anorexia , neurological signs , such as disorientation , abnormal gait or posture , tremor etc . If any of the preceding conditions are observed and the veterinarians deemed it necessary , the animal was euthanized . Any deteriorating condition deemed particularly distressful to the animal ( as assessed by research and/or veterinary staff ) was a condition for euthanasia . At the end of the study , or anytime during the study if recommended by a Center veterinarian , an animal was euthanized . Animals were euthanized by an IV overdose ( greater than or equal to 50 mg/kg or to effect ) of sodium pentobarbital or equivalent as approved by a clinical veterinarian , preceded by ketamine ( at least 15 mg/kg body weight , IM ) . It is possible that the final blood draw , performed following anesthesia but prior to the IV overdose , may result in death by exsanguination . All euthanasia was performed in accordance with the recommendations of the Panel on Euthanasia of the American Veterinary Medical Association . Death would be defined by stoppage of the heart as determined by a qualified and experienced person using a stethoscope to monitor heart sounds from the chest area , as well as all other vital signs that can be monitored by observation . Necropsy was performed by a qualified pathologist of the WNPRC .
|
A paucity of SIV-specific CD8+ T cells in lymphoid follicles and complete absence within most follicular germinal centers during early infection may set the stage for the establishment of persistent chronic infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"blood",
"cells",
"hiv",
"infections",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"microbiology",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"lymph",
"nodes",
"rna",
"viruses",
"lymphatic",
"system",
"cytotoxic",
"t",
"cells",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"infectious",
"diseases",
"staining",
"white",
"blood",
"cells",
"animal",
"cells",
"medical",
"microbiology",
"t",
"cells",
"microbial",
"pathogens",
"viral",
"replication",
"siv",
"antibody-producing",
"cells",
"cell",
"staining",
"cell",
"biology",
"b",
"cells",
"anatomy",
"virology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"viral",
"diseases",
"lentivirus",
"organisms"
] |
2019
|
Low levels of SIV-specific CD8+ T cells in germinal centers characterizes acute SIV infection
|
Genome-wide association studies ( GWAS ) have in recent years discovered thousands of associated markers for hundreds of phenotypes . However , associated loci often only explain a relatively small fraction of heritability and the link between association and causality has yet to be uncovered for most loci . Rare causal variants have been suggested as one scenario that may partially explain these shortcomings . Specifically , Dickson et al . recently reported simulations of rare causal variants that lead to association signals of common , tag single nucleotide polymorphisms , dubbed “synthetic associations” . However , an open question is what practical implications synthetic associations have for GWAS . Here , we explore the signatures exhibited by such “synthetic associations” and their implications based on patterns of genetic variation observed in human populations , thus accounting for human evolutionary history –a force disregarded in previous simulation studies . This is made possible by human population genetic data from HapMap 3 consisting of both resequencing and array-based genotyping data for the same set of individuals from multiple populations . We report that synthetic associations tend to be further away from the underlying risk alleles compared to “natural associations” ( i . e . associations due to underlying common causal variants ) , but to a much lesser extent than previously predicted , with both the age and the effect size of the risk allele playing a part in this phenomenon . We find that while a synthetic association has a lower probability of capturing causal variants within its linkage disequilibrium block , sequencing around the associated variant need not extend substantially to have a high probability of capturing at least one causal variant . We also show that the minor allele frequency of synthetic associations is lower than of natural associations for most , but not all , loci that we explored . Finally , we find the variance in associated allele frequency to be a potential indicator of synthetic associations .
Recent years have seen a plethora of genome-wide association studies ( GWAS ) finding thousands of common markers that are associated with hundreds of diseases and other traits . GWAS were initially founded on the Common Disease-Common Variant hypothesis [1]–[3] , which predicted that common complex diseases are most likely caused by a few common variants . As a consequence , the design of most GWAS consisted of genotyping common tag single nucleotide polymorphisms ( SNPs ) and comparing their allele frequencies between cases and controls . Some limitations of this design have been the topic of much recent discussion , with the gap between association and causality and the relatively small portion of heritable variation explained by associated markers drawing the most concern [4]–[7] . Several hypotheses aiming to explain the missing heritability have been proposed , including the roles of structural variants , gene-gene interactions , gene-environment interactions , epigenetics , and complex inheritance [4]–[7] . In addition , rare variants of relatively high penetrance contributing to disease risk [8] , [9] has also been suggested as a source of missing heritability since rare variants have not been directly observed in most GWAS , and they might be differently tagged by common markers [10]–[12] . Given this renewed interest in such variants , an investigation into their effect on GWAS association signals is warranted . A recent simulation-based study showed that rare causal variants can often create “synthetic associations , ” namely significant associations of common markers induced by the combined effect of one or more rare causal variants [13] . Dickson et al . further showed that a synthetically associated common marker could be substantially further away than expected had the underlying causal variant been common , and that synthetic associations are expected to be on average of lower minor allele frequency ( MAF ) than associations due to underlying common causal variants [13] . These predictions may partially explain why resequencing fine-mapping efforts , which are based on patterns of linkage disequilibrium ( LD ) of common variants , have often been unsuccessful in uncovering causal variants [10] , [13] , [14] . As the development of new methods and study designs for associating rare causal variants is underway [12] , [15]–[24] , the predictions of Dickson et al . are influencing the choice of study design , as well as the interpretation of traditional , genotyping-based GWAS ( e . g . [25] , [26] ) . A few instances of rare causal variants have already been well established [27]–[29] , including potentially causal rare variants in NOD2 that contribute to Crohn's disease risk [30]–[33] . In this example , since an associated common marker in the same gene is in LD with at least two of the rare variants , it is possible that they contribute to the marker's association signal [30] , thus inducing a synthetic association . As only a few examples of rare causal variants contributing to complex disease are well established , the jury is still out on their prevalence and on how often they lead to synthetic associations , with several recent studies arguing that the phenomenon is not necessarily widespread [34]–[36] . In light of this uncertainty , a detailed investigation of the signatures of synthetic associations and their implications is crucial for interpreting the results of genotyping-based GWAS and for considering the alternative of association studies based on whole-genome or whole-exome sequencing . Two of the key questions with regards to “synthetic associations” are ( 1 ) what are the implications for the resequencing distance for fine-mapping of significant associations ? and ( 2 ) how different is the MAF of synthetic associations from that of “natural associations” ( i . e . associations where the underlying causal variants are common ) ? While these questions have been addressed in studies of simulated data [13] , [36] , those simulations did not account for the nature of disease loci and risk variants , nor did they account for the specific nature of human genetic variation . In the former , it has been shown that the effect size and frequency of the disease variants can alter the power of the test [37] . While , in the latter , the mark left by human evolutionary history on patterns of genetic variation can greatly influence the nature of significant association signals , which we address in the present study . For example , when considering samples from European populations , which have been the populations of choice of most GWAS , it is crucial to account for their recent explosive population growth that has led to an inflation in the proportion of rare variants and to an altered haplotype and LD structure [38]–[41] , as well as to account for the well-established effects of the earlier Out-of-Africa event on these genetic patterns [42]–[48] . Here , we focus on the question of how empirical LD patterns can affect signals of “synthetic association” by investigating them in real human population genetic data . Through this , we aim to derive a better understanding of synthetic associations and their practical implications . Using empirical resequencing data , we randomly assume certain variants as increasing disease risk , determine cases and controls accordingly , and conduct an association study using genotyping data of the same individuals from arrays that have been employed in most GWAS . To illuminate and quantify signatures that are specific to “synthetic associations” , we repeat the process for rare and common causal variants and contrast the characteristics of synthetic associations with those of natural associations . We aim to elucidate how far associations are from the underlying causal variants , how their frequencies are distributed and , more importantly , how these different signatures should alter the design of fine-mapping studies . To examine possible heterogeneity in these signatures across the genome and across populations with different evolutionary histories , we repeated the analysis for several resequencing loci on different chromosomes and for two populations , one West African and one North European . The novelty of this study is in elucidating implications of synthetic associations and how they may affect fine-mapping strategies with patterns of LD as observed in human populations .
To empirically investigate the signatures of “synthetic associations” , we needed to examine scenarios in human genetic data where the presumed disease risk variants—rare or common—are known . Thus , we considered “disease loci” in the ENCODE regions that were sequenced as part of HapMap 3 [49] . The advantages of using these resequencing data are overcoming ascertainment biases that plague genotyping arrays [45] , [50]–[52] and observing variants of much lower allele frequency . Equipped with resequencing data for over 110 individuals in each population , we studied variants that appeared at least twice in 220 chromosomes . We randomly assigned variants within each disease locus as being causal and considered individuals carrying any one of these variants to have elevated disease risk . We then probabilistically assigned individuals to be either cases or controls based on their assigned risk . To mimic the case of many rare variants of large effect size underlying synthetic associations , and to contrast it with that of a few common variants of moderately low effect sizes underlying natural associations , we investigated three scenarios: ( i ) 2 common causal variants with a genotypic relative risk ( GRR ) of 1 . 5 , ( ii ) 5 and ( iii ) 9 rare causal variants with a genotypic relative risk of 3 . We verified that our results are not an artifact of the number of causal variants , as illustrated in the following , by comparing with a less realistic scenario of 5 common causal variants . We also considered a random assignment of cases and controls , which provides a null distribution in the absence of any risk alleles . After obtaining a set of cases and a set of controls , we performed an association study using the genotyping array data for the same individuals from HapMap 3 [49] , without considering any of the resequencing data in which disease loci have been emulated ( Materials and Methods ) . This mimics the conditions and variant-type of actual genotyping-based GWAS , which typically utilize array data of mainly common markers , most often using the same or similar arrays to those we have used for our analyses ( a combination of Affymetrix Human SNP array 6 . 0 and Illumina Human1M ) . We report results for association testing of all genotyped markers located within 3 cM of the resequenced disease locus , after verifying that the vast majority of significant associations are within those bounds ( Materials and Methods ) . Similar to the requirement of genome-wide significance in a GWAS , we required significance following multiple-hypothesis correction for the entire region tested , such that our results can be extrapolated to genome-wide studies . We repeated the association testing for 5 different disease loci ( Table 1 ) and for 50 sets of random assignments of causal variants in each locus . For each of these sets , we repeated the association testing in 10 replicates , varying between them only the stochastic assignment of cases and controls , for a total of 500 association tests in each locus for each of the three scenarios of causal variants . We also considered separately both a European ( CEU ) and a West African ( YRI ) population . Because of the relatively small sample size of ∼110 individuals , we simulated a larger sample using HAPGEN [53] , which maintains the genetic variation observed in the original data , including patterns of LD and MAF ( Materials and Methods ) . All scenarios show significant associations much more often than the false discovery rate of 5% ( Table S1 ) . To determine whether “synthetic associations” due to underlying rare variants tend to be further away than “natural associations” due to underlying common variants , we considered for each association test the distance between any association and the causal variant with which it is in strongest LD ( Materials and Methods ) . We found that the median distance , over the many hundreds of associations found across the 500 tests , is variable across the five loci and—to some extent—between the two populations ( Figure 1 ) . Synthetic associations tend to be much further than natural associations , as previously predicted [13] , though for one region ( disease locus #1 ) both synthetic and natural associations are in close proximity to the causal variants ( Figure 1 ) . Alternatively , when considering the distance between an association and the closest causal variant ( rather than the one in strongest LD ) , the distance of synthetic associations is reduced , yet generally remains greater than that of natural associations ( Figure S1 ) . Taken together , these results lead us to ask what factors contribute to this increased distance , and , more importantly , to what extent this increased distance should impact the choice of fine-mapping strategies . We explored several plausible explanations for this increased distance . Firstly , we ensured that the increased distance of rare causal variants is not due to more variants in those scenarios ( 5 and 9 ) than in the scenario of common causal variants ( 2 ) by repeating our analysis for cases with 5 common causal variants . We observed no increase in association distance of resultant natural associations ( Figure S2 ) , revealing that the increased distance is not due to the increased number of causal variants . Secondly , we investigated the hypothesis that increased marker effect size can cause greater association distances since association power is proportional to effect size times the correlation between the causal variant and the marker [37] . We investigated this hypothesis by increasing the effect size of common causal variants to equal that in the scenario of rare causal variants , though such an effect size might be considered unrealistic for common variants . The median association distance of the resulting natural associations indeed increases for all regions and populations , but is still considerably lower than synthetic associations in most cases ( Figure 1 ) . We next tested whether the age of the mutation played a role in increasing association distances for synthetic associations . As rare variants are , on average , resultant of more recent mutations compared to common variants , recombination would have had less time to operate , thus resulting in diminished decay of LD and haplotype structure around rare variants [39] . To test whether the age of the mutation plays a part in explaining our results , we partitioned rare causal variants into two age groups: i ) variants due to relatively more recent mutations and ii ) variants due to relatively older mutations . Variants with minor alleles present in only a single population fell into the former category , while those with minor alleles present in more than one population fell into the latter ( Materials and Methods ) . We observed a larger distance between an associated marker and the causal variant with which it is in highest LD for more recent mutations than for older mutations ( Figure 2 ) . Out of the 4 disease loci for which enough data was available to perform this analysis , 3 in YRI and 2 in CEU exhibit a median distance from older rare causal variants that is at least 41% less than the median distance from more recent causal variants . Combined , these results suggest that the increased distance of synthetic associations compared to natural associations is partially due to the young age of the mutations that give rise to rare risk alleles , as well as due to the higher effect size that is claimed to be implicated for rare risk alleles . The main concern regarding synthetic associations is how their signatures alter the search for the actual causal variant ( s ) . Specifically , how far should one sequence around an association in order to capture causal variants ? We addressed this question using two approaches . We first computed for each scenario of causal variants the fraction of tests ( out of all tests with any significant association ) that had at least one associated marker within any given distance of the causal variant with which it is in highest LD . We found that for common causal variants , a shorter resequencing distance of 0 . 01 cM is enough to capture a causal variant in 90% of the tests in CEU and 77% for YRI ( Figure 3 ) . For rare causal variants , combined over all disease loci , at least 90% of tests discovered an association within 0 . 1 cM of a causal variant ( Figure 3 ) . Secondly , we investigated a scenario in which fine-mapping consists of sequencing the LD block of associations as observed in the data . Hence , we estimated the probability that an associated marker is in the same LD block as any of the causal variants , with the definition of LD blocks being based only on markers from the genotyping arrays , which are relatively common ( Materials and Method ) . On average , the LD blocks spanned 0 . 007 cM for CEU and 0 . 005 cM for YRI , after the addition of a flanking region of 0 . 0005 cM . We found that in CEU , 94% of associated markers had a common causal variant in the same LD block , while the same was true for only 78% of associated markers in the rare causal variant case . A similar trend was observed for YRI , albeit less marked , where 79% of natural associations captured a causal variant , but only 73% of synthetic associations captured a causal variant . Finally , we explored the minor allele frequency ( MAF ) of associated markers and found that natural associations are of higher frequency on average than synthetic associations ( Figure 4 ) . Summing over all disease loci and populations , <1% of natural associations had MAF below 0 . 1 , while this proportion increased to 15–28% for synthetic associations . Dissecting the signal further by region and population , we found that while some regions display less than 2 . 4% difference between the median MAF of natural associations and synthetic associations ( disease locus #1 in YRI , #2 in CEU ) , others display an almost 200% difference ( #4 in CEU ) . Synthetic associations also display a larger standard deviation in associated MAF across different associations in different sets and replicates as compared to natural associations , with all but one region displaying a difference ranging from 17%–70% ( Table 2 ) .
With the use of HapMap 3 resequencing and genotyping data from five different genomic regions and two populations [49] , we considered several scenarios of disease risk loci , and performed association tests to investigate the signatures of synthetic associations and how they alter one's approach for studying them . We found that the median distance of synthetic associations , while greater than that of natural associations , still never exceeds 0 . 15 cM ( ∼150 kb ) for any of the 10 locus-by-population settings . Even if we instead consider the worst-case scenario of the largest distance between any association and any causal variant , its median still never exceeds 0 . 41 cM ( ∼410 kb ) . These results are in clear contrast to the results of a previous simulation-based study that showed the median of the largest distance to be 5 cM ( 5 Mb ) [13] . The difference between the two studies may be attributed to differences in the frequencies of rare causal variants . We considered rare alleles of frequency in the range 0 . 005–0 . 04 ( average across all variants of 0 . 019 ) , while Dickson et al . simulated allele frequencies in the range 0 . 005–0 . 02 [13] ( average of 0 . 0125 assuming uniform sampling ) . However , when we restricted to a narrower range of frequencies up to 0 . 02 ( average of 0 . 012 ) , we still observed no locus for which the median distance of synthetic association exceeds 0 . 5 cM ( ‘All variants’ in Figure 2 ) . It is unlikely that any remaining slight difference in risk allele frequency would result in over an order of magnitude difference in association distance . A more substantial difference between the two studies lies in the data analyzed . Dickson et al . conducted simulations of constant effective population size , uniform recombination rate , and neutral loci , with association testing based on a simulated “genotyping array” that follows a uniform ascertainment bias [13] . Here , we have analyzed data with empirically observed LD patterns , and have based association testing on data from real genotyping arrays as designed for GWAS . Put together , while theory posits that a median distance of synthetic associations of 5 cM is possible , characteristics of empirical data suggests that such cases will not be common , and that even under the worst-case scenario the vast majority of synthetic associations are at least an order of magnitude closer . By considering which of the rare polymorphisms are population-specific , and hence likely to be more recent , we illustrated that the increase in association distance is partially due to the age of the mutation . This is likely a result of recombination having had less time to break down the haplotype surrounding more recent mutations . We also considered common causal variants with a higher effect size and showed that association distance is increased . As rare causal variants contributing to an association signal are claimed to have higher effect sizes than common causal variants , the increased distance for synthetic associations can thus partially be due to the larger effect size . Additionally , the contribution of multiple rare causal variants to a single signal of association may also increase association distance –a source we have yet to fully explore in detail . To assess the impact of this increased association distance , we explored the probability that an association test had at least one association where the causal variant with which it was in highest LD lay within a given distance from the association . We found that for rare causal variants a window size of 0 . 1 cM was sufficient to capture at least one causal variant in such a manner in at least 90% of the tests for all regions and populations ( Figure 3 ) . Alternatively , by following an LD block based approach for fine-mapping , 73–79% of synthetic associations capture at least one of the rare causal variants within the same LD block . This suggests that traditional LD block-based fine-mapping offers a pretty high probability of discovering some of the causal variants , though there could still be added benefit from sequencing a larger region . Preliminary analysis suggests that it is difficult to predict the optimal region to resequence given a specific disease locus , as no single factor , such as pair-wise LD decay , can sufficiently predict this distance ( data not shown ) . Further work is thus necessary in order to determine which factors that influence synthetic associations , such as the age of mutation , causal variant effect size , haplotype structure and the stochastic coupling of multiple rare variants on the background of a common marker , play a role in an observed association signal . In a further analysis , we found that the causal variants being rare entails that the associated markers will themselves be of lower frequency ( Figure 4 ) , a result consistent with previous simulation studies [13] , [36] . When narrowing the number of associations to only the most significant , we found that this further reduced the allele frequency of synthetic associations ( Figure S3 ) . In addition , we found that the frequency of synthetic associations often had a larger standard deviation than natural associations ( Table 2 ) . These results have two implications . Firstly , it suggests that synthetic associations as compared to natural associations are likely to have underestimated effect sizes of the causal variant due to reduced associated allele frequencies [54] ( especially when analyzing the most significant association ) and from incomplete LD with the causal variant . Secondly , this suggests that the standard deviation of the associated minor allele frequency can offer a way to flag for underlying rare causal variants that induce potential synthetic associations; given a larger standard deviation of associated frequencies , it would be advised to follow a wider fine-mapping study design . Due to the >1000-fold human population growth in the last hundreds of generations , the amount of rare variation is much greater than expected [38]–[41] . This explosive addition of rare variation entails an LD structure that is yet to be quantified , but certainly disparate than the extensively studied LD structure of common variants . In addition , the earlier founder events as modern humans migrated out of Africa and settled across the globe have been shown to greatly alter patterns of genetic variation [42]–[45] , [55] . For this reason we studied both a West African population and a population of European ancestry , with differences in our results between the two reinforcing the importance of taking demographic history into consideration by studying empirical data . The effect of evolutionary history on signatures of synthetic and natural associations is further supported by the highly variable behavior across genomic regions of all the signatures we observed . In conclusion , this study delivered a characterization of several signatures of synthetic associations and assessed their impact on the search for the causal variant ( s ) underlying the signal . While our study does not take part in the debate on how frequently synthetic associations occur , it is relevant in any situation in which they do . We illustrated that because synthetic associations are likely to be more distant from causal variants , fine-mapping studies should look further than when searching for common causal variants , but to a much lesser extent than previously suggested . We also propose the larger standard deviation of associated allele frequencies as a way to detect potential rare causal variants at play . Additional analysis is warranted though , to elucidate the quantitative relationship between genetic architecture , demographic history , allele frequency and association signals . Finally , although the debate remains open as to the contribution of rare risk alleles to human complex diseases and to the ensuing abundance of synthetic associations [34]–[36] , [56] , our results offer new guiding principles for determining a length of a region to fine-map , and for considering the alternative of an association study based on whole-genome or whole-exome sequencing .
We obtained from HapMap 3 [49] genotyping array data for YRI ( Yoruba in Ibadan , Nigeria ) and CEU ( individuals in Utah with Northern and Western European ancestry from the Centre d'Etude du Polymorphisme Humain collection ) and resequencing data of five ENCODE regions , each 100 kb in length ( Table 1 ) , for 115 YRI and 111 CEU individuals . We also obtained resequencing data for 60 TSI ( Toscani in Italia ) samples and 60 LWK ( Luhya in Webuye , Kenya ) , which we used for the variant age analysis ( below ) . We considered each resequencing region as a disease locus from which to select causal variants . Using resequencing data facilitates higher concentration of rare variants and is free of the ascertainment biases associated with genotyping arrays [45] , [50]–[52] . Due to the low sample size , we employed HAPGEN [53] to simulate 10 , 000 individuals for each population –a strategy previously employed to investigate the estimation of relative risks [54] . HAPGEN simulates additional haplotypes by treating each new haplotype as a mosaic of already present haplotypes . We refer readers to [53] for additional details on HAPGEN . We first phased and imputed missing data with BEAGLE v3 . 3 [57] . We then simulated additional data for each resequencing region and the 3 cM-flanking window for each region using HAPGEN with a recombination map from the March 2006 human reference sequence ( NCBI Build 36 , hg18 ) and a null mutation rate as input parameters . We ensured that the LD patterns of the original data ( for rare and common variants ) were maintained ( Figure S4 ) . We also ensured that allele frequencies in the simulated data do not change drastically from the original data as no variants were observed that were initially of very low frequency and attained a much higher frequency and vice versa in the simulated dataset ( Figure S5 , S6 ) . Association tests were performed using the simulated data from the HapMap 3 genotyping array data , excluding any causal variants that happen to be in the genotyping array data . We report results for an association study for SNPs located in the disease locus and in flanking regions of 3 cM on each side ( from which no causal variants are chosen ) , as almost no associations were observed to fall beyond that distance ( data not shown ) . In our study , rare causal variants have risk allele frequencies in the simulated data between 0 . 005 and 0 . 04 ( we note that a portion of this range is defined as “low frequency” , rather than rare , by some studies ) , and common causal variants have risk allele frequencies in the simulated data between 0 . 1 and 0 . 3 . In testing for association , we considered all SNPs of all allele frequencies from the genotyping data . All coordinates and genetic distances in this paper are according to the March 2006 human reference sequence ( NCBI Build 36 , hg18 ) . We considered each individual as a case or a control with a probability proportional to the individual's assigned risk , which is elevated if the individual has one or more risk alleles . We set the baseline risk as 0 . 15 and the genotypic relative risk to 1 . 5 for the scenario of common causal variants . We also explored an unrealistic genotypic relative risk of 3 for common causal variants to investigate the influence of effect size on association distance . For rare causal variants , we assigned a higher genotypic relative risk of 3 . While the use of a fixed GRR for variants of differing allele frequencies results in differing portions of variance explained by each variant , it is a more realistic disease model . By fixing variance explained , rarer variants would tend to have higher , and perhaps somewhat unrealistic , GRRs . Because we have fixed GRR and allowed the proportion of variance explained to vary , an association test will have more power in detecting variants of higher allele frequency given a fixed GRR . For the common causal variants scenario , we randomly assigned 2 SNPs from the resequencing data as causal , while we assigned either 5 or 9 for the rare causal variants scenario . To ensure that the number of causal variants did not affect our results , we also studied a scenario with 5 common causal variants in loci where this was feasible . For each scenario of a certain type and number of causal variants , 50 sets of causal variants were randomly selected , with replacement between groups . Each of these 50 sets allows for a possibly different risk for each individual . For each of these 50 sets , we repeated 10 replicates of randomly assigning cases and controls according to the same individual assigned risk . In each of the 500 association tests ( 50 different variant groups and their 10 phenotypic replicates ) , we randomly chose 1000 cases and 1000 controls according to the individual's assigned risk . This ensures that the same number of cases and controls were shared across all analyses , thereby having comparable statistical power . For each scenario of type and number of causal variants , we pooled together the results from these 500 tests for the statistics and figures presented in this study . Similarly , we generated 500 tests for each disease locus with randomly assigned case/control status to serve as a control . All association tests were done with PLINK's logistic regression function [58] . Significance thresholds were determined with a region-wide Bonferroni correction . For the control scenario of random assignment of cases and controls , 2 . 12% of the association tests showed a significant association as compared with the expectation of 5% . We determined genetic distances based on the Oxford genetic map based on HapMap2 data [50] , [59] . For SNPs missing from HapMap2 , we estimated the position as the linear interpolation of the genetic positions of the two closest SNPs . The association distances were determined by computing the genetic distance between an associated SNP and the causal variant with which it was in highest LD , measured in r2 . Pairwise r2 values were calculated in pLINK [58] . To partition rare variants based on the age of the mutation , we first narrowed the range of the risk allele frequency in the simulated data to 0 . 005 and 0 . 02 in order to ensure a roughly equal partition into the two age groups . We discarded disease locus #1 from this analysis because it had too few rare variants to allow their portioning into two groups ( Table 1 ) . Rare variants in the 111 CEU individuals were defined to be relatively more recent if only the major allele was observed in the resequencing of 115 YRI individuals and 60 TSI individuals in the original data; the variant was defined as relatively older otherwise . We repeated the above analyses for each of these groups separately , such that in each association testing either all causal variants are older or all are more recent . We repeated the same analysis in YRI with CEU and 60 LWK as out groups . We duly note that polymorphisms absent from the limited number of samples may not be monomorphic in the population as a whole , hence not all mutations leading to relatively older variants precede those leading to variants in the relatively more recent class . Yet , this represents only a small fraction of variants and variants in the relatively older class are expected to be older on average than those belonging to the more recent class . It is also important to note that false positive variant calls are added to the more recent group despite the erroneous call . This scenario is highly unlikely in our analyses due to the stringent quality control measures taken in HapMap 3 [45] and the exclusion of singletons in our study . For each of these two scenarios of causal variants , we similarly chose 50 sets of causal variant groups with 10 phenotypic replicates each and obtained maximal distances as above . For comparison , we repeated the analysis for random rare causal variants in the narrowed range of frequency of 0 . 005–0 . 02 used here , irrespective of mutation age . For each association test we explored whether a causal variant with which an association is in highest LD ( measured in r2 ) is within a given genetic distance from the association . For each simulated scenario and resequencing window size ranging from 0 cM to 10 cM , we calculated the proportion of tests that have at least one such association . For the second analysis , we observed over all significant associations if any causal variant was in the same LD block as an association . LD blocks were estimated in pLINK with the genotyping data [58] and 0 . 0005 cM was added to the start and end coordinates in order to compensate for the uncertainty in these estimates .
|
Genome-wide association studies ( GWAS ) , based on the hypothesis that common genetic variation underlies complex diseases , have found many sites in the genome associated with complex diseases . However , both the fraction of variation explained by these sites and the number of studies identifying causal variants remain low . While there are many possible explanations for these issues , we focus on one theory that suggests rare variants also play a significant role in complex diseases . We investigated the effect of rare causal variants as compared to common causal variants in simulated data with patterns of variation observed in actual human genetic data . As suggested by previous studies , we found that rare causal variants show different signatures in GWAS results . We explore in this study the implications of these differences in influencing the search for causal variants underlying the signal of association .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genome-wide",
"association",
"studies",
"genetic",
"polymorphism",
"genetics",
"population",
"genetics",
"biology",
"human",
"genetics",
"genetics",
"and",
"genomics"
] |
2012
|
Predicting Signatures of “Synthetic Associations” and “Natural Associations” from Empirical Patterns of Human Genetic Variation
|
CD4 T cells are critical for control of persistent infections; however , the key signals that regulate CD4 T help during chronic infection remain incompletely defined . While several studies have addressed the role of inhibitory receptors and soluble factors such as PD-1 and IL-10 , significantly less work has addressed the role of T cell co-stimulatory molecules during chronic viral infection . Here we show that during a persistent infection with lymphocytic choriomeningitis virus ( LCMV ) clone 13 , mice lacking the glucocorticoid-induced tumor necrosis factor receptor related protein ( GITR ) exhibit defective CD8 T cell accumulation , increased T cell exhaustion and impaired viral control . Differences in CD8 T cells and viral control between GITR+/+ and GITR-/- mice were lost when CD4 T cells were depleted . Moreover , mixed bone marrow chimeric mice , as well as transfer of LCMV epitope-specific CD4 or CD8 T cells , demonstrated that these effects of GITR are largely CD4 T cell-intrinsic . GITR is dispensable for initial CD4 T cell proliferation and differentiation , but supports the post-priming accumulation of IFNγ+IL-2+ Th1 cells , facilitating CD8 T cell expansion and early viral control . GITR-dependent phosphorylation of the p65 subunit of NF-κB as well as phosphorylation of the downstream mTORC1 target , S6 ribosomal protein , were detected at day three post-infection ( p . i . ) , and defects in CD4 T cell accumulation in GITR-deficient T cells were apparent starting at day five p . i . Consistently , we pinpoint IL-2-dependent CD4 T cell help for CD8 T cells to between days four and eight p . i . GITR also increases the ratio of T follicular helper to T follicular regulatory cells and thereby enhances LCMV-specific IgG production . Together , these findings identify a CD4 T cell-intrinsic role for GITR in sustaining early CD8 and late humoral responses to collectively promote control of chronic LCMV clone 13 infection .
During chronic viral infections , exemplified by the clone 13 variant of lymphocytic choriomeningitis virus ( LCMV cl 13 ) , persistent antigen presentation results in the functional exhaustion of the T cell response , characterized by persistent upregulation of inhibitory molecules and a progressive loss of T cell effector functions [1] . Ultimately , LCMV cl 13 is cleared by 60–90 days p . i . due to both T- and B-cell responses . While there have been a number of studies on the role of T cell inhibitory receptors and anti-inflammatory cytokines during chronic infection [2–7] , rather less has been done to study the role of co-stimulatory receptors in this context . Co-stimulatory TNFR family members are of particular interest in this regard because they are often induced upon antigen receptor signaling , leading to their co-expression with inhibitory receptors during a persistent infection [8–10] . CD4 T cell help is critical for the control of chronic infections . The removal of CD4 T cells from mice prior to infection with LCMV cl 13 [11–13] , or the loss of CD4 T cells during progressive HIV infection [14] leads to increased viral burden , immune dysregulation , and functional T cell exhaustion . While CD4 cells are clearly implicated in the control of chronic viral infections , the co-stimulatory signals that contribute to CD4 T cell help remain poorly defined . Evidence to date suggests that there is significant heterogeneity in the potency and mechanisms of T cell modulation by members of the TNFR superfamily during chronic viral infection [8 , 9 , 15–18] . The Glucocorticoid-induced TNFR related protein ( GITR ) and its ligand ( GITRL ) are induced upon activation of a number of immune cell types [19] . GITR is expressed at low levels on resting T cells , but its expression rapidly increases upon activation . Although constitutively expressed on Foxp3+ regulatory T cells ( Treg ) , GITR is dispensable for Treg function but can play a role in the accumulation of both regulatory and effector T cells in vivo [20 , 21] . In many experimental systems , it has been difficult to determine the key cell types through which GITR mediates its effects . Controversy over the relative role of GITR on effector versus regulatory T cells persists , though these effects may be context-dependent [19 , 22 , 23] . GITR-deficient mice are viable , reproduce normally , and have no obvious immunological defects in the naïve state [24] . Studies using GITR-deficient mice or an agonist anti-GITR antibody have shown an immune stimulatory role for GITR in the context of viral infections , but the mechanisms underlying these phenomena are not entirely clear [19] . Here we address the role of GITR during chronic viral infection using GITR-deficient mice infected with LCMV cl 13 [25] . We find that GITR-/- mice have 2–3-fold fewer LCMV-specific CD8 T cells post-priming and throughout the infection , with concomitant increases in viral load . A previous study showed that GITR on TCR transgenic CD8 T cells contributes to the control of influenza infection [26] . Surprisingly however , during LCMV cl 13 infection , the effect of GITR was largely CD4 T cell-intrinsic and independent of Tregs . GITR-deficiency impaired the accumulation of IFNγ+IL-2+ T helper 1 ( Th1 ) cells as well as follicular helper T cells ( Tfh ) and the production of LCMV-specific IgG . Mixed bone marrow chimeras and adoptive transfer studies of LCMV-specific CD4 and CD8 T cells demonstrate that the effects of GITR are CD4 T cell-intrinsic . Neutralization of IL-2 in GITR+/+ mice reduced the immune response to the level observed in GITR-/- mice , whereas IL-2 neutralization in GITR-/- mice had no effect , revealing a central role for this cytokine in GITR-dependent control of LCMV cl 13 . Taken together , our studies provide evidence for a critical role for GITR in the post-priming accumulation of IL-2+ Th1 cells to help cell-mediated immunity as well as Tfh accumulation to sustain the humoral arm of the immune response to control a persistent viral infection .
To evaluate the role of GITR in a chronic infection , we infected wild-type mice ( GITR+/+ ) or mice with ablated Tnfrsf18 ( GITR-/- ) [24] with LCMV cl 13 . At both acute and chronic time points ( days eight and 45 p . i . ) , GITR-/- mice had significantly increased viral load ( Fig . 1A ) . These effects were particularly striking in the kidney , where GITR-/- mice had 35-fold higher viral load than GITR+/+ mice at day eight p . i . GITR-/- mice had substantially fewer LCMV-specific CD8 T cells in peripheral blood ( Fig . 1B , for gating strategy see S1 Fig . ) . Following initial infection and T cell priming at day five p . i . , there was no difference in the number of splenic LCMV-specific CD8 T cells between GITR+/+ and GITR-/- mice ( Fig . 1C ) . However , between days five and eight p . i . , a significant impairment in the frequency and total number of splenic LCMV-specific CD8 T cells in GITR-/- mice materializes , and this effect becomes even more striking at day 45 p . i . ( Fig . 1 , D–F ) . Co-signaling molecules PD-1 and Tim-3 impair CD8 T cell responses to LCMV cl 13 [4 , 5] . The majority of the LCMV-specific CD8 T cells in GITR+/+ and GITR-/- mice were PD-1+ Tim-3+ after LCMV cl 13 infection , but CD8 T cells from GITR-/- mice expressed significantly higher levels of PD-1 and Tim-3 per cell at days eight and 45 p . i . ( Fig . 2 , A and B ) . CD8 T cells from GITR+/+ or GITR-/- mice produced similar amounts of IFNγ following peptide restimulation at day eight p . i . , but by day 45 p . i . , the GITR-/- mice had fewer IFNγ+ CD8 T cells with lower IFNγ production per cell ( Fig . 2 , C and D ) . Moreover , at all time points analyzed , fewer of the IFNγ-producing CD8 T cells from GITR-/- mice exhibited a multifunctional phenotype as defined by TNF and surface CD107a expression ( a surrogate marker for cellular cytotoxic activity ) ( Fig . 2 , E and F ) . These findings reveal a key role for GITR in the CD8 T cell response and control of chronic LCMV infection . We also examined the role of GITR with the acutely infecting Armstrong variant of LCMV and observed approximately two- to three-fold more GP33–41-specific CD8 T cells in GITR+/+ compared to GITR-/- at day eight p . i . ( p<0 . 01 in two of the three experiments , total of ten mice per group ) , however both GITR+/+ and GITR-/- mice cleared LCMV Armstrong infection by day eight p . i . Previous work has shown an intrinsic role for GITR in sustaining the survival of TCR transgenic CD8 T cells during acute influenza virus infection [26] . To determine if CD8 T cell-intrinsic effects were also responsible for defects in LCMV cl 13 control in GITR-/- mice , we crossed GITR-/- with P14 mice , which have a transgenic TCR specific for the Db-restricted LCMV GP33–41 epitope [27] . We transferred 103 purified CD8 T cells from CD45 . 1 GITR+/+ or GITR-/- P14 littermates into CD45 . 2 congenic mice one day prior to LCMV cl 13 infection ( Fig . 3A ) . P14 T cells lacking GITR showed no impairment in expansion , and in fact showed a slight increase in frequency ( Fig . 3B ) . There was also no difference in IFNγ production or degranulation , PD-1 or Tim-3 expression between the GITR-sufficient or GITR-deficient P14 T cells ( Fig . 3 , C–E ) . Moreover , the absence of GITR only on the transferred P14 T cells had no effect on viral control at day eight p . i . ( Fig . 3F ) . Thus , the effect of GITR in potentiating the CD8 T cell response to LCMV cl 13 is CD8 T cell-extrinsic . To evaluate the importance of CD4 T cells in this model , we depleted CD4 cells from GITR+/+ and GITR-/- mice prior to LCMV cl 13 infection ( Fig . 4A ) . CD4-depleted GITR+/+ mice had fewer and less functional LCMV-specific CD8 T cells , and 100-fold higher viral load relative to non-CD4-depleted GITR+/+ mice , consistent with previous reports of CD4 depletion increasing the severity of LCMV cl 13 infection [11–13] . However , the qualitative and quantitative differences between GITR+/+ and GITR-/- CD8 T cell responses and viral control at day eight p . i . were largely lost in CD4-depleted mice ( Fig . 4 , B–E ) , suggesting that CD4 T cells are necessary for the effect of GITR on the CD8 T cell response to LCMV cl 13 as well as viral control . We next analyzed the effect of GITR on CD4 T cell responses . GITR-/- mice had similar total numbers of CD4 T cells in the spleen following LCMV infection at day eight p . i . However , fewer of the GITR-/- CD4 T cells expressed the signature Th1 transcription factor T-bet ( Fig . 4F , for gating strategy see S2 Fig . ) and there was a striking three-fold reduction in the percent of CD4 T cells that were IFNγ+ or IFNγ+IL-2+ following GP61–80-restimulation at day eight p . i . , a defect that was maintained out to day 45 p . i . ( Fig . 4 , G and H ) . In contrast , GITR+/+ and GITR-/- littermates had similar proportions and absolute numbers of Foxp3+ Tregs ( Fig . 4F ) . Moreover , depletion of Tregs did not recapitulate the effects of total CD4 T cell depletion on the CD8 T cell response ( S3 Fig . ) . IL-2 is important for CD8 T cell accumulation and function in the context of chronic LCMV infection [28–30] . Therefore , we evaluated whether the effects of GITR on the T cell response were dependent on IL-2 . Of note , we did not detect any IL-2+ CD8 T cells after ex vivo peptide stimulation at day eight p . i . , consistent with previous reports [31] . As the CD8 T cell deficit in GITR-/- mice is not realized until after day five p . i . , we administered 0 . 5mg of blocking anti-IL-2 to mice on days four and six p . i . , and evaluated T cell responses at day eight p . i . IL-2 blockade reduced the LCMV-specific Th1 and CD8 T cell responses in GITR+/+ mice to the level observed in GITR-/- mice , whereas IL-2 blockade had no impact on the T cell responses in GITR-/- mice ( Fig . 4 , I and J ) . Thus , GITR is critical for Th1 responses and IL-2-dependent help for CD8 T cells . GITR impacts viral load both early and late during LCMV cl 13 infection ( Fig . 1A ) . As Tfh have emerged as an important CD4 T cell subset that promotes late B cell responses to LCMV cl 13 infection [32 , 33] , we examined the effects of GITR on Tfh . At day eight p . i . , we noted elevated levels of GITR expression on Tfh compared to non-Tfh CD4+ SMARTA cells ( SMARTA mice express a transgenic TCR specific for the I-Ab-restricted LCMV GP61–80 ) ( Fig . 5A ) [34] . Evaluation of the Tfh response in peripheral blood of GITR+/+ and GITR-/- mice between day seven and 45 p . i . indicated a significantly lower frequency of Tfh in GITR-/- mice compared to GITR+/+ mice at late ( days 21 and 45 p . i . ) but not early times during the response to LCMV cl 13 ( Fig . 5B ) . T follicular regulatory cells ( Tfr ) are Foxp3+ T cells that localize to germinal centers to regulate the response [35 , 36] . The spleens of GITR+/+ and GITR-/- mice had similar proportions of Tfh and Tfr at day eight p . i . , but at day 45 p . i . GITR-/- mice exhibited significantly fewer Tfh and there was a significant increase in the proportion of Tfr in GITR-/- compared to GITR+/+ mice ( Fig . 5 , C and D ) . Moreover , GITR-/- mice had significant impairment in the LCMV-specific IgG response at day 21–45 p . i . ( Fig . 5E ) . Although CD4 T cells are required for the effect of GITR on the CD8 T cell response to LCMV , it was conceivable that GITR on another cell population could indirectly affect CD4 T cell help [19] . To evaluate a cell-intrinsic effect of GITR , we used mixed bone marrow chimeras , in which GITR+/+ and GITR-/- cells compete in the same mouse . If GITR intrinsically affects a specific cell type , then the GITR-/- cells will be at a competitive disadvantage , whereas if the effect of GITR is extrinsic then GITR+/+ and GITR-/- cells would be equally affected by the lack of GITR in another cell compartment and their ratio would be constant . To this end , we reconstituted lethally irradiated C57BL/6 mice with 1:1 mixture of CD45 . 2 GITR+/+:CD45 . 1 GITR-/- bone marrow cells ( Fig . 6A ) . Although we used CD45 . 2 recipients , we have found that in separate control experiments using identical radiation schedules , >95% of lymphocytes were of donor origin at three months post-reconstitution . Consistent with the minimal contribution from residual T cells following lethal irradiation , we verified that the CD45 . 2:CD45 . 1 ratio in peripheral blood was 1:1 , with similar proportions of CD4 , CD8 T cells and Foxp3+ Tregs prior to infection ( Fig . 6B ) . At day eight p . i . , there was a slight but significant effect of GITR on total CD4 cells , with fewer CD4 T cells in the GITR-/- compartment ( Fig . 6C ) . Among the CD4 cells , there was a two-fold reduction in GITR-/- Tregs and a three-fold reduction in GITR-/- Tfh ( Fig . 6 , C and D ) . Additionally , there was a three-fold reduction in IFNγ-producing GITR-/- Th1 cells after LCMV GP61–80 restimulation , of which fewer co-produced IL-2 ( Fig . 6 , E and F ) . Taken together , these data demonstrate a key CD4 T cell-intrinsic role of GITR during LCMV cl 13 infection , with the most dramatic effects on the Th1 and Tfh responses . At day eight p . i . , there was no difference in the frequency of GITR+/+ and GITR-/- NP394–404-specific CD8 T cells , whereas the frequency of GITR-/- GP33–41-specific T cells was significantly reduced compared to the GITR+/+ population ( Fig . 6G ) , however in contrast to the complete knockout mouse ( Fig . 2 ) , the levels of PD-1 and Tim-3 were similar between the GITR+/+ and GITR-/- compartments ( Fig . 6H ) . The frequency of CD8 T cells with the ability to produce cytokines and degranulate after GP33–41 peptide restimulation was also decreased in GITR-/- compared to GITR+/+ ( Fig . 6 , I–L ) . These data show that under competitive conditions , GITR intrinsically affects GP33–41 but not NP396–404-specific accumulation , whereas effects of GITR on T cell inhibitory receptor levels are cell-extrinsic , and likely due to increased viral load in the complete GITR-deficient mouse . To distinguish between the effects of GITR on CD4 T cell accumulation versus differentiation , we transferred a 1:1 mixture of GITR+/+:GITR-/- TCR transgenic SMARTA cells one day prior to LCMV cl 13 infection ( Fig . 7A ) . We used the congenic marker CD45 . 1 and GITR expression to identify GITR+/+ and GITR-/- SMARTA cells within the same mouse ( Fig . 7B ) . Differences in GITR+/+ and GITR-/- SMARTA accumulation showed a trend starting at day five p . i . , with significant differences by day eight p . i . , with a 3 . 2:1 ratio ( Fig . 7 , C and D ) . The ratio of GITR+/+:GITR-/- CD44hi , Tfh , and Th1 SMARTA cells was uniformly 3:1 , indicating that GITR is not skewing the differentiation of particular CD4 effector T cell sub-populations , rather it contributes to the accumulation of all CD4 T cells ( Fig . 7E ) . While there was a lower frequency of GITR- CD4 T cells producing IFNγ after GP61–80 peptide restimulation , GITR+ and GITR- IFNγ+ SMARTA produced a similar amount of IFNγ per cell . There was also a consistent trend toward fewer GITR- IL-2+ IFNγ+ SMARTA cells relative to the GITR+ SMARTA ( Fig . 7 , F and G ) . To determine whether GITR co-stimulation enhanced early division or post-priming accumulation of CD4 T cells , we transferred a mixture of 5×105 each of GITR+/+ and GITR-/- CFSE-labeled CD45 . 1 SMARTA cells into naive C57BL/6 ( CD45 . 2 ) mice . CFSE dilution at days three and five p . i . was comparable between GITR- SMARTA and GITR+ SMARTA ( Fig . 7H ) , suggesting that GITR does not influence initial division , but rather affects post-priming accumulation of CD4 T cells . The accumulation of T cells during their clonal expansion is well known to require survival signals , such as those mediated by NF-κB , as well as upregulation of protein translation and glycolysis , as mediated by mTOR [37–40] . Therefore , to investigate a possible mechanism underlying the enhanced accumulation of GITR+ SMARTA , a 1:1 mixture of GITR+/+ and GITR-/- CD45 . 1 SMARTA were transferred into C57BL/6 ( CD45 . 2 ) mice and phsopho ( p ) -p65 NF-κB and ( p ) S6 , a downstream mTOR target were evaluated in SMARTA cells at day three p . i . without further stimulation . Consistent with GITR-dependent NF-κB activation in CD8 T cells [26 , 41–43] , ( p ) p65 NF-κB was increased in GITR+/+ compared to GITR-/- SMARTA cells at day three p . i . and ( p ) S6 was also increased , consistent with GITR-dependent mTORC1 activation ( Fig . 7I ) . mTORC1 can be activated by Akt , a kinase that critically regulates T cell metabolism and expansion [37 , 38] . Akt activation has been noted downstream of other TNFRs , such as OX40 [37 , 44] . Therefore , we evaluated Akt activation following GITR cross-linking . GITR+/+ SMARTA splenocytes were activated in vitro with LCMV GP61–80 and IL-2 to induce GITR expression , then serum starved for 12 hours , followed by treatment with agonistic anti-GITR ( DTA-1 ) or Rat IgG control . DTA-1 enhanced Akt activation as demonstrated by increased ( p ) Akt Thr308 at 15 and 30 minutes post-stimulation ( Fig . 7J ) . Taken together , these studies show that GITR intrinsically affects post-priming accumulation but not early division of CD4 T cells with effects on NF-κB and the Akt-mTORC1-S6 signaling axis detected at day three p . i .
The mechanism by which CD4 T cell responses contribute to the control of persistent infection , or how co-stimulation regulates CD4 T cell responses to infection remains incompletely defined [45] . Here we show that GITR directly augments the CD4 Th1 response to LCMV cl 13 , which in turn impacts IL-2-dependent CD8 T cell expansion and viral control . The critical role of GITR on the CD4 T cells for control of LCMV cl 13 is demonstrated by the loss of GITR-mediated protection when CD4 T cells are depleted . Signaling downstream of GITR could be detected as early as day three p . i . , consistent with an effect of GITR on CD4 T cell accumulation starting around day five and delayed effects on CD8 T cells appearing between days five and eight p . i . These findings are also consistent with strong induction of GITRL by day two p . i . with LCMV cl 13 [46] . These results demonstrate an important role for GITR early in the immune response that impacts post-priming accumulation of CD4 T cells to allow help for CD8 T cells . CD4 T cell-intrinsic GITR also reciprocally affects the Tfh and Tfr responses and , consequently , LCMV-specific IgG production . As Tfh and antibody responses are thought to contribute to the late control of persistent LCMV infection [32] , this effect of GITR on Tfh and Tfr likely impacts the late control of the infection rather than the early control , when Tfh and anti-LCMV IgG levels are comparable between GITR+/+ and GITR-/- mice . Previous work demonstrated a CD8 T cell-intrinsic role for GITR on survival of adoptively transferred CD8 T cells during acute respiratory infection of mice with influenza A virus , with no impact on cell division [26] . Here we saw similar effects in the competitive bone marrow chimera experiment ( Fig . 6 ) . However , the CD4-depletion studies and the P14 GITR+/+ and GITR-/- adoptive transfer studies support a CD8 T cell-extrinsic effect of GITR . LCMV cl 13 infection results in systemic prolonged inflammation , perhaps providing additional signals to compensate for the lack of GITR on CD8 T cells during the early stages of LCMV cl 13 infection . Nonetheless , the effects of GITR on CD8 T cells can be revealed under competitive conditions , but are clearly not sufficient to compensate for the requirement for GITR on the CD4 T cells to allow help for CD8 T cell and antibody responses to chronic LCMV infection . LCMV-specific CD4 T cells can rescue exhausted CD8 T cell responses to LCMV cl 13 , although the precise mechanism of this CD4 T cell helper function remains open to speculation [47] . CD25-deficient CD8 T cells have defective persistence during chronic LCMV infection [30] , and IL-2 can be used therapeutically to augment the CD8 T cell response to Armstrong ( acute ) and cl 13 ( chronic ) LCMV infection [28 , 29] . Here we show that IL-2 blockade in GITR+/+ mice resulted in an immune response indistinguishable from that in GITR-/- mice , in which IL-2 blockade had no effect . These findings imply that the effect of GITR on CD4 help for CD8 T cells is IL-2 dependent . Although we have not directly shown that the CD4 T cells are the source of IL-2 , taken together with the CD4 depletion studies , the IL-2 blocking data are consistent with a lack of early IL-2+ Th1 accumulation as the underlying cause of the defective CD8 T cell accumulation and viral control in GITR-/- mice at day eight p . i . GITR-deficiency did not result in a change in the proportion or number of Tregs . However , under conditions of competition , there was a two-fold reduction in the number of GITR-/- Tregs , consistent with previous findings that GITR may have a role in peripheral maintenance of Tregs [20] . When CD4 T cells were depleted , the dominant effect was a net increase in viral load , consistent with a loss of T cell help rather than the removal of a regulatory CD4 T cell population in GITR-/- mice . Nonetheless , it is possible that GITR on effector cells could render them resistant to Tregs [20] . Therefore we used DEREG mice to further assess the role of Tregs in our model ( S3 Fig . ) . Diphtheria toxin ( DT ) was used to deplete Tregs from depletion of regulatory T cells ( DEREG ) mice , which express the human DT receptor under the Foxp3 promoter . DT was highly toxic in the LCMV cl 13 model , even in the non-DEREG mice . Viral load in this model was three to four orders of magnitude higher than in non-DT treated GITR+/+ mice , making the results difficult to interpret . Moreover , four of 26 LCMV cl 13 infected GITR+/+ mice treated with DT died , whereas all GITR-/- mice survived , likely due to reduced levels of inflammatory cytokines . Although the data from these experiments are difficult to interpret given the high level of toxicity , it is clear that Th1 as well as CD8 T cell differences between GITR+/+ and GITR-/- mice are largely retained even when >90% of Tregs were depleted . These data argue against a major role for GITR on the CD4 or CD8 effector T cells in preventing the action of Tregs in this model . Late in LCMV cl 13 infection Tfh accumulate and provide help for antibody responses [32] . GITR-/- mice have impaired Tfh responses with a slight concomitant increase in the Tfr population [35 , 36] and a modest decrease in LCMV-specific antibodies . Competitive bone marrow chimeras revealed that GITR also has cell-intrinsic effects on Tfh . GITRL is largely absent on total B cells , whereas a low level of GITRL can be detected on GL-7+ Fas+ germinal center B cells at days eight and 21 p . i . [46] . It is unlikely that the defect in LCMV-specific IgG production is due to direct effects of GITR on the B cell response , as GITR is largely dispensable for B cell development and activation [48] . Competitive adoptive transfer experiments showed that the absence of GITR on CD4 T cells resulted in a three-fold defect in total LCMV-specific CD4 T cells , Th1 , and Tfh sub-populations equivalently , indicating that the Th1 deficit in GITR-/- mice is not due to altered CD4 T cell differentiation . We demonstrated that differences in GITR+/+ and GITR-/- CD4 numbers are not due to differences in early CD4 T cell proliferation , but instead due to defects in post-priming accumulation of CD4 T cells . GITR induces classical NF-κB to induce the pro-survival molecule Bcl-xL in CD8 T cells [26] . T cell accumulation also requires upregulation of glycolytic and biosynthetic pathways , such as induced by the Akt-mTORC1 pathway [38] . Here we show that GITR can directly activate Akt and induce phosphorylation of the downstream mTORC1 target S6 , as well as activate the classical NF-κB pathway in CD4 T cells as early as day three p . i . Collectively , these signals allow the accumulation of IL-2+ Th1 cells , thereby allowing them to help the early CD8 T cell response , as well as to sustain the CD4 response as it progresses towards a Tfh fate . In sum , we report here a critical CD4 T cell-intrinsic role for GITR in control of chronic LCMV infection . We showed that GITR co-stimulation activates the Akt-mTORC1-S6 signaling axis as well as classical NF-κB signaling to collectively promote CD4 T cell expansion . GITR-deficient mice have defective post-priming accumulation of IL-2+ Th1 cells , which precedes an impaired CD8 T cell response and loss of viral control . The effects of GITR on both Th1 and CD8 T cells are IL-2-dependent . We also reveal a novel role for GITR in sustaining the Tfh response and LCMV-specific IgG production . These findings define a critical role for CD4 T cell-intrinsic GITR signaling in early accumulation of CD4 T cells , help for CD8 T cells and viral control .
Biotinylated H-2Db/GP33–41: KAVYNFATM , H-2Db/GP276–286: SGVENPGGYCL , and H-2Db/NP396–404: FQPQNGQFI monomers were obtained from the National Institute for Allergy and Infectious Disease tetramer facility ( Emory University , Atlanta , GA ) , and tetramerized with streptavidin APC ( Molecular Probes ) . Anti-CD107a ( clone 15A7 ) , anti-CD8 ( clone 53–6 . 7 ) , anti-CXCR5 ( clone 2G8 ) , and anti-TNF ( clone MP6-XT22 ) were purchased from BD Biosciences . Anti-phospho-Akt Thr308 ( C31E5E ) and anti-phospho-S6 ribosomal protein Ser235/236 ( clone D57 . 2 . 2E ) were purchased from Cell Signaling Technologies . Lectin PNA from Arachis hypogaea ( Cat . #L21409 ) and viability stain ( Cat . #L34955 ) were purchased from Molecular Probes . Anti-CD4 ( clone RM4–5 ) and CD45 . 1 ( clone A20 ) were purchased from BioLegend . FITC IgG isotype control , anti-Fas ( clone 15A7 ) , anti-CD25 ( clone PC61 . 5 ) , anti-LAP ( clone TW7–16B4 ) , anti-PD-1 ( clone J43 ) , anti-Tim-3 ( clone RMT3–23 ) , anti-ICOS ( clone 7E . 17G9 ) , anti-CD3 ( clone 1245–2C11 ) , anti-CD45 . 1 ( clone A20 ) , anti-CD45 . 2 ( clone 104 ) , anti-CTLA-4 ( clone UC10–4B9 ) , anti-B220 ( clone RA3–6B2 ) , anti-CD16/32 ( clone 93 ) , anti-Foxp3 ( clone FJK-16s ) , anti T-bet ( clone eBio4B10 ) , anti-IL-2 ( clone JES6–5H4 ) , anti-IFNγ ( clone XMG1 . 2 ) , phospho-NF-κB p65 Ser529 ( clone MCFA30 ) , Streptavidin-PE and -APC , and fixable viability dye eFluor 506 ( Cat . #65–0866–14 ) were purchased from eBioscience . Anti-IL-2 ( clone S4B6–1 ) was purchased from Bio X Cell ( West Lebanon , NH ) . Un-nicked Diphtheria Toxin was purchased from List Biological Laboratories , Inc . ( Campbell , CA ) and was used to deplete Foxp3+ Tregs from DEREG mice [49] by administering 1μg i . p . DT per two days from day -2 to day six p . i . LCMV Armstrong and cl 13 ( obtained from P . Ohashi , University of Toronto and M . Oldstone , Scripps Research institute ) were propagated on BHK and L929 cells , respectively ( kindly provided by Pamela Ohashi , Princess Margaret Cancer Centre , Toronto , Ontario ) . GITR+/+ and GITR-/- mice were infected intravenously with 2 × 106 ffu of LCMV Armstrong or LCMV clone 13 where indicated . GITR-/- mice were kindly provided by Dr . C . Riccardi ( University of Perugia ) and Dr . P . Pandolfi ( Harvard Medical School ) . GITR-/- mice were analyzed by SNP analysis across 1500 SNPs differing in the 129 and B6 genome and found to be a minimum of ninety two per cent C57BL/6 ( TCAG , University of Toronto ) . Effects of GITR on CD4 and CD8 T cell effector responses were similar when compared to commercial C57BL/6 mice or when we used F1 or F2 littermate controls . However , the proportion of Tregs in the mice after infection differed in commercial mice versus littermates , therefore all Treg data are reported from experiments with littermate controls . DEREG [49] , now available from Jackson Laboratories , P14 [27] and SMARTA [34] mice were obtained from Dr . P . Ohashi ( Princess Margaret Hospital , University of Toronto ) and additionally crossed to GITR-/- CD45 . 1 mice and compared with their F2 littermates . All animals were housed under specific pathogen free conditions at the Terrence Donnelly Centre for Cellular and Biomolecular Research ( University of Toronto ) . Where indicated , CD4 T cells were depleted by administering 0 . 5mg of purified anti-CD4 ( clone GK1 . 5 ) two days before LCMV cl 13 infection , or IL-2 was blocked by administering 0 . 5mg anti-IL-2 ( clone S4B6–1 ) given at days four and six p . i . All animal procedures were approved by the animal care committee of the University of Toronto ( Protocol approval number: 20009988 ) in accordance with the Canadian Council on Animal care , a National Regulatory body through which the University of Toronto holds a certificate of good animal care ( http://www . ccac . ca/en_/assessment/certification ) . Splenocytes were cultured in complete medium with Brefeldin A and monensin ( BD , Cat . #554715 ) with 1μg/mL Db-restricted peptides: NP396–404 , GP33–41 , or GP276–286 ( Anaspec , Fremont , CA ) for five hours , or 4μg/mL I-Ab-restricted GP61–80 with 20U/mL rmIL-2 ( eBioscience ) for five hours . Cells were then surface stained , fixed and permeabilized ( BD ) , and stained for intracellular cytokine production . Organs were harvested and immediately placed on dry ice . Organs were later thawed , homogenized , and supernatant dilutions ( range: 100–105 ) were used to infect an MC57 cell monolayer under a 2% methylcellulose-MEM overlay . 48 hours later , monolayers were fixed with 4% PFA , permeabilized with 1% Triton X-100 , and stained with Rat anti-LCMV mAb ( clone VL-4 ) . Following secondary Goat anti-Rat-HRP , a colorimetric reaction with o-Phenylenediamine ( Sigma-Aldrich ) was used to quantify LCMV-infected foci . LCMV-specific IgG ELISAs were performed as previously described [50] , with modification to the purification of LCMV . LCMV was purified on a CsCl gradient of 1 . 45g/L and 1 . 2g/L in 10mM Tris ( pH 8 . 1 ) . Plates were coated with 10ng of inactivated purified LCMV cl 13 and blocked with Superblock in TBST ( Thermo Fisher Scientific ) , followed by one-hour incubation with dilutions of plasma from infected animals . 1:2 , 000 HRP-tagged anti-mouse secondary antibody was then added and incubated for one hour . 3 , 3’ , 5 , 5’-tetramethylbenzidine ( Sigma Aldrich ) was used for colorimetric reaction and absorbance was measured at 456nm . Purification of CD8 T cells for P14 adoptive transfers or CD4 T cells for SMARTA adoptive transfers were achieved with negative selection kits according to manufacturer’s protocols ( StemCell Technologies , Vancouver , BC , Cat . 19753A and 19752 , respectively ) . For P14 and SMARTA transfers , cells were transferred intraperitoneally one day prior to LCMV infection , and where indicated , were labeled with 10μM Carboxyfluorescein Diacetate Succinimidyl Ester ( CFSE ) . GITR+/+: GITR-/- mixed bone marrow chimeras were generated by intravenously reconstituting lethally irradiated C57BL/6 mice with a 1:1 mixture of GITR-/- CD45 . 1: GITR+/+ CD45 . 2 bone marrow cells for a total of 5 × 106 cells . Following irradiation and reconstitution , mice were given 2mg/mL neomycin sulfate ( Bio-Shop , Burlington , ON ) for two consecutive weeks , after which mice were rested for an additional 90 days before use . For in vitro studies , SMARTA splenocytes were stimulated with 4μg/mL I-Ab-restricted GP61–80 with 40U/mL recombinant murine IL-2 in complete RPMI containing 10% FCS for two days . Cells were rested overnight in serum-free Aim V medium with 1X β-Mercaptoethanol ( Invitrogen ) . Live SMARTA cells were stimulated with 10μg/mL anti-GITR ( DTA-1 , hybridoma provided by S . Sakaguchi , Osaka University ) or Rat IgG isotype control ( Jackson ) each followed by 5μg/mL Goat anti-Rat ( Jackson ) . Samples were prepared for flow cytometry using BD Phosflow Buffer I ( BD , Cat . #557870 ) . Similar protocols were used for direct ex vivo staining of freshly isolated splenocytes at day three p . i . Samples were acquired with FACS Canto , LSR II , or LSR Fortessa ( BD Biosciences ) with FACSDiva software . Flow cytometry data were analyzed with FlowJo v9 ( Tree Star , Inc . , Ashland , OR ) . All statistical analyses were performed using GraphPad Prism v6 ( La Jolla , CA ) . Unpaired Student’s t-tests were used to compare two groups , with P values indicated on figures: * P < 0 . 05 , ** P < 0 . 01 , *** P < 0 . 001 , **** P < 0 . 0001 .
|
The natural rodent pathogen LCMV clone 13 causes a persistent viral infection in mice and has successfully predicted several immunological factors that are relevant to human chronic viral infection such as HIV . LCMV clone 13 infection is ultimately controlled by cell-mediated and humoral immune responses by day 60–90 post-infection in CD4 T cell-sufficient mice . While it has been known for several years that CD4 T cell help is critical for control of LCMV clone 13 , research to date has been largely limited to the regulatory factors that contribute to late CD4 T cell dysfunction , with little knowledge of the role of T cell co-stimulatory factors in sustaining CD4 T cells to help cell-mediated and humoral immune responses . Using GITR-deficient mice , we show that the co-stimulatory molecule GITR plays a critical cell-intrinsic role in early CD4 T cell accumulation to support cytotoxic T cell responses and late LCMV-specific IgG production . The early effects of Th1 on CTL responses are IL-2-dependent . Mice lacking GITR have up to 35-fold higher viral burden relative to GITR-sufficient controls . Taken together , we demonstrate a critical cell-intrinsic role for GITR is sustaining CD4 T cell responses to control chronic LCMV infection . Thus , GITR on CD4 T cells may critically contribute to the initial viral set-point in infections such as HIV .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
GITR Intrinsically Sustains Early Type 1 and Late Follicular Helper CD4 T Cell Accumulation to Control a Chronic Viral Infection
|
The centrality-lethality rule , which notes that high-degree nodes in a protein interaction network tend to correspond to proteins that are essential , suggests that the topological prominence of a protein in a protein interaction network may be a good predictor of its biological importance . Even though the correlation between degree and essentiality was confirmed by many independent studies , the reason for this correlation remains illusive . Several hypotheses about putative connections between essentiality of hubs and the topology of protein–protein interaction networks have been proposed , but as we demonstrate , these explanations are not supported by the properties of protein interaction networks . To identify the main topological determinant of essentiality and to provide a biological explanation for the connection between the network topology and essentiality , we performed a rigorous analysis of six variants of the genomewide protein interaction network for Saccharomyces cerevisiae obtained using different techniques . We demonstrated that the majority of hubs are essential due to their involvement in Essential Complex Biological Modules , a group of densely connected proteins with shared biological function that are enriched in essential proteins . Moreover , we rejected two previously proposed explanations for the centrality-lethality rule , one relating the essentiality of hubs to their role in the overall network connectivity and another relying on the recently published essential protein interactions model .
An intriguing question in the analysis of biological networks is whether biological characteristics of a protein , such as essentiality , can be explained by its placement in the network , i . e . , whether topological prominence implies biological importance . One of the first connections between the two in the context of a protein interaction network , the so-called centrality-lethality rule , was observed by Jeong and colleagues [3] , who demonstrated that high-degree nodes or hubs in a protein interaction network of Saccharomyces cerevisiae contain more essential proteins than would be expected by chance . Since then the correlation between degree and essentiality was confirmed by other studies [4]–[7] , but until recently there was no systematic attempt to examine the reasons for this correlation . In particular , what is the main topological determinant of essentiality ? Is it the number of immediate neighbors or some other , more global topological property that essential proteins may have in a protein interaction network ? Jeong and colleagues [3] suggested that overrepresentation of essential proteins among high-degree nodes can be attributed to the central role that hubs play in mediating interactions among numerous , less connected proteins . Indeed , the removal of hubs disrupts the connectivity of the network , as measured by the network diameter or the size of the largest connected component , more than the removal of an equivalent number of random nodes [3] , [8] . Therefore , under the assumption that an organism's function depends on the connectivity among various parts of its interactome , hubs would be predominantly essential because they play a central role in maintaining this connectivity . Recently , He and colleagues challenged the hypothesis of essentiality being a function of a global network structure and proposed that the majority of proteins are essential due to their involvement in one or more essential protein–protein interactions that are distributed uniformly at random along the network edges [9] . Under this hypothesis , hubs are proposed to be predominantly essential because they are involved in more interactions and thus are more likely to be involved in one which is essential . In this work we carefully evaluate each of the proposed explanations for the centrality-lethality rule . Recently several hypotheses that linked structural properties of protein interaction networks to biological phenomena have come under scrutiny , with the main concern being that the observed properties are due to experimental artifacts and/or other biases present in the networks and as such lack any biological implication . To limit the impact of such biases on the results reported in our study we use six variants of the genomewide protein interaction network for Saccharomyces cerevisiae compiled from diverse sources of interaction evidence [10]–[15] . To assess whether the essentiality of hubs is related to their role in maintaining network connectivity we performed two tests . First , if this were the case , then we would expect essential hubs to be more important for maintaining network connectivity than nonessential hubs . We found that this is not the case . Next , in addition to node degree , we consider several other measures of topological prominence , and we demonstrate that some of them are better predictors of the role that a node plays in network connectivity than node degree . Thus , if essentiality were related to maintaining network connectivity , then one would expect essentiality to be better correlated with these centrality measures than with the node degree . However , we found that node degree is a better predictor of essentiality than any other measure tested . To reject the essential protein interaction model [9] , we used a hypothesis testing approach . Namely , we observed that this model implies that the probability that a protein is essential is independent of the probability that another noninteracting protein is essential . However , in the tested networks the essentiality of noninteracting proteins that share interaction partners is correlated . Thus , we reject the independence assumption and , as a result , the essential protein interaction model with high confidence . Motivated by our findings we propose an alternative explanation for the centrality-lethality rule . Our explanation draws on a growing realization that phenotypic effect of gene-knockout experiments is a function of a group of functionally related genes , such as genes whose gene products are members of the same multiprotein complex [16] . It is well known that densely connected subnetworks are enriched in proteins that share biological function . Therefore , one would expect that dense subnetworks of protein interaction networks should be either enriched or depleted in essential proteins . Indeed , Hart and colleagues observed that essential proteins are not distributed evenly among the set of automatically indentified multiprotein complexes [17] . In this work we observe that the same phenomenon holds for potentially larger groups of densely connected and functionally related proteins , which we call COmplex BIological Modules ( or COBIMs ) . We demonstrate that due to the uneven distribution of essential proteins among COBIMs the majority of the essential proteins lie in those COBIMs that are enriched in essential proteins , which we call Essential COmplex BIological Modules ( or ECOBIMs ) . By the very definition , ECOBIMs contain , relative to their size , more essential nodes than a random group of proteins of the same size . But what fraction of all essential hubs are members of such ECOBIMs ? How does this number relate to what is expected by chance ? In fact , how does the enrichment of hubs that are members/nonmembers of ECOBIMs in essential proteins relate to the enrichment values expected by chance under a suitable randomization protocol ? We propose that membership in ECOBIMs largely accounts for the enrichment of hubs in essential proteins . In support of this hypothesis , we found that the fraction of essential proteins among non-ECOBIM hubs is , depending on the network , only 13–35% , which is almost as low as the network average . Furthermore the essentiality of nodes that are not members of ECOBIMs is only weakly correlated with their degree . Finally , using a randomization experiment we demonstrated that these properties are characteristic of the protein interaction network and are unlikely in a corresponding randomized network .
Our source of protein interaction data for the yeast Saccharomyces cerevisiae is numerous small-scale studies and seven high-throughput experiments [15] , [18]–[23] . Interactions reported in targeted studies are believed to be biologically relevant as they are usually subjected to a variety of validation methods . Recently , Reguly et al . [11] curated about 30 , 000 literature abstracts to compile a network of protein interactions reported in small-scale experiments . We refer to this network as the LC network ( Literature Curated network ) . It was suggested that the centrality-lethality phenomenon is an artifact of a possible bias present in the networks mainly derived from small-scale experiments [24] . Namely , essential proteins are the focus of more studies and therefore tend to have a higher degree in these networks . Therefore , to complement the LC network , we included in our study two networks that contain interactions reported in both small-scale studies and high-throughput experiments . The DIP CORE network is derived from the pool of protein interactions deposited in the DIP database using a computational method of Deane et al . [10] that recruits evolutionary information to filter out unreliable interactions . The HC network ( the High Confidence network ) recently published by Batada et al . [12] is derived by intersecting small-scale data with the above-mentioned seven high-throughput datasets . More specifically , an interaction is included in the final network only if it was independently reported at least twice . We also include two networks derived solely from high-throughput experimental data . The Y2H network is obtained from the genomewide yeast-two-hybrid interaction screen of Ito et al . [15] and contains high-confidence interactions that were experimentally detected at least three times . Recently , Collins et al . [13] published a statistical scoring scheme that maps raw complex purification experimental data to interaction confidence scores . The authors applied their method to raw purification data from two recent genomewide complex purification experiments [22] , [23] . We refer to a network that contains all interactions with a confidence score above a certain threshold as the TAP-MS network . Finally , we include a network of interactions predicted in silico using the computational approach of Jansen et al . [14] . The method trains a Bayesian network that combines a variety of genomic features such as mRNA coexpression , colocalization , etc . to derive interaction confidence scores for protein pairs . The authors used protein interactions derived from a set of manually curated protein complexes as the set of positive training examples and pairs of proteins localized to different cellular compartments as the set of negative training examples . We refer to this network as the BAYESIAN network . Table 1 summarizes the structural properties of the six networks just described . ( Here and throughout the paper we analyze the largest connected component of each protein interaction network . ) Table 2 shows the overlap , fraction of interactions in common , between the networks . Given the differences in the experimental techniques used to construct these networks and the fact that the edges in the TAP-MS and BAYESIAN networks correspond to membership in multiprotein complexes , in the Y2H to physical contacts , and in the DIP CORE , LC , and HC networks to a mix of these two things , it is not surprising that the networks differ significantly in terms of density , cliquishness , and other parameters . The biggest outlier is the Y2H network . In fact , for this network , the relation between essentiality and lethality is less prominent as discussed in the next section . In their influential paper , Jeong et al . [3] observed that the degree of a node in a yeast protein interaction network correlates with the phenotypic effect of its deletion . More specifically , the authors observed that high-degree nodes are three times more likely to be essential than nodes having few interaction partners . It was further hypothesized that high-degree nodes tend to be essential due to the central role that they play in maintaining the overall connectivity of the network by mediating interactions among other less connected proteins . Consequently , high-degree nodes are also referred to as hubs , and the observed phenomenon is known as the centrality-lethality rule . To confirm the centrality-lethality rule in the tested networks we used the results of a systematic gene deletion screen [25] in which 1 , 105 yeast genes were found to be essential for growth on rich glucose media . There are numerous ways of exposing positive correlation between degree and essentiality , two of which are used in this paper . First , one can ask whether hubs , nodes with a degree greater than or equal to a certain threshold , are more likely to be essential than an average network node , i . e . , whether the fraction of essential proteins among hubs is greater than the network average . To choose an appropriate threshold value we relied on Figure 1A , which shows the enrichment values for nodes with a degree greater than or equal to k as a function of k . In some networks the steady increase of enrichment values is interrupted for very large values of k . Therefore , we chose the threshold value so that approximately 20% of the network nodes are hubs . ( For the DIP CORE network the value of k is 7 , for the LC network it is 10 , for the HC network it is 10 , for the TAP-MS network it is 24 , for the BAYESIAN network it is 12 , and for the Y2H network it is 3 . ) However , we repeated the experiments with hubs defined as 10% ( data not shown ) and found that our conclusions are robust to the specific choice of the threshold . From Figure 1A it is clear that the enrichment values increase with k . Therefore , one can use a nonparametric measure of association , such as Kendall's tau and Spearman's rho rank correlation coefficients [26] , to assess the correlation between degree and essentiality over all network nodes . As shown in Figure 1B these two measures agree in their estimates of the strength of the correlation; therefore all further evaluations were done with the Kendall's tau rank correlation coefficient . Then , to assess the correlation between other centrality measures and essentiality after correcting for correlation with degree , we used a partial Kendall's tau rank correlation . It should be noted that in contrast to other networks the Y2H network exhibits only a weak correlation between degree and essentiality . This is in agreement with the study of Batada et al . [4] . They observed a highly significant difference in the average degree of essential and nonessential proteins in the LC network but found that the difference almost disappears when the analysis is restricted to interactions detected by only the yeast-two-hybrid experiments . A network centrality index assigns a centrality value to each node in the network that quantifies its topological prominence . Topological prominence can be defined in a number of ways , and over the years many centrality indices were introduced that emphasize different aspects of network topology [27] . In a local centrality index , the node's centrality value is mainly influenced by the topology of its local neighborhood . A well known example of a local centrality index is degree centrality , where the node's centrality value is equal to the number of its immediate neighbors . Betweenness indices , on the other hand , assign centrality values based on the node's role in maintaining the connectivity between pairs of other nodes in the network . A well-known example of a betweenness centrality index is shortest-path betweenness centrality , where the node's centrality value is proportional to the fraction of shortest paths that pass through it . Even though degree centrality is a local centrality index , in some networks hubs may play an important role in maintaining the overall connectivity of the network . For example , it was demonstrated that in some scale-free networks the removal of hubs affects the ability of other nodes to communicate much more than the removal of random nodes [8] . To clarify the topological role of hubs in the tested networks , we compared degree centrality to two other local indices ( eigenvector centrality ( EC ) [28] and subgraph centrality ( SC ) [29] ) and to two betweenness indices ( shortest-path betweenness centrality ( SPBC ) [30] and current-flow betweenness centrality ( CFC ) [31] ) . ( See Figure 2 for an illustration , and Materials and Methods for a more detailed description of the centrality measures used in this study . ) Since betweenness indices rank nodes based on their role in mediating communication between pairs of other nodes in the network , it is interesting to compare the effectiveness of high-degree nodes and nodes with high betweenness centrality in disconnecting the network . One common way to measure the impact of the nodes' removal on the network connectivity is by monitoring the decrease in the size of the largest connected component . Figure 3A–F shows , for the six protein interaction networks , how the removal of the most central nodes , random nodes , and essential proteins affects the network connectivity . As expected , removing nodes with high local centrality values is much less disruptive than removing those with high betweenness centrality values . Interestingly , degree centrality is as efficient in shattering the network as betweenness in the DIP CORE , LC , and Y2H networks , is as inefficient as the local indices in the TAP-MS network , and is somewhere between the local and betweenness indices in the HC and BAYESIAN networks . The local measures strongly agree in their ranking of network nodes in all networks except the Y2H network . The agreement is the strongest in the TAP-MS network; as a result the curves for the EC and SC measures overlap completely in Figure 3D . While the removal of a set of nodes may not disconnect various parts of the network , it may impair significantly the “quality of communication” between them . For example , there can be an increase in the length of the shortest path or decrease in the number of alternative paths between pairs of nodes in the network . Therefore , we introduced two additional measures , which we call network integrity measures , to capture various aspects of the effect of the nodes' removal on the ability of other nodes to communicate . ( See Materials and Methods for a description of the network integrity measures . ) We find that even when these more sensitive measures are used the observations made above about the disruptive power of hubs relative to other most central proteins hold ( Table S1 ) . Next , we examined whether the disruption power of hubs comes mainly from essential hubs . First , we observe that the removal of all essential proteins from the huge connected component is less disruptive than the removal of an equivalent number of the most central nodes according to any index ( Figure 3A–F ) . Moreover , as shown in Table 3 , the removal of essential nodes is not more disruptive than the removal of an equivalent number of random nonessential nodes that have the same degree distribution . We conclude that even though in most networks , the DIP CORE , LC , HC , and Y2H networks , the removal of high-degree nodes is disruptive , this disruption is not related to the essentiality of these nodes . On the contrary , essential genes are indistinguishable in that respect from the random nonessential genes with the same degree distribution . Above we demonstrated that various centrality indices vary considerably in their ability to predict disruption in the overall connectivity of the network . Next we asked whether this difference is reflected in the enrichment levels . Figure 4 shows the fraction of essential proteins among hubs and an equivalent number of most central proteins according to five centrality measures . We observe that the local centrality indices have enrichment levels comparable to those of betweenness indices and in some cases even higher . But most notably , degree centrality fares better than any other centrality index in five networks but is narrowly beaten by shortest-path centrality for the Y2H network . The superiority of degree centrality is even more apparent when Kendall's tau rank correlation coefficient is used to measure correlation between centrality values and essentiality over all network nodes ( compare Table 2 to Table 4 ) . As there is considerable correlation between degree centrality and other centrality indices , we used Kendall's tau partial rank correlation coefficient to see whether any of the indices is correlated with essentiality beyond its correlation with degree centrality index . We found that , controlling for the correlation with degree , the correlation with essentiality is reduced to statistically insignificant values for betweenness centrality indices and is greatly reduced for local indices ( Table 4 ) . The above observations indicate that the main topological determinant of essentiality is the node's local neighborhood rather than its role in maintaining the overall connectivity of the network . In particular , even though removing the nodes with high betweenness centrality indices is much more effective in shattering some of our protein interaction networks , their correlation with essentiality is reduced to statistically insignificant levels by subtracting their correlation with degree centrality . Recently He and colleagues [9] proposed an explanation for the centrality-lethality rule in terms of essential protein interactions: a protein is essential either due to its involvement in one or more essential protein interactions or due to other factors . The authors argue that the determination of protein essentiality in the protein interaction network can be captured by a simple random process: ( i ) distribute essential protein interactions along the edges of the network uniformly at random with probability α; ( ii ) distribute essential proteins among the nodes of the network uniformly at random with probability β . Thus , according to the model , the probability ( PE ) of a protein with k neighbors being essential is PE = 1− ( 1−α ) k ( 1−β ) , and the natural logarithm of the fraction of nonessential proteins among proteins of degree k has a linear dependency on k: log ( 1−PE ) = log ( 1−α ) k+log ( 1−β ) . We note that from the assumptions of the essential protein interaction model it follows that if two proteins do not interact then the essentiality of one protein in such a pair does not depend on the essentiality of the other protein . Furthermore , this independence should also be observed when proteins share interaction neighbors . To test whether this holds in real data , we computed the number of nonadjacent protein pairs , with three or more neighbors ( one or more neighbors in the Y2H network ) , that are either both essential or both nonessential in the tested networks and compared these numbers to the expected number of such pairs under the model . ( The model parameters were estimated using three different strategies as described in the Materials and Methods . In their paper , He et al . point out that their model may not work in networks where the edges represent membership in the same protein complex . Thus , we excluded the TAP-MS and BAYESIAN networks from the analysis . ) As shown in Table 5 , the model does not capture the correlation in essentiality observed in the tested networks; i . e . , there is a statistically significant difference between the number of such pairs observed in real data and the number expected under the model . Consequently , the essential interaction model is rejected with high confidence . In the previous section we showed that proteins that share neighbors are more likely to have the same essentiality ( be both essential or both nonessential ) than expected under the essential PPI model . Moreover , it was observed in another study that essential proteins are not distributed uniformly among in the set of automatically derived multiprotein complexes [17] . This suggests that densely connected subnetworks are polarized toward being either highly enriched or significantly depleted of essential proteins . Furthermore , it is well known that densely connected subnetworks are enriched in proteins that share biological function . Therefore , one should expect that protein interaction networks contain densely connected functional modules that are highly enriched in essential proteins . Some large multiprotein complexes , for example , those involved in transcription regulation , are known to be highly enriched in essential proteins , but how general is this phenomenon and can it account for the centrality-lethality rule ? To investigate the above question , we introduce a notion of Essential Complex Biological Modules , which are groups of proteins with shared biological function that extensively interact with each other and are enriched in essential proteins . First , we describe an automatic method for the extraction of ECOBIMs from a protein interaction network . Next , we argue that the membership in ECOBIMs accounts to large extent for the centrality-lethality rule in the tested networks . Finally , we address statistical issues related to our selection procedure by applying suitable randomization protocols . We developed an automatic method for extraction of ECOBIMs from a protein interaction network . In this work proteins are deemed to share biological function if they are annotated with the same GO biological process term from a set of 192 terms that were selected by a group of experts to represent relevant aspects of molecular biology [32] . Therefore , our method is applied to subnetworks induced by proteins annotated with the same GO biological process term , one subnetwork at a time . The high-level idea behind the method is to first identify groups of densely connected proteins , which we call Complex Biological Modules ( or COBIMs ) , and then identify a subset of COBIMs as ECOBIMs based on the distribution of essential proteins among the COBIM nodes . More specifically , our heuristic selects a subset of COBIMs that are enriched in essential proteins . ( The method is schematically shown in Figure 5 and is described in detail in the Materials and Methods section . Figure S1 shows the fraction of nodes that are members of r or more COBIMs for various values of r . ) To examine to what extent the membership in ECOBIMs accounts for the centrality-lethality rule we partitioned hubs into two groups , those that are members of one or more ECOBIMs ( ECOBIM hubs ) and those that are not ( non-ECOBIM hubs ) , and compared their enrichment values . As shown in Figure 6 ECOBIM hubs are highly enriched in essential proteins , whereas non-ECOBIM hubs are depleted in essential proteins as compared to the network average enrichment values . But most importantly , as discussed in the next paragraph , the difference in the fraction of essential proteins among ECOBIM hubs and non-ECOBIM hubs is not a result of our greedy ECOBIM selection procedure or particular degree sequence of essential proteins in the network . We next asked whether there is a correlation between degree and lethality for network nodes that are not members of the ECOBIMs . As shown in Table 6 the correlation between essentiality and degree for non-ECOBIM nodes is much less than that for all network nodes . One may ask to what extent the difference in the behavior of ECOBIM hubs and non-ECOBIM hubs is due to the particular selection procedure that we employ to identify the putative ECOBIMs . More specifically , there are two concerns that need to be addressed . First , our method is guided by the enrichment in essential proteins when selecting ECOBIMs from COBIMs . Therefore , it is expected that the fraction of essential proteins among ECOBIM hubs should be higher than that among non-ECOBIM hubs . Second , our method considers only annotated yeast genes . Therefore , one might argue that the difference in behavior is due to the fact that ECOBIM hubs are necessarily annotated while non-ECOBIM hubs may include both annotated and unannotated genes . To address the first concern we performed a control experiment where essential proteins were assigned to a random set of nodes having the same degree distribution as the true set of essential proteins in the network . ( A total of 100 , 000 random assignments were performed , which resulted in 100 , 000 sets of ECOBIMs . ) To address the second concern , we restricted the random assignment to annotated genes only . As shown in Table 6 , the ECOBIMs resulting from the true assignment of essential proteins have dramatically different properties than these resulting from the random assignment of essential proteins . In particular , the fraction of essential proteins among non-ECOBIM hubs under the true assignment of essential proteins is significantly lower than that under the randomized assignment of essential proteins , even though the same selection procedure is used in both cases . Therefore , we conclude that the observed difference is the result of the particular distribution of essential proteins among the nodes of the network and not an artifact of our selection procedure . The same holds for the reduction in correlation between degree and essentiality for non-ECOBIM nodes . The identified ECOBIMs mostly correspond to large essential multiprotein complexes such as the anaphase promoting complex ( APC ) and the DAM1 protein complex but not exclusively complexes . For example , one of the largest ECOBIMs identified in the LC network contains multiprotein complexes involved in the process of RNA polymerase 2 transcription [33] , such as RNA polymerase 2 , general transcription factors , the mediator complex , etc . The ECOBIMs with at least 20 members are shown in Table 7; all ECOBIMs and their member proteins are given in Table S2 . Moreover , the ECOBIMs are remarkably different than non-ECOBIM COBIMs . As shown in Table 8 , the distribution of essential proteins among the COBIM nodes is highly uneven . In particular , the observed difference between fractions of essential proteins among the ECOBIM nodes and among non-ECOBIM COBIM nodes can not be accounted for neither by degrees of essential COBIM nodes nor by the particular ECOBIM selection procedure . The last claim is validated by performing 100 , 000 randomized assignments of essential proteins that preserve degrees and the number of essential COBIM nodes , selecting the ECOBIMs and computing the corresponding fractions . As shown in Table 8 , the values obtained under the true assignment of essential proteins are significantly different from those obtained under the randomized assignment of essential proteins . So far , we demonstrated that the high correlation between degree and essentially can be predominantly attributed to the ECOBIMs . In addition , it is well known that certain functions that are essential to the cell , for example , transcription regulation or cell-cycle regulation , rely on large multiprotein complexes . Indeed , many of the GO terms that are overrepresented among ECOBIM nodes are of this type , as seen in Figure 7 . Do ECOBIMs play a distinguished role in those essential processes , or are they merely a byproduct of the above-mentioned observation ? In particular , is the difference in the enrichment in essential proteins exclusively due to the fact that some essential GO processes contain ECOBIMs while others do not ? To elucidate the role of the ECOBIMs we examined all GO processes that contain at least one ECOBIM . Table 9 shows the results for the DIP core network sorted by the percentage of essential proteins in a given GO process . ( The data for the other networks are given in Table S3 . ) Observe that the enrichment of ECOBIMs in essential genes is typically much higher than the average enrichment in the corresponding GO . Thus , the ECOBIMs are not merely representatives of the average structure of the corresponding GO subnetwork . The uneven distribution of essential proteins is also observed even when the corresponding GO process is extremely enriched in essential proteins such as rRNA metabolic process ( GO:0016072 ) or transcription initiation ( GO:0006352 ) . The percentage of essential proteins among network nodes annotated to either one of these two processes is more than 80% , and all COBIMs are selected as ECOBIMs . The process with the next highest percentage of essential proteins , transcription from RNA polymerase III promoter ( GO:0006383 ) , contains both types of COBIMs . Interestingly , for this process , all ECOBIM nodes are essential , but none of the remaining COBIM nodes is . In fact , if a GO process contains both ECOBIM and non-ECOBIM COBIMS , then such polarization is frequent albeit rarely that extreme ( Table 9 ) . Removal of any protein from a Complex Biological Module is expected to perturb or even disable the whole module . Thus , within a large spectrum of essential GO processes , a cell can tolerate large perturbations of some modules but very little perturbations of ECOBIMs . This last observation can also explain the poor correlation between degree and essentiality in Y2H networks , as it indicates that ECOBIMs are likely to contain large , stable multiprotein modules , typically multiprotein complexes . However , interactions recovered by the Y2H technique correspond to physical contacts and as such do not encompass all members of a complex . Moreover , due to its binary nature , the Y2H technique may completely miss interactions in complexes that require cooperative binding [34] .
The enrichment of high-degree nodes in essential proteins , known as the centrality-lethality rule , suggests that the topological prominence of a protein in a protein interaction network may be a good predictor of its biological importance . There exist numerous measures of topological prominence , called network centrality indices; local centrality indices assign centrality values based on the topology of the node's local neighborhood , whereas betweenness centrality indices assign centrality values based on the node's role in maintaining the connectivity between pairs of other nodes in the network . Even though by definition degree centrality is a local measure , depending on the structure of the network , hubs may play an important role in maintaining the overall connectivity of the network . In this paper we sought to identify the main topological determinant of essentiality and to give a biological explanation for the connection between the network topology and essentiality . To address this question we performed a rigorous analysis of six protein interaction networks for Saccharomyces cerevisiae compiled from diverse sources of interaction evidence . To clarify the topological roles of essential proteins in general and essential hubs in particular , we compared degree centrality to other local and betweenness centrality indices . We found that while in some networks high-degree nodes are as important in maintaining the overall network connectivity as nodes having high betweenness centrality values , this property is not due to essential proteins . On the contrary , essential proteins are indistinguishable in that respect from nonessential proteins having the same degree distribution . We also found that degree centrality is a better predictor of essentiality than any other measure tested and that correlation of betweenness indices with essentiality is entirely due to their correlation with degree centrality . Thus , we conclude that the topological determinant of essentiality is the node's local neighborhood rather than its role in maintaining the overall connectivity of the network . Next we examined whether the essential interactions model , recently proposed to explain the centrality-lethality rule , is valid in the tested networks . We found that the model's central assumption that the majority of proteins are essential due to their involvement in one or more essential protein interactions , which are distributed uniformly at random along the edges of the network , violates basic clustering patterns of essential proteins in the networks that we examined . The uniform distribution of essential protein interactions implies that , as long as two proteins do not interact , the essentiality of one protein in the pair is independent of the essentiality of the other protein . However , in real protein interaction networks the essentiality of pairs of proteins that share many neighbors is correlated , and the number of nonadjacent protein pairs that share three or more neighbors and are either both essential or both nonessential significantly deviates from the expected number of such pairs under the model . Consequently , we rejected the essential interactions explanation with high confidence . We stress that we do not reject the existence of essential protein interactions but rather the assumption that these interactions are evenly distributed along the edges of the network and explain the degree distribution of essential proteins . The above observations led us to propose an alternative explanation for the centrality-lethality rule . Our explanation builds on a growing body of evidence that gene knock-out phenotypes for genes whose gene products are members of the same multiprotein complex are correlated [16] , [17] . In particular , Hart et al . demonstrated that essential proteins are not distributed evenly among the set of automatically identified multiprotein complexes; rather there are “surprisingly” many complexes where the majority of members are essential and “surprisingly” many complexes where the majority of members are not essential [17] . Here we hypothesized and then computationally confirmed that the same phenomenon holds for potentially larger groups of densely connected and functionally related proteins that we called Complex Biological Modules and abbreviated as COBIMs . But more importantly , we were able to demonstrate that membership in ECOBIMs , those COBIMs that are enriched in essential proteins , provides a good explanation for the correlation between degree and essentiality in the protein interaction networks considered in this study . In particular , we showed that non-ECOBIM hubs are depleted in essential proteins and for non-ECOBIM proteins the correlation between degree and essentiality is greatly reduced . Moreover , by applying suitable randomization protocols we showed that the different characteristics of ECOBIM and non-ECOBIM hubs ( or in general ECOBIM and non-ECOBIM proteins ) are not a mere consequence of their degrees or the particular computational method that we adopted for selecting the ECOBIMs . In the past , several attempts were made to classify high-degree nodes using additional biological data to obtain a deeper insight into biological and physiological properties that hubs were reported to possess . Here we discuss how our findings fit the results reported in two such studies [35] , [36] . Han et al . utilized mRNA expression data to classify hubs into party and date hubs , where the party hubs show a significant agreement in the mRNA expression levels , or are coexpressed , with their interacting partners , whereas the date hubs are not coexpressed with their neighbors [35] . The removal of the date hubs was observed to shatter the network much more efficiently than the removal of party hubs . On the basis of this and other observations made in the paper , the date hubs were proposed to “…participate in a wide range of integrated connections required for the global organization of biological modules in the whole proteome network…” However , the fraction of essential proteins among the party hubs was even slightly higher than that among the date hubs . This is consistent with one of the conclusions made in this paper , namely , essentiality is not a byproduct of the node's ability to maintain the overall connectivity of the network . Furthermore , it has been proposed that “party hubs represent integral elements within distinct modules” and “tend to function at a lower level of the organization of the proteome” [35] . Such a description is consistent with the properties COBIM hubs where COBIMs hubs are explicitly defined as hubs that are members of highly connected modules . Similarly to the party hubs , the average enrichment of COBIM hubs in essential proteins is slightly higher than that of non-COBIM hubs ( data not shown ) . We also demonstrated that essential proteins clearly cluster within ECOBIMs rather than being uniformly distributed over all COBIMs . In the second study Kim et al . utilized structural data to classify hubs into singlish-interface and multiinterface hubs , where singlish-interface hubs would interact with their partners through one or two distinct interfaces , whereas the multiinterface hubs would interact with their partners through three or more distinct interfaces [36] . In this case , however , the classification produced significantly different enrichment levels , with a multiinterface hub being twice as likely to be essential as a singlish-interface hub or an average network node . The authors suggested that multiinterface hubs most likely correspond to members of large and stable multiprotein complexes . Consequently , this would imply that stable multiprotein complexes are enriched in essential proteins . This view is consistent with the results of this paper with additional caveats as discussed below . It is well known that certain biological functions essential for the cell depend on large multiprotein complexes . ( Consider , for example , RNA Polymerase II transcription machinery [33] or ribosome biogenesis and assembly [37] . ) Indeed , many ECOBIMs indentified by our approach are associated with such processes . However , even within such essential processes , ECOBIMs distinguish themselves as being more enriched in essential proteins than the remaining proteins within the same process . The enrichment in essential proteins of non-ECOBIM COBIMs is usually at the same level and frequently significantly lower than the average enrichment within the corresponding GO process . Thus , within a large spectrum of essential GO processes , a cell can tolerate large perturbations of non-ECOBIM modules but very little perturbation of ECOBIMs . Some COBIMs do not contain any essential proteins . In such a case , the whole module can be nonessential , and the fact that a cell can tolerate the removal of any of member of such a COBIM does not exclude the possibility that this COBIM corresponds to a stable complex .
In this work we compare the degree centrality measure to two other local measures ( eigenvector centrality ( EC ) [28] and subgraph centrality ( SC ) [29] ) and to two betweenness measures ( shortest-path betweenness centrality ( SPBC ) [30] and current-flow betweenness centrality ( CFC ) [31] ) . The computation of the eigenvector centrality values can be cast as an iterative process: ( i ) start with an initial vector of centrality scores ; ( ii ) in iteration k+1 update the centrality score of a node i using the scores of its neighbors from the previous iteration and then normalize the scores . It can be shown that this process converges to the eigenvector that corresponds to the largest eigenvalue of the adjacency matrix of the network . The subgraph centrality value of a node is equal to the number of closed walks that start and terminate at the node . As there is an infinite number of such walks , to obtain finite index values the number of closed walks of length k is weighted by 1/k ! . Therefore , short walks dominate the subgraph centrality values . For the shortest-path betweenness index , the node's centrality value is equal to the average fraction of shortest paths that pass through the node . The current-flow centrality measure extends the shortest-path centrality measure by taking into account other paths in addition to shortest paths . This is achieved through a current-flow paradigm where the network is viewed as a resistor network with each edge having a unit capacity . For every pair of nodes s and t , one unit of current is shipped from s to t , and the centrality of a node is set to the average amount of current that passes through that node . We demonstrate the difference between the five centrality measures on a toy network in Figure 1A . In this network two cliques K50 and K10 are interconnected by an edge ( A1 , B1 ) and through a node D . The nodes of K50 are labeled A1…A50 , and the nodes of K10 are labeled B1…B10 . An additional node C attaches to K50 through A2 . Figure 1B shows the ranking of network nodes based on the centrality values assigned by the five centrality measures . We introduced two measures , which we call network integrity measures , to capture various effects of node removal on the ability of other nodes to communicate . An integrity measure maps a set of nodes , S , to a value between 0 and 1 , with the value of 0 being assigned when the removal of S completely disrupts the communication and the value of 1 being assigned when it causes no disruption . Our first measure , shortest-path integrity , quantifies the increase in the length of the shortest path due to the removal of S and is given by , where d ( s , t ) is the length of the shortest path between s and t in the original network , ds ( s , t ) is the length of the shortest path between s and t after the removal of S , and C is a constant . In this work we chose the value of C to be twice the diameter of the original network . Our second measure , edge-disjoint paths integrity , quantifies the decrease in the number of edge-disjoint paths and is given by , where fs ( s , t ) is the number of edge-disjoint paths between s and t in the modified network and f ( s , t ) is this value in the original network . To evaluate the model on the tested networks we used three strategies to estimate the model's parameters: a network simulation procedure , line fitting to points ( log ( 1−PE ) , k ) for k≤k0 , and weighted line fitting to points ( log ( 1−PE ) , k ) for all values of k . ( In weighted line fitting , the contribution of ( log ( 1−PE ) , k ) to the error function is weighted by the fraction of nodes having degree k . ) The first two strategies are described by He et al . [9] . They deem the agreement of parameter values estimated using the network simulation and line fitting strategies to be one of the strongest indications for the validity of the model . But in the tested networks the parameter values estimated using different strategies , as shown in Table S4 , vary considerably . Our method for automatic extraction of putative ECOBIMs is applied to subnetworks induced by proteins annotated with the same biological process GO term . In this work we used a set of 192 biological process terms , which were selected by a group of experts to represent relevant aspects of molecular biology . Thus , the method was applied to 192 subnetworks , one subnetwork at a time . From each GO subnetwork the method extracts groups of densely connected proteins . An ideal dense network is a clique , a complete network where every pair of nodes is adjacent . Over the years various generalizations of the clique concept were proposed in the literature to model a wider set of dense networks . Here we adopt one such generalization based on k-connectivity . We say that a pair of nodes is k-connected if there are k node-disjoint paths in the network between them . We say that a network is k-connected if every pair of nodes is k-connected . For example , a ( k+1 ) -clique , a clique with ( k+1 ) nodes , is k-connected . In fact , it is the smallest k-connected graph . Our method utilizes the following approach to find regions of GO subnetworks that are k-connected: start with a seed that is a ( k+1 ) -clique and iteratively extend the seed through addition of proteins that have at least k neighbors already in the seed . In addition to being k-connected our COBIMs satisfy the following property: nodes can be removed from a COBIM one by one such that the network induced by the remaining nodes is still k-connected . We note that not every k-connected network has this property . Consider , for example , a cycle . The cycle is 2-connected . However , removal of any node results in a path which is 1-connected . The value of parameter k was chosen so that the fraction of COBIM nodes is about 25% of the number of nodes in the network . As shown in Table S5 this results in the following values of k for the tested networks: for the DIP CORE network k = 3 , for the LC and HC networks k = 4 , for the TAP-MS network k = 11 , for the BAYESIAN network k = 4 , and for the Y2H network k = 1 . We also sampled values of k in the neighborhood of selected values and found that the results reported in this paper are however robust with respect to the selected value of k . We note that approaches similar to ours have been previously used by Palla et al [38] and Chesler et al . [39] . Once the COBIMs are computed , the method selects a subset of COBIMs based on the distribution of essential proteins among the COBIM nodes . Namely , the heuristic selects all COBIMs with a fraction of essential proteins that is significantly higher than what would be expected from a uniform distribution of essential genes among the COBIM nodes . More specifically , a COBIM with n nodes and m essential nodes is selected iff:where N is the total number of COBIM nodes and M is the number of essential COBIM nodes .
|
Analysis of protein interaction networks in the budding yeast Saccharomyces cerevisiae has revealed that a small number of proteins , the so-called hubs , interact with a disproportionately large number of other proteins . Furthermore , many hub proteins have been shown to be essential for survival of the cell—that is , in optimal conditions , yeast cannot grow and multiply without them . This relation between essentiality and the number of neighbors in the protein–protein interaction network has been termed the centrality-lethality rule . However , why are such hubs essential ? Jeong and colleagues [1] suggested that overrepresentation of essential proteins among high-degree nodes can be attributed to the central role that hubs play in mediating interactions among numerous , less connected proteins . Another view , proposed by He and Zhang , suggested that that the majority of proteins are essential due to their involvement in one or more essential protein–protein interactions that are distributed uniformly at random along the network edges [2] . We find that none of the above reasons determines essentiality . Instead , the majority of hubs are essential due to their involvement in Essential Complex Biological Modules , a group of densely connected proteins with shared biological function that are enriched in essential proteins . This study sheds new light on the topological complexity of protein interaction networks .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"computational",
"biology/systems",
"biology"
] |
2008
|
Why Do Hubs in the Yeast Protein Interaction Network Tend To Be Essential: Reexamining the Connection between the Network Topology and Essentiality
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.